diff --git "a/data/vsibench_train/qa_mobj_3dBbox_json_conversation.json" "b/data/vsibench_train/qa_mobj_3dBbox_json_conversation.json" deleted file mode 100644--- "a/data/vsibench_train/qa_mobj_3dBbox_json_conversation.json" +++ /dev/null @@ -1,63554 +0,0 @@ -[ - { - "id": 0, - "data_source": "Mono3DRefer", - "scene_name": "000006", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[549.69, 193.27], [572.12, 193.26], [572.68, 194.17], [548.5, 194.19], [549.69, 172.01], [572.12, 172.02], [572.68, 171.28], [548.5, 171.27]]}, {\"category\": \"car\", \"corners_3d\": [[542.08, 205.33], [575.11, 205.64], [543.33, 209.06], [506.16, 208.65], [542.08, 170.08], [575.11, 169.93], [543.33, 168.3], [506.16, 168.49]]}, {\"category\": \"car\", \"corners_3d\": [[227.28, 236.89], [223.37, 241.03], [50.69, 246.21], [67.9, 241.35], [227.28, 186.19], [223.37, 186.54], [50.69, 186.98], [67.9, 186.57]]}, {\"category\": \"car\", \"corners_3d\": [[359.92, 203.91], [329.17, 203.75], [369.41, 201.43], [397.2, 201.56], [359.92, 170.57], [329.17, 170.65], [369.41, 171.78], [397.2, 171.71]]}]\n```" - } - ] - }, - { - "id": 1, - "data_source": "Mono3DRefer", - "scene_name": "000007", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[341.86, 213.8], [330.84, 213.81], [345.1, 211.54], [355.5, 211.52], [341.86, 176.33], [330.84, 176.33], [345.1, 176.14], [355.5, 176.14]]}]\n```" - } - ] - }, - { - "id": 2, - "data_source": "Mono3DRefer", - "scene_name": "000007", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[569.12, 218.69], [614.14, 218.64], [616.66, 224.89], [565.48, 224.96], [569.12, 175.01], [614.14, 175.01], [616.66, 175.31], [565.48, 175.31]]}, {\"category\": \"car\", \"corners_3d\": [[505.76, 202.52], [481.85, 202.54], [490.29, 200.32], [512.41, 200.3], [505.76, 180.42], [481.85, 180.43], [490.29, 179.86], [512.41, 179.86]]}, {\"category\": \"car\", \"corners_3d\": [[562.72, 193.94], [542.22, 193.94], [546.08, 192.58], [565.24, 192.57], [562.72, 175.93], [542.22, 175.93], [546.08, 175.73], [565.24, 175.73]]}]\n```" - } - ] - }, - { - "id": 3, - "data_source": "Mono3DRefer", - "scene_name": "000007", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[569.12, 218.69], [614.14, 218.64], [616.66, 224.89], [565.48, 224.96], [569.12, 175.01], [614.14, 175.01], [616.66, 175.31], [565.48, 175.31]]}, {\"category\": \"car\", \"corners_3d\": [[505.76, 202.52], [481.85, 202.54], [490.29, 200.32], [512.41, 200.3], [505.76, 180.42], [481.85, 180.43], [490.29, 179.86], [512.41, 179.86]]}, {\"category\": \"car\", \"corners_3d\": [[562.72, 193.94], [542.22, 193.94], [546.08, 192.58], [565.24, 192.57], [562.72, 175.93], [542.22, 175.93], [546.08, 175.73], [565.24, 175.73]]}, {\"category\": \"cyclist\", \"corners_3d\": [[341.86, 213.8], [330.84, 213.81], [345.1, 211.54], [355.5, 211.52], [341.86, 176.33], [330.84, 176.33], [345.1, 176.14], [355.5, 176.14]]}]\n```" - } - ] - }, - { - "id": 4, - "data_source": "Mono3DRefer", - "scene_name": "000008", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[219.56, 403.08], [402.7, 423.05], [-270.35, 828.85], [-570.8, 707.32], [219.56, 191.33], [402.7, 192.94], [-270.35, 225.51], [-570.8, 215.76]]}, {\"category\": \"car\", \"corners_3d\": [[487.41, 375.31], [335.78, 359.89], [519.79, 293.74], [624.54, 300.0], [487.41, 182.63], [335.78, 181.88], [519.79, 178.69], [624.54, 178.99]]}, {\"category\": \"car\", \"corners_3d\": [[938.81, 324.02], [1089.87, 331.55], [1281.04, 436.98], [1022.67, 416.76], [938.81, 195.87], [1089.87, 197.02], [1281.04, 213.07], [1022.67, 209.99]]}, {\"category\": \"car\", \"corners_3d\": [[651.17, 240.9], [721.28, 243.06], [685.57, 262.64], [598.07, 259.14], [651.17, 176.35], [721.28, 176.46], [685.57, 177.47], [598.07, 177.29]]}, {\"category\": \"car\", \"corners_3d\": [[779.48, 208.92], [741.67, 208.23], [758.31, 204.44], [792.29, 204.99], [779.48, 169.36], [741.67, 169.42], [758.31, 169.79], [792.29, 169.74]]}, {\"category\": \"car\", \"corners_3d\": [[885.38, 231.89], [944.13, 233.3], [956.12, 240.95], [889.82, 239.15], [885.38, 178.24], [944.13, 178.37], [956.12, 179.07], [889.82, 178.9]]}]\n```" - } - ] - }, - { - "id": 5, - "data_source": "Mono3DRefer", - "scene_name": "000009", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[611.77, 222.54], [658.74, 222.84], [656.38, 230.01], [602.73, 229.62], [611.77, 177.08], [658.74, 177.1], [656.38, 177.71], [602.73, 177.68]]}, {\"category\": \"car\", \"corners_3d\": [[617.92, 193.46], [600.25, 193.36], [608.01, 192.28], [624.78, 192.37], [617.92, 177.33], [600.25, 177.31], [608.01, 177.07], [624.78, 177.09]]}, {\"category\": \"car\", \"corners_3d\": [[591.34, 194.13], [575.14, 194.04], [582.96, 193.04], [598.42, 193.12], [591.34, 178.82], [575.14, 178.8], [582.96, 178.52], [598.42, 178.54]]}]\n```" - } - ] - }, - { - "id": 6, - "data_source": "Mono3DRefer", - "scene_name": "000013", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[517.18, 242.31], [456.21, 242.2], [482.72, 231.21], [534.05, 231.29], [517.18, 185.4], [456.21, 185.38], [482.72, 183.39], [534.05, 183.41]]}]\n```" - } - ] - }, - { - "id": 7, - "data_source": "Mono3DRefer", - "scene_name": "000015", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1126.33, 310.24], [1049.09, 316.32], [987.82, 304.3], [1060.99, 299.24], [1126.33, 134.86], [1049.09, 132.65], [987.82, 137.01], [1060.99, 138.84]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[662.87, 222.63], [687.56, 222.75], [688.35, 224.25], [662.76, 224.11], [662.87, 172.13], [687.56, 172.1], [688.35, 171.76], [662.76, 171.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[686.46, 222.11], [712.88, 222.28], [714.18, 223.9], [686.71, 223.72], [686.46, 169.57], [712.88, 169.52], [714.18, 169.04], [686.71, 169.09]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[534.82, 222.48], [562.5, 222.53], [559.84, 224.34], [530.94, 224.29], [534.82, 168.8], [562.5, 168.79], [559.84, 168.22], [530.94, 168.24]]}]\n```" - } - ] - }, - { - "id": 8, - "data_source": "Mono3DRefer", - "scene_name": "000015", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-207.47, 813.2], [-659.7, 741.26], [233.05, 375.99], [414.79, 384.0], [-207.47, 229.31], [-659.7, 223.86], [233.05, 196.23], [414.79, 196.83]]}]\n```" - } - ] - }, - { - "id": 9, - "data_source": "Mono3DRefer", - "scene_name": "000015", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-207.47, 813.2], [-659.7, 741.26], [233.05, 375.99], [414.79, 384.0], [-207.47, 229.31], [-659.7, 223.86], [233.05, 196.23], [414.79, 196.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1126.33, 310.24], [1049.09, 316.32], [987.82, 304.3], [1060.99, 299.24], [1126.33, 134.86], [1049.09, 132.65], [987.82, 137.01], [1060.99, 138.84]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[662.87, 222.63], [687.56, 222.75], [688.35, 224.25], [662.76, 224.11], [662.87, 172.13], [687.56, 172.1], [688.35, 171.76], [662.76, 171.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[686.46, 222.11], [712.88, 222.28], [714.18, 223.9], [686.71, 223.72], [686.46, 169.57], [712.88, 169.52], [714.18, 169.04], [686.71, 169.09]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[534.82, 222.48], [562.5, 222.53], [559.84, 224.34], [530.94, 224.29], [534.82, 168.8], [562.5, 168.79], [559.84, 168.22], [530.94, 168.24]]}]\n```" - } - ] - }, - { - "id": 10, - "data_source": "Mono3DRefer", - "scene_name": "000016", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[214.82, 256.56], [126.12, 254.22], [422.8, 215.85], [473.48, 216.49], [214.82, 94.76], [126.12, 96.93], [422.8, 132.74], [473.48, 132.14]]}]\n```" - } - ] - }, - { - "id": 11, - "data_source": "Mono3DRefer", - "scene_name": "000016", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[671.52, 195.39], [694.4, 195.54], [686.47, 197.88], [661.31, 197.69], [671.52, 175.01], [694.4, 175.02], [686.47, 175.25], [661.31, 175.23]]}, {\"category\": \"car\", \"corners_3d\": [[701.64, 201.63], [733.67, 202.01], [720.85, 205.75], [684.95, 205.27], [701.64, 172.64], [733.67, 172.64], [720.85, 172.61], [684.95, 172.62]]}, {\"category\": \"car\", \"corners_3d\": [[263.83, 324.62], [385.98, 335.19], [18.02, 461.91], [-148.41, 430.04], [263.83, 192.82], [385.98, 194.21], [18.02, 210.88], [-148.41, 206.69]]}, {\"category\": \"car\", \"corners_3d\": [[515.81, 205.51], [494.7, 205.41], [507.56, 203.25], [527.31, 203.34], [515.81, 182.46], [494.7, 182.42], [507.56, 181.79], [527.31, 181.82]]}]\n```" - } - ] - }, - { - "id": 12, - "data_source": "Mono3DRefer", - "scene_name": "000016", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[648.81, 189.15], [628.49, 189.33], [587.48, 187.01], [605.33, 186.87], [648.81, 159.62], [628.49, 159.48], [587.48, 161.37], [605.33, 161.47]]}]\n```" - } - ] - }, - { - "id": 13, - "data_source": "Mono3DRefer", - "scene_name": "000016", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[671.52, 195.39], [694.4, 195.54], [686.47, 197.88], [661.31, 197.69], [671.52, 175.01], [694.4, 175.02], [686.47, 175.25], [661.31, 175.23]]}, {\"category\": \"truck\", \"corners_3d\": [[214.82, 256.56], [126.12, 254.22], [422.8, 215.85], [473.48, 216.49], [214.82, 94.76], [126.12, 96.93], [422.8, 132.74], [473.48, 132.14]]}, {\"category\": \"car\", \"corners_3d\": [[701.64, 201.63], [733.67, 202.01], [720.85, 205.75], [684.95, 205.27], [701.64, 172.64], [733.67, 172.64], [720.85, 172.61], [684.95, 172.62]]}, {\"category\": \"car\", \"corners_3d\": [[263.83, 324.62], [385.98, 335.19], [18.02, 461.91], [-148.41, 430.04], [263.83, 192.82], [385.98, 194.21], [18.02, 210.88], [-148.41, 206.69]]}, {\"category\": \"car\", \"corners_3d\": [[515.81, 205.51], [494.7, 205.41], [507.56, 203.25], [527.31, 203.34], [515.81, 182.46], [494.7, 182.42], [507.56, 181.79], [527.31, 181.82]]}]\n```" - } - ] - }, - { - "id": 14, - "data_source": "Mono3DRefer", - "scene_name": "000016", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[214.82, 256.56], [126.12, 254.22], [422.8, 215.85], [473.48, 216.49], [214.82, 94.76], [126.12, 96.93], [422.8, 132.74], [473.48, 132.14]]}, {\"category\": \"tram\", \"corners_3d\": [[648.81, 189.15], [628.49, 189.33], [587.48, 187.01], [605.33, 186.87], [648.81, 159.62], [628.49, 159.48], [587.48, 161.37], [605.33, 161.47]]}]\n```" - } - ] - }, - { - "id": 15, - "data_source": "Mono3DRefer", - "scene_name": "000016", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[671.52, 195.39], [694.4, 195.54], [686.47, 197.88], [661.31, 197.69], [671.52, 175.01], [694.4, 175.02], [686.47, 175.25], [661.31, 175.23]]}, {\"category\": \"car\", \"corners_3d\": [[701.64, 201.63], [733.67, 202.01], [720.85, 205.75], [684.95, 205.27], [701.64, 172.64], [733.67, 172.64], [720.85, 172.61], [684.95, 172.62]]}, {\"category\": \"car\", \"corners_3d\": [[263.83, 324.62], [385.98, 335.19], [18.02, 461.91], [-148.41, 430.04], [263.83, 192.82], [385.98, 194.21], [18.02, 210.88], [-148.41, 206.69]]}, {\"category\": \"car\", \"corners_3d\": [[515.81, 205.51], [494.7, 205.41], [507.56, 203.25], [527.31, 203.34], [515.81, 182.46], [494.7, 182.42], [507.56, 181.79], [527.31, 181.82]]}, {\"category\": \"tram\", \"corners_3d\": [[648.81, 189.15], [628.49, 189.33], [587.48, 187.01], [605.33, 186.87], [648.81, 159.62], [628.49, 159.48], [587.48, 161.37], [605.33, 161.47]]}]\n```" - } - ] - }, - { - "id": 16, - "data_source": "Mono3DRefer", - "scene_name": "000016", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[671.52, 195.39], [694.4, 195.54], [686.47, 197.88], [661.31, 197.69], [671.52, 175.01], [694.4, 175.02], [686.47, 175.25], [661.31, 175.23]]}, {\"category\": \"truck\", \"corners_3d\": [[214.82, 256.56], [126.12, 254.22], [422.8, 215.85], [473.48, 216.49], [214.82, 94.76], [126.12, 96.93], [422.8, 132.74], [473.48, 132.14]]}, {\"category\": \"car\", \"corners_3d\": [[701.64, 201.63], [733.67, 202.01], [720.85, 205.75], [684.95, 205.27], [701.64, 172.64], [733.67, 172.64], [720.85, 172.61], [684.95, 172.62]]}, {\"category\": \"car\", \"corners_3d\": [[263.83, 324.62], [385.98, 335.19], [18.02, 461.91], [-148.41, 430.04], [263.83, 192.82], [385.98, 194.21], [18.02, 210.88], [-148.41, 206.69]]}, {\"category\": \"car\", \"corners_3d\": [[515.81, 205.51], [494.7, 205.41], [507.56, 203.25], [527.31, 203.34], [515.81, 182.46], [494.7, 182.42], [507.56, 181.79], [527.31, 181.82]]}, {\"category\": \"tram\", \"corners_3d\": [[648.81, 189.15], [628.49, 189.33], [587.48, 187.01], [605.33, 186.87], [648.81, 159.62], [628.49, 159.48], [587.48, 161.37], [605.33, 161.47]]}]\n```" - } - ] - }, - { - "id": 17, - "data_source": "Mono3DRefer", - "scene_name": "000017", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[728.53, 198.46], [749.17, 198.07], [791.62, 200.23], [769.91, 200.69], [728.53, 174.48], [749.17, 174.45], [791.62, 174.59], [769.91, 174.62]]}]\n```" - } - ] - }, - { - "id": 18, - "data_source": "Mono3DRefer", - "scene_name": "000018", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[473.33, 213.73], [439.96, 214.77], [387.65, 206.76], [415.82, 206.07], [473.33, 167.66], [439.96, 167.52], [387.65, 168.54], [415.82, 168.63]]}]\n```" - } - ] - }, - { - "id": 19, - "data_source": "Mono3DRefer", - "scene_name": "000018", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[523.67, 215.4], [548.1, 215.23], [550.6, 219.48], [523.72, 219.69], [523.67, 194.27], [548.1, 194.19], [550.6, 196.32], [523.72, 196.43]]}]\n```" - } - ] - }, - { - "id": 20, - "data_source": "Mono3DRefer", - "scene_name": "000018", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[572.64, 215.17], [597.35, 214.97], [606.36, 219.24], [579.18, 219.48], [572.64, 185.59], [597.35, 185.53], [606.36, 186.81], [579.18, 186.89]]}]\n```" - } - ] - }, - { - "id": 21, - "data_source": "Mono3DRefer", - "scene_name": "000018", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[473.33, 213.73], [439.96, 214.77], [387.65, 206.76], [415.82, 206.07], [473.33, 167.66], [439.96, 167.52], [387.65, 168.54], [415.82, 168.63]]}, {\"category\": \"car\", \"corners_3d\": [[523.67, 215.4], [548.1, 215.23], [550.6, 219.48], [523.72, 219.69], [523.67, 194.27], [548.1, 194.19], [550.6, 196.32], [523.72, 196.43]]}]\n```" - } - ] - }, - { - "id": 22, - "data_source": "Mono3DRefer", - "scene_name": "000018", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[572.64, 215.17], [597.35, 214.97], [606.36, 219.24], [579.18, 219.48], [572.64, 185.59], [597.35, 185.53], [606.36, 186.81], [579.18, 186.89]]}, {\"category\": \"tram\", \"corners_3d\": [[473.33, 213.73], [439.96, 214.77], [387.65, 206.76], [415.82, 206.07], [473.33, 167.66], [439.96, 167.52], [387.65, 168.54], [415.82, 168.63]]}]\n```" - } - ] - }, - { - "id": 23, - "data_source": "Mono3DRefer", - "scene_name": "000018", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[572.64, 215.17], [597.35, 214.97], [606.36, 219.24], [579.18, 219.48], [572.64, 185.59], [597.35, 185.53], [606.36, 186.81], [579.18, 186.89]]}, {\"category\": \"car\", \"corners_3d\": [[523.67, 215.4], [548.1, 215.23], [550.6, 219.48], [523.72, 219.69], [523.67, 194.27], [548.1, 194.19], [550.6, 196.32], [523.72, 196.43]]}]\n```" - } - ] - }, - { - "id": 24, - "data_source": "Mono3DRefer", - "scene_name": "000018", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[572.64, 215.17], [597.35, 214.97], [606.36, 219.24], [579.18, 219.48], [572.64, 185.59], [597.35, 185.53], [606.36, 186.81], [579.18, 186.89]]}, {\"category\": \"tram\", \"corners_3d\": [[473.33, 213.73], [439.96, 214.77], [387.65, 206.76], [415.82, 206.07], [473.33, 167.66], [439.96, 167.52], [387.65, 168.54], [415.82, 168.63]]}, {\"category\": \"car\", \"corners_3d\": [[523.67, 215.4], [548.1, 215.23], [550.6, 219.48], [523.72, 219.69], [523.67, 194.27], [548.1, 194.19], [550.6, 196.32], [523.72, 196.43]]}]\n```" - } - ] - }, - { - "id": 25, - "data_source": "Mono3DRefer", - "scene_name": "000019", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[743.73, 274.21], [844.19, 274.64], [944.6, 322.76], [796.32, 321.82], [743.73, 183.4], [844.19, 183.45], [944.6, 188.46], [796.32, 188.36]]}, {\"category\": \"car\", \"corners_3d\": [[555.95, 202.38], [575.26, 202.4], [571.92, 204.41], [551.3, 204.39], [555.95, 183.87], [575.26, 183.87], [571.92, 184.62], [551.3, 184.61]]}]\n```" - } - ] - }, - { - "id": 26, - "data_source": "Mono3DRefer", - "scene_name": "000019", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[65.05, 636.57], [-475.4, 636.84], [243.82, 329.1], [425.74, 329.07], [65.05, -44.73], [-475.4, -44.86], [243.82, 99.54], [425.74, 99.55]]}]\n```" - } - ] - }, - { - "id": 27, - "data_source": "Mono3DRefer", - "scene_name": "000019", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[639.76, 208.59], [674.74, 208.59], [683.33, 213.35], [643.7, 213.34], [639.76, 170.73], [674.74, 170.73], [683.33, 170.45], [643.7, 170.45]]}]\n```" - } - ] - }, - { - "id": 28, - "data_source": "Mono3DRefer", - "scene_name": "000019", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[65.05, 636.57], [-475.4, 636.84], [243.82, 329.1], [425.74, 329.07], [65.05, -44.73], [-475.4, -44.86], [243.82, 99.54], [425.74, 99.55]]}, {\"category\": \"car\", \"corners_3d\": [[743.73, 274.21], [844.19, 274.64], [944.6, 322.76], [796.32, 321.82], [743.73, 183.4], [844.19, 183.45], [944.6, 188.46], [796.32, 188.36]]}, {\"category\": \"car\", \"corners_3d\": [[555.95, 202.38], [575.26, 202.4], [571.92, 204.41], [551.3, 204.39], [555.95, 183.87], [575.26, 183.87], [571.92, 184.62], [551.3, 184.61]]}]\n```" - } - ] - }, - { - "id": 29, - "data_source": "Mono3DRefer", - "scene_name": "000019", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[743.73, 274.21], [844.19, 274.64], [944.6, 322.76], [796.32, 321.82], [743.73, 183.4], [844.19, 183.45], [944.6, 188.46], [796.32, 188.36]]}, {\"category\": \"van\", \"corners_3d\": [[639.76, 208.59], [674.74, 208.59], [683.33, 213.35], [643.7, 213.34], [639.76, 170.73], [674.74, 170.73], [683.33, 170.45], [643.7, 170.45]]}, {\"category\": \"car\", \"corners_3d\": [[555.95, 202.38], [575.26, 202.4], [571.92, 204.41], [551.3, 204.39], [555.95, 183.87], [575.26, 183.87], [571.92, 184.62], [551.3, 184.61]]}]\n```" - } - ] - }, - { - "id": 30, - "data_source": "Mono3DRefer", - "scene_name": "000019", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[65.05, 636.57], [-475.4, 636.84], [243.82, 329.1], [425.74, 329.07], [65.05, -44.73], [-475.4, -44.86], [243.82, 99.54], [425.74, 99.55]]}, {\"category\": \"van\", \"corners_3d\": [[639.76, 208.59], [674.74, 208.59], [683.33, 213.35], [643.7, 213.34], [639.76, 170.73], [674.74, 170.73], [683.33, 170.45], [643.7, 170.45]]}]\n```" - } - ] - }, - { - "id": 31, - "data_source": "Mono3DRefer", - "scene_name": "000019", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[65.05, 636.57], [-475.4, 636.84], [243.82, 329.1], [425.74, 329.07], [65.05, -44.73], [-475.4, -44.86], [243.82, 99.54], [425.74, 99.55]]}, {\"category\": \"car\", \"corners_3d\": [[743.73, 274.21], [844.19, 274.64], [944.6, 322.76], [796.32, 321.82], [743.73, 183.4], [844.19, 183.45], [944.6, 188.46], [796.32, 188.36]]}, {\"category\": \"van\", \"corners_3d\": [[639.76, 208.59], [674.74, 208.59], [683.33, 213.35], [643.7, 213.34], [639.76, 170.73], [674.74, 170.73], [683.33, 170.45], [643.7, 170.45]]}, {\"category\": \"car\", \"corners_3d\": [[555.95, 202.38], [575.26, 202.4], [571.92, 204.41], [551.3, 204.39], [555.95, 183.87], [575.26, 183.87], [571.92, 184.62], [551.3, 184.61]]}]\n```" - } - ] - }, - { - "id": 32, - "data_source": "Mono3DRefer", - "scene_name": "000020", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[685.23, 237.94], [750.18, 237.71], [805.1, 258.79], [719.19, 259.2], [685.23, 180.98], [750.18, 180.95], [805.1, 183.58], [719.19, 183.63]]}]\n```" - } - ] - }, - { - "id": 33, - "data_source": "Mono3DRefer", - "scene_name": "000021", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1061.65, 468.95], [1156.53, 469.75], [1615.01, 725.63], [1436.47, 722.86], [1061.65, 188.66], [1156.53, 188.7], [1615.01, 202.36], [1436.47, 202.21]]}]\n```" - } - ] - }, - { - "id": 34, - "data_source": "Mono3DRefer", - "scene_name": "000021", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1053.58, 209.26], [1100.77, 213.06], [900.13, 213.18], [871.87, 209.35], [1053.58, 121.93], [1100.77, 116.6], [900.13, 116.44], [871.87, 121.8]]}]\n```" - } - ] - }, - { - "id": 35, - "data_source": "Mono3DRefer", - "scene_name": "000021", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[459.26, 272.77], [359.43, 271.22], [440.44, 247.02], [516.38, 247.9], [459.26, 181.11], [359.43, 180.98], [440.44, 178.98], [516.38, 179.05]]}, {\"category\": \"car\", \"corners_3d\": [[523.09, 248.36], [445.34, 247.45], [497.57, 231.44], [558.96, 232.0], [523.09, 165.33], [445.34, 165.42], [497.57, 167.02], [558.96, 166.96]]}, {\"category\": \"car\", \"corners_3d\": [[724.76, 209.47], [769.23, 209.57], [790.56, 215.58], [738.76, 215.46], [724.76, 172.07], [769.23, 172.07], [790.56, 171.94], [738.76, 171.94]]}, {\"category\": \"car\", \"corners_3d\": [[582.33, 217.39], [535.01, 217.1], [554.98, 211.06], [595.92, 211.28], [582.33, 167.02], [535.01, 167.06], [554.98, 167.85], [595.92, 167.82]]}, {\"category\": \"car\", \"corners_3d\": [[710.28, 201.82], [750.31, 201.9], [765.5, 206.14], [719.6, 206.04], [710.28, 164.45], [750.31, 164.42], [765.5, 163.19], [719.6, 163.22]]}, {\"category\": \"car\", \"corners_3d\": [[598.17, 207.15], [562.3, 207.0], [573.24, 203.66], [605.63, 203.78], [598.17, 166.85], [562.3, 166.87], [573.24, 167.46], [605.63, 167.44]]}]\n```" - } - ] - }, - { - "id": 36, - "data_source": "Mono3DRefer", - "scene_name": "000021", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1061.65, 468.95], [1156.53, 469.75], [1615.01, 725.63], [1436.47, 722.86], [1061.65, 188.66], [1156.53, 188.7], [1615.01, 202.36], [1436.47, 202.21]]}, {\"category\": \"van\", \"corners_3d\": [[1053.58, 209.26], [1100.77, 213.06], [900.13, 213.18], [871.87, 209.35], [1053.58, 121.93], [1100.77, 116.6], [900.13, 116.44], [871.87, 121.8]]}]\n```" - } - ] - }, - { - "id": 37, - "data_source": "Mono3DRefer", - "scene_name": "000021", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1061.65, 468.95], [1156.53, 469.75], [1615.01, 725.63], [1436.47, 722.86], [1061.65, 188.66], [1156.53, 188.7], [1615.01, 202.36], [1436.47, 202.21]]}, {\"category\": \"car\", \"corners_3d\": [[459.26, 272.77], [359.43, 271.22], [440.44, 247.02], [516.38, 247.9], [459.26, 181.11], [359.43, 180.98], [440.44, 178.98], [516.38, 179.05]]}, {\"category\": \"car\", \"corners_3d\": [[523.09, 248.36], [445.34, 247.45], [497.57, 231.44], [558.96, 232.0], [523.09, 165.33], [445.34, 165.42], [497.57, 167.02], [558.96, 166.96]]}, {\"category\": \"car\", \"corners_3d\": [[724.76, 209.47], [769.23, 209.57], [790.56, 215.58], [738.76, 215.46], [724.76, 172.07], [769.23, 172.07], [790.56, 171.94], [738.76, 171.94]]}, {\"category\": \"car\", \"corners_3d\": [[582.33, 217.39], [535.01, 217.1], [554.98, 211.06], [595.92, 211.28], [582.33, 167.02], [535.01, 167.06], [554.98, 167.85], [595.92, 167.82]]}, {\"category\": \"car\", \"corners_3d\": [[710.28, 201.82], [750.31, 201.9], [765.5, 206.14], [719.6, 206.04], [710.28, 164.45], [750.31, 164.42], [765.5, 163.19], [719.6, 163.22]]}, {\"category\": \"car\", \"corners_3d\": [[598.17, 207.15], [562.3, 207.0], [573.24, 203.66], [605.63, 203.78], [598.17, 166.85], [562.3, 166.87], [573.24, 167.46], [605.63, 167.44]]}]\n```" - } - ] - }, - { - "id": 38, - "data_source": "Mono3DRefer", - "scene_name": "000021", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[459.26, 272.77], [359.43, 271.22], [440.44, 247.02], [516.38, 247.9], [459.26, 181.11], [359.43, 180.98], [440.44, 178.98], [516.38, 179.05]]}, {\"category\": \"car\", \"corners_3d\": [[523.09, 248.36], [445.34, 247.45], [497.57, 231.44], [558.96, 232.0], [523.09, 165.33], [445.34, 165.42], [497.57, 167.02], [558.96, 166.96]]}, {\"category\": \"van\", \"corners_3d\": [[1053.58, 209.26], [1100.77, 213.06], [900.13, 213.18], [871.87, 209.35], [1053.58, 121.93], [1100.77, 116.6], [900.13, 116.44], [871.87, 121.8]]}, {\"category\": \"car\", \"corners_3d\": [[724.76, 209.47], [769.23, 209.57], [790.56, 215.58], [738.76, 215.46], [724.76, 172.07], [769.23, 172.07], [790.56, 171.94], [738.76, 171.94]]}, {\"category\": \"car\", \"corners_3d\": [[582.33, 217.39], [535.01, 217.1], [554.98, 211.06], [595.92, 211.28], [582.33, 167.02], [535.01, 167.06], [554.98, 167.85], [595.92, 167.82]]}, {\"category\": \"car\", \"corners_3d\": [[710.28, 201.82], [750.31, 201.9], [765.5, 206.14], [719.6, 206.04], [710.28, 164.45], [750.31, 164.42], [765.5, 163.19], [719.6, 163.22]]}, {\"category\": \"car\", \"corners_3d\": [[598.17, 207.15], [562.3, 207.0], [573.24, 203.66], [605.63, 203.78], [598.17, 166.85], [562.3, 166.87], [573.24, 167.46], [605.63, 167.44]]}]\n```" - } - ] - }, - { - "id": 39, - "data_source": "Mono3DRefer", - "scene_name": "000021", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1061.65, 468.95], [1156.53, 469.75], [1615.01, 725.63], [1436.47, 722.86], [1061.65, 188.66], [1156.53, 188.7], [1615.01, 202.36], [1436.47, 202.21]]}, {\"category\": \"car\", \"corners_3d\": [[459.26, 272.77], [359.43, 271.22], [440.44, 247.02], [516.38, 247.9], [459.26, 181.11], [359.43, 180.98], [440.44, 178.98], [516.38, 179.05]]}, {\"category\": \"car\", \"corners_3d\": [[523.09, 248.36], [445.34, 247.45], [497.57, 231.44], [558.96, 232.0], [523.09, 165.33], [445.34, 165.42], [497.57, 167.02], [558.96, 166.96]]}, {\"category\": \"van\", \"corners_3d\": [[1053.58, 209.26], [1100.77, 213.06], [900.13, 213.18], [871.87, 209.35], [1053.58, 121.93], [1100.77, 116.6], [900.13, 116.44], [871.87, 121.8]]}, {\"category\": \"car\", \"corners_3d\": [[724.76, 209.47], [769.23, 209.57], [790.56, 215.58], [738.76, 215.46], [724.76, 172.07], [769.23, 172.07], [790.56, 171.94], [738.76, 171.94]]}, {\"category\": \"car\", \"corners_3d\": [[582.33, 217.39], [535.01, 217.1], [554.98, 211.06], [595.92, 211.28], [582.33, 167.02], [535.01, 167.06], [554.98, 167.85], [595.92, 167.82]]}, {\"category\": \"car\", \"corners_3d\": [[710.28, 201.82], [750.31, 201.9], [765.5, 206.14], [719.6, 206.04], [710.28, 164.45], [750.31, 164.42], [765.5, 163.19], [719.6, 163.22]]}, {\"category\": \"car\", \"corners_3d\": [[598.17, 207.15], [562.3, 207.0], [573.24, 203.66], [605.63, 203.78], [598.17, 166.85], [562.3, 166.87], [573.24, 167.46], [605.63, 167.44]]}]\n```" - } - ] - }, - { - "id": 40, - "data_source": "Mono3DRefer", - "scene_name": "000022", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[180.92, 229.55], [218.74, 229.33], [172.43, 237.17], [129.13, 237.46], [180.92, 199.17], [218.74, 199.06], [172.43, 202.7], [129.13, 202.84]]}, {\"category\": \"car\", \"corners_3d\": [[121.6, 230.61], [159.82, 230.4], [104.49, 238.37], [60.74, 238.63], [121.6, 197.97], [159.82, 197.88], [104.49, 201.34], [60.74, 201.46]]}]\n```" - } - ] - }, - { - "id": 41, - "data_source": "Mono3DRefer", - "scene_name": "000023", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[678.14, 194.44], [687.61, 194.45], [690.85, 195.44], [680.94, 195.44], [678.14, 165.12], [687.61, 165.12], [690.85, 164.77], [680.94, 164.77]]}]\n```" - } - ] - }, - { - "id": 42, - "data_source": "Mono3DRefer", - "scene_name": "000023", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[398.49, 205.78], [373.14, 205.78], [388.43, 203.65], [412.13, 203.65], [398.49, 183.15], [373.14, 183.15], [388.43, 182.48], [412.13, 182.48]]}]\n```" - } - ] - }, - { - "id": 43, - "data_source": "Mono3DRefer", - "scene_name": "000023", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[601.62, 187.38], [626.77, 187.38], [630.02, 190.22], [599.96, 190.22], [601.62, 160.13], [626.77, 160.13], [630.02, 157.65], [599.96, 157.65]]}]\n```" - } - ] - }, - { - "id": 44, - "data_source": "Mono3DRefer", - "scene_name": "000023", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[398.49, 205.78], [373.14, 205.78], [388.43, 203.65], [412.13, 203.65], [398.49, 183.15], [373.14, 183.15], [388.43, 182.48], [412.13, 182.48]]}, {\"category\": \"cyclist\", \"corners_3d\": [[678.14, 194.44], [687.61, 194.45], [690.85, 195.44], [680.94, 195.44], [678.14, 165.12], [687.61, 165.12], [690.85, 164.77], [680.94, 164.77]]}]\n```" - } - ] - }, - { - "id": 45, - "data_source": "Mono3DRefer", - "scene_name": "000023", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[601.62, 187.38], [626.77, 187.38], [630.02, 190.22], [599.96, 190.22], [601.62, 160.13], [626.77, 160.13], [630.02, 157.65], [599.96, 157.65]]}, {\"category\": \"cyclist\", \"corners_3d\": [[678.14, 194.44], [687.61, 194.45], [690.85, 195.44], [680.94, 195.44], [678.14, 165.12], [687.61, 165.12], [690.85, 164.77], [680.94, 164.77]]}]\n```" - } - ] - }, - { - "id": 46, - "data_source": "Mono3DRefer", - "scene_name": "000023", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[601.62, 187.38], [626.77, 187.38], [630.02, 190.22], [599.96, 190.22], [601.62, 160.13], [626.77, 160.13], [630.02, 157.65], [599.96, 157.65]]}, {\"category\": \"car\", \"corners_3d\": [[398.49, 205.78], [373.14, 205.78], [388.43, 203.65], [412.13, 203.65], [398.49, 183.15], [373.14, 183.15], [388.43, 182.48], [412.13, 182.48]]}]\n```" - } - ] - }, - { - "id": 47, - "data_source": "Mono3DRefer", - "scene_name": "000023", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[601.62, 187.38], [626.77, 187.38], [630.02, 190.22], [599.96, 190.22], [601.62, 160.13], [626.77, 160.13], [630.02, 157.65], [599.96, 157.65]]}, {\"category\": \"car\", \"corners_3d\": [[398.49, 205.78], [373.14, 205.78], [388.43, 203.65], [412.13, 203.65], [398.49, 183.15], [373.14, 183.15], [388.43, 182.48], [412.13, 182.48]]}, {\"category\": \"cyclist\", \"corners_3d\": [[678.14, 194.44], [687.61, 194.45], [690.85, 195.44], [680.94, 195.44], [678.14, 165.12], [687.61, 165.12], [690.85, 164.77], [680.94, 164.77]]}]\n```" - } - ] - }, - { - "id": 48, - "data_source": "Mono3DRefer", - "scene_name": "000025", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[898.5, 428.76], [1129.95, 435.2], [2045.13, 954.91], [1323.31, 900.35], [898.5, 216.98], [1129.95, 218.09], [2045.13, 307.71], [1323.31, 298.31]]}, {\"category\": \"car\", \"corners_3d\": [[445.47, 268.36], [538.18, 269.38], [482.7, 310.64], [351.77, 308.58], [445.47, 182.21], [538.18, 182.31], [482.7, 186.35], [351.77, 186.15]]}, {\"category\": \"car\", \"corners_3d\": [[603.96, 217.87], [562.9, 217.48], [582.26, 211.91], [618.31, 212.21], [603.96, 174.13], [562.9, 174.12], [582.26, 173.96], [618.31, 173.97]]}, {\"category\": \"car\", \"corners_3d\": [[725.11, 240.08], [787.22, 240.55], [805.35, 250.84], [733.73, 250.23], [725.11, 178.99], [787.22, 179.03], [805.35, 179.97], [733.73, 179.91]]}, {\"category\": \"car\", \"corners_3d\": [[721.13, 227.85], [769.91, 228.33], [780.05, 236.93], [723.69, 236.29], [721.13, 186.67], [769.91, 186.79], [780.05, 188.95], [723.69, 188.79]]}]\n```" - } - ] - }, - { - "id": 49, - "data_source": "Mono3DRefer", - "scene_name": "000025", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[659.57, 205.31], [650.54, 205.27], [651.16, 204.42], [659.95, 204.45], [659.57, 175.63], [650.54, 175.63], [651.16, 175.56], [659.95, 175.56]]}]\n```" - } - ] - }, - { - "id": 50, - "data_source": "Mono3DRefer", - "scene_name": "000025", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[898.5, 428.76], [1129.95, 435.2], [2045.13, 954.91], [1323.31, 900.35], [898.5, 216.98], [1129.95, 218.09], [2045.13, 307.71], [1323.31, 298.31]]}, {\"category\": \"car\", \"corners_3d\": [[445.47, 268.36], [538.18, 269.38], [482.7, 310.64], [351.77, 308.58], [445.47, 182.21], [538.18, 182.31], [482.7, 186.35], [351.77, 186.15]]}, {\"category\": \"car\", \"corners_3d\": [[603.96, 217.87], [562.9, 217.48], [582.26, 211.91], [618.31, 212.21], [603.96, 174.13], [562.9, 174.12], [582.26, 173.96], [618.31, 173.97]]}, {\"category\": \"car\", \"corners_3d\": [[725.11, 240.08], [787.22, 240.55], [805.35, 250.84], [733.73, 250.23], [725.11, 178.99], [787.22, 179.03], [805.35, 179.97], [733.73, 179.91]]}, {\"category\": \"cyclist\", \"corners_3d\": [[659.57, 205.31], [650.54, 205.27], [651.16, 204.42], [659.95, 204.45], [659.57, 175.63], [650.54, 175.63], [651.16, 175.56], [659.95, 175.56]]}, {\"category\": \"car\", \"corners_3d\": [[721.13, 227.85], [769.91, 228.33], [780.05, 236.93], [723.69, 236.29], [721.13, 186.67], [769.91, 186.79], [780.05, 188.95], [723.69, 188.79]]}]\n```" - } - ] - }, - { - "id": 51, - "data_source": "Mono3DRefer", - "scene_name": "000026", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[679.88, 212.4], [634.35, 211.6], [665.35, 205.98], [704.71, 206.56], [679.88, 155.84], [634.35, 156.18], [665.35, 158.6], [704.71, 158.35]]}]\n```" - } - ] - }, - { - "id": 52, - "data_source": "Mono3DRefer", - "scene_name": "000026", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[555.62, 190.37], [571.94, 190.31], [576.15, 191.25], [558.97, 191.32], [555.62, 174.6], [571.94, 174.6], [576.15, 174.69], [558.97, 174.7]]}]\n```" - } - ] - }, - { - "id": 53, - "data_source": "Mono3DRefer", - "scene_name": "000026", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[679.88, 212.4], [634.35, 211.6], [665.35, 205.98], [704.71, 206.56], [679.88, 155.84], [634.35, 156.18], [665.35, 158.6], [704.71, 158.35]]}, {\"category\": \"car\", \"corners_3d\": [[555.62, 190.37], [571.94, 190.31], [576.15, 191.25], [558.97, 191.32], [555.62, 174.6], [571.94, 174.6], [576.15, 174.69], [558.97, 174.7]]}]\n```" - } - ] - }, - { - "id": 54, - "data_source": "Mono3DRefer", - "scene_name": "000029", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[652.7, 202.63], [680.76, 202.56], [690.16, 205.01], [659.8, 205.09], [652.7, 175.19], [680.76, 175.18], [690.16, 175.38], [659.8, 175.38]]}]\n```" - } - ] - }, - { - "id": 55, - "data_source": "Mono3DRefer", - "scene_name": "000033", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[73.59, 323.89], [-68.09, 323.92], [135.37, 278.52], [234.46, 278.51], [73.59, 196.13], [-68.09, 196.13], [135.37, 189.14], [234.46, 189.14]]}, {\"category\": \"car\", \"corners_3d\": [[583.07, 188.97], [600.5, 188.96], [601.18, 189.88], [582.76, 189.89], [583.07, 175.25], [600.5, 175.25], [601.18, 175.38], [582.76, 175.38]]}]\n```" - } - ] - }, - { - "id": 56, - "data_source": "Mono3DRefer", - "scene_name": "000035", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[587.84, 198.29], [612.68, 198.27], [614.05, 200.33], [587.21, 200.35], [587.84, 173.74], [612.68, 173.74], [614.05, 173.81], [587.21, 173.81]]}, {\"category\": \"car\", \"corners_3d\": [[470.67, 239.31], [529.35, 239.31], [503.5, 260.58], [426.04, 260.57], [470.67, 176.73], [529.35, 176.73], [503.5, 177.97], [426.04, 177.97]]}, {\"category\": \"car\", \"corners_3d\": [[284.02, 231.85], [235.17, 232.01], [277.81, 224.7], [320.53, 224.57], [284.02, 187.04], [235.17, 187.08], [277.81, 185.32], [320.53, 185.29]]}]\n```" - } - ] - }, - { - "id": 57, - "data_source": "Mono3DRefer", - "scene_name": "000035", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[666.03, 216.09], [675.77, 215.46], [717.94, 217.77], [708.32, 218.46], [666.03, 160.99], [675.77, 161.16], [717.94, 160.53], [708.32, 160.34]]}]\n```" - } - ] - }, - { - "id": 58, - "data_source": "Mono3DRefer", - "scene_name": "000035", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[434.16, 223.3], [431.61, 224.14], [422.1, 224.16], [424.8, 223.32], [434.16, 173.77], [431.61, 173.79], [422.1, 173.79], [424.8, 173.77]]}]\n```" - } - ] - }, - { - "id": 59, - "data_source": "Mono3DRefer", - "scene_name": "000035", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[587.84, 198.29], [612.68, 198.27], [614.05, 200.33], [587.21, 200.35], [587.84, 173.74], [612.68, 173.74], [614.05, 173.81], [587.21, 173.81]]}, {\"category\": \"car\", \"corners_3d\": [[470.67, 239.31], [529.35, 239.31], [503.5, 260.58], [426.04, 260.57], [470.67, 176.73], [529.35, 176.73], [503.5, 177.97], [426.04, 177.97]]}, {\"category\": \"car\", \"corners_3d\": [[284.02, 231.85], [235.17, 232.01], [277.81, 224.7], [320.53, 224.57], [284.02, 187.04], [235.17, 187.08], [277.81, 185.32], [320.53, 185.29]]}, {\"category\": \"cyclist\", \"corners_3d\": [[666.03, 216.09], [675.77, 215.46], [717.94, 217.77], [708.32, 218.46], [666.03, 160.99], [675.77, 161.16], [717.94, 160.53], [708.32, 160.34]]}]\n```" - } - ] - }, - { - "id": 60, - "data_source": "Mono3DRefer", - "scene_name": "000035", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[587.84, 198.29], [612.68, 198.27], [614.05, 200.33], [587.21, 200.35], [587.84, 173.74], [612.68, 173.74], [614.05, 173.81], [587.21, 173.81]]}, {\"category\": \"car\", \"corners_3d\": [[470.67, 239.31], [529.35, 239.31], [503.5, 260.58], [426.04, 260.57], [470.67, 176.73], [529.35, 176.73], [503.5, 177.97], [426.04, 177.97]]}, {\"category\": \"car\", \"corners_3d\": [[284.02, 231.85], [235.17, 232.01], [277.81, 224.7], [320.53, 224.57], [284.02, 187.04], [235.17, 187.08], [277.81, 185.32], [320.53, 185.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[434.16, 223.3], [431.61, 224.14], [422.1, 224.16], [424.8, 223.32], [434.16, 173.77], [431.61, 173.79], [422.1, 173.79], [424.8, 173.77]]}]\n```" - } - ] - }, - { - "id": 61, - "data_source": "Mono3DRefer", - "scene_name": "000035", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[666.03, 216.09], [675.77, 215.46], [717.94, 217.77], [708.32, 218.46], [666.03, 160.99], [675.77, 161.16], [717.94, 160.53], [708.32, 160.34]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[434.16, 223.3], [431.61, 224.14], [422.1, 224.16], [424.8, 223.32], [434.16, 173.77], [431.61, 173.79], [422.1, 173.79], [424.8, 173.77]]}]\n```" - } - ] - }, - { - "id": 62, - "data_source": "Mono3DRefer", - "scene_name": "000035", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[587.84, 198.29], [612.68, 198.27], [614.05, 200.33], [587.21, 200.35], [587.84, 173.74], [612.68, 173.74], [614.05, 173.81], [587.21, 173.81]]}, {\"category\": \"car\", \"corners_3d\": [[470.67, 239.31], [529.35, 239.31], [503.5, 260.58], [426.04, 260.57], [470.67, 176.73], [529.35, 176.73], [503.5, 177.97], [426.04, 177.97]]}, {\"category\": \"car\", \"corners_3d\": [[284.02, 231.85], [235.17, 232.01], [277.81, 224.7], [320.53, 224.57], [284.02, 187.04], [235.17, 187.08], [277.81, 185.32], [320.53, 185.29]]}, {\"category\": \"cyclist\", \"corners_3d\": [[666.03, 216.09], [675.77, 215.46], [717.94, 217.77], [708.32, 218.46], [666.03, 160.99], [675.77, 161.16], [717.94, 160.53], [708.32, 160.34]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[434.16, 223.3], [431.61, 224.14], [422.1, 224.16], [424.8, 223.32], [434.16, 173.77], [431.61, 173.79], [422.1, 173.79], [424.8, 173.77]]}]\n```" - } - ] - }, - { - "id": 63, - "data_source": "Mono3DRefer", - "scene_name": "000036", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[570.17, 273.17], [669.9, 273.18], [693.32, 312.5], [554.51, 312.47], [570.17, 178.34], [669.9, 178.34], [693.32, 180.49], [554.51, 180.49]]}, {\"category\": \"car\", \"corners_3d\": [[371.21, 218.88], [341.28, 218.9], [362.77, 215.1], [390.23, 215.09], [371.21, 196.44], [341.28, 196.45], [362.77, 194.5], [390.23, 194.49]]}, {\"category\": \"car\", \"corners_3d\": [[315.21, 220.99], [286.28, 220.85], [309.97, 217.81], [337.12, 217.93], [315.21, 197.92], [286.28, 197.85], [309.97, 196.26], [337.12, 196.33]]}, {\"category\": \"car\", \"corners_3d\": [[439.21, 212.82], [412.98, 212.8], [431.27, 209.21], [455.14, 209.22], [439.21, 191.78], [412.98, 191.77], [431.27, 190.07], [455.14, 190.08]]}, {\"category\": \"car\", \"corners_3d\": [[981.84, 229.33], [960.91, 225.81], [1075.61, 226.23], [1104.42, 229.82], [981.84, 185.09], [960.91, 184.32], [1075.61, 184.42], [1104.42, 185.19]]}, {\"category\": \"car\", \"corners_3d\": [[1188.51, 222.02], [1155.43, 218.9], [1255.91, 219.39], [1296.27, 222.59], [1188.51, 178.71], [1155.43, 178.34], [1255.91, 178.4], [1296.27, 178.78]]}, {\"category\": \"car\", \"corners_3d\": [[1126.07, 365.63], [1373.6, 382.9], [2283.7, 742.04], [1517.84, 630.85], [1126.07, 211.85], [1373.6, 215.34], [2283.7, 287.99], [1517.84, 265.5]]}]\n```" - } - ] - }, - { - "id": 64, - "data_source": "Mono3DRefer", - "scene_name": "000039", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[357.53, 420.45], [115.12, 422.16], [336.63, 306.98], [466.63, 306.48], [357.53, 208.59], [115.12, 208.84], [336.63, 192.21], [466.63, 192.14]]}, {\"category\": \"car\", \"corners_3d\": [[683.78, 240.76], [743.91, 240.77], [776.49, 257.31], [701.72, 257.3], [683.78, 183.77], [743.91, 183.77], [776.49, 186.43], [701.72, 186.43]]}, {\"category\": \"car\", \"corners_3d\": [[569.54, 213.24], [533.7, 213.25], [542.94, 208.29], [574.38, 208.29], [569.54, 178.55], [533.7, 178.55], [542.94, 177.85], [574.38, 177.85]]}, {\"category\": \"car\", \"corners_3d\": [[703.1, 220.53], [661.78, 220.59], [653.4, 214.63], [689.58, 214.59], [703.1, 180.97], [661.78, 180.98], [653.4, 179.96], [689.58, 179.96]]}, {\"category\": \"car\", \"corners_3d\": [[586.03, 201.06], [561.59, 201.1], [562.81, 199.24], [585.64, 199.2], [586.03, 178.8], [561.59, 178.8], [562.81, 178.41], [585.64, 178.4]]}, {\"category\": \"car\", \"corners_3d\": [[640.06, 198.56], [664.1, 198.57], [667.58, 200.49], [641.75, 200.48], [640.06, 175.39], [664.1, 175.39], [667.58, 175.58], [641.75, 175.58]]}]\n```" - } - ] - }, - { - "id": 65, - "data_source": "Mono3DRefer", - "scene_name": "000040", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[510.19, 210.97], [475.01, 211.06], [485.06, 207.14], [516.61, 207.06], [510.19, 179.74], [475.01, 179.76], [485.06, 179.05], [516.61, 179.03]]}]\n```" - } - ] - }, - { - "id": 66, - "data_source": "Mono3DRefer", - "scene_name": "000040", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[702.99, 198.4], [695.78, 198.49], [691.74, 198.18], [698.88, 198.09], [702.99, 168.96], [695.78, 168.94], [691.74, 168.99], [698.88, 169.0]]}]\n```" - } - ] - }, - { - "id": 67, - "data_source": "Mono3DRefer", - "scene_name": "000040", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[510.19, 210.97], [475.01, 211.06], [485.06, 207.14], [516.61, 207.06], [510.19, 179.74], [475.01, 179.76], [485.06, 179.05], [516.61, 179.03]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[702.99, 198.4], [695.78, 198.49], [691.74, 198.18], [698.88, 198.09], [702.99, 168.96], [695.78, 168.94], [691.74, 168.99], [698.88, 169.0]]}]\n```" - } - ] - }, - { - "id": 68, - "data_source": "Mono3DRefer", - "scene_name": "000041", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[439.82, 204.93], [446.31, 203.7], [509.04, 203.69], [505.06, 204.93], [439.82, 178.67], [446.31, 178.44], [509.04, 178.44], [505.06, 178.66]]}]\n```" - } - ] - }, - { - "id": 69, - "data_source": "Mono3DRefer", - "scene_name": "000042", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[415.89, 198.44], [390.17, 198.54], [433.57, 190.52], [451.17, 190.48], [415.89, 165.04], [390.17, 165.01], [433.57, 167.45], [451.17, 167.47]]}]\n```" - } - ] - }, - { - "id": 70, - "data_source": "Mono3DRefer", - "scene_name": "000042", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[578.08, 200.43], [604.1, 200.42], [604.82, 202.76], [576.57, 202.79], [578.08, 174.72], [604.1, 174.72], [604.82, 174.88], [576.57, 174.88]]}]\n```" - } - ] - }, - { - "id": 71, - "data_source": "Mono3DRefer", - "scene_name": "000042", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[578.08, 200.43], [604.1, 200.42], [604.82, 202.76], [576.57, 202.79], [578.08, 174.72], [604.1, 174.72], [604.82, 174.88], [576.57, 174.88]]}, {\"category\": \"tram\", \"corners_3d\": [[415.89, 198.44], [390.17, 198.54], [433.57, 190.52], [451.17, 190.48], [415.89, 165.04], [390.17, 165.01], [433.57, 167.45], [451.17, 167.47]]}]\n```" - } - ] - }, - { - "id": 72, - "data_source": "Mono3DRefer", - "scene_name": "000043", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[466.1, 261.27], [500.3, 259.93], [508.0, 264.81], [471.98, 266.31], [466.1, 175.09], [500.3, 175.05], [508.0, 175.18], [471.98, 175.21]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[439.08, 262.0], [470.29, 261.14], [472.34, 267.39], [438.92, 268.37], [439.08, 170.56], [470.29, 170.58], [472.34, 170.42], [438.92, 170.4]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[712.57, 229.77], [692.19, 230.24], [680.17, 227.8], [699.79, 227.37], [712.57, 162.1], [692.19, 162.01], [680.17, 162.47], [699.79, 162.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[747.37, 228.99], [725.05, 229.61], [712.43, 227.59], [734.09, 227.02], [747.37, 160.24], [725.05, 160.1], [712.43, 160.56], [734.09, 160.69]]}]\n```" - } - ] - }, - { - "id": 73, - "data_source": "Mono3DRefer", - "scene_name": "000045", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[326.66, 242.74], [269.47, 242.58], [327.32, 231.41], [375.45, 231.53], [326.66, 177.89], [269.47, 177.87], [327.32, 177.07], [375.45, 177.08]]}, {\"category\": \"car\", \"corners_3d\": [[386.86, 229.57], [346.85, 229.32], [384.49, 222.68], [419.93, 222.87], [386.86, 188.24], [346.85, 188.17], [384.49, 186.37], [419.93, 186.42]]}, {\"category\": \"car\", \"corners_3d\": [[428.78, 221.93], [387.76, 221.82], [419.38, 215.63], [455.26, 215.71], [428.78, 179.41], [387.76, 179.39], [419.38, 178.56], [455.26, 178.58]]}, {\"category\": \"car\", \"corners_3d\": [[486.23, 209.94], [456.25, 209.86], [473.61, 206.46], [500.86, 206.52], [486.23, 181.98], [456.25, 181.96], [473.61, 181.13], [500.86, 181.14]]}, {\"category\": \"car\", \"corners_3d\": [[505.66, 205.94], [480.22, 205.86], [494.55, 203.22], [517.98, 203.29], [505.66, 182.05], [480.22, 182.03], [494.55, 181.29], [517.98, 181.31]]}, {\"category\": \"car\", \"corners_3d\": [[164.05, 278.24], [75.85, 277.99], [206.04, 253.0], [273.47, 253.14], [164.05, 197.54], [75.85, 197.48], [206.04, 191.63], [273.47, 191.66]]}]\n```" - } - ] - }, - { - "id": 74, - "data_source": "Mono3DRefer", - "scene_name": "000045", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[694.54, 185.11], [713.41, 185.13], [733.95, 189.34], [708.58, 189.3], [694.54, 158.77], [713.41, 158.75], [733.95, 153.92], [708.58, 153.96]]}]\n```" - } - ] - }, - { - "id": 75, - "data_source": "Mono3DRefer", - "scene_name": "000045", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[326.66, 242.74], [269.47, 242.58], [327.32, 231.41], [375.45, 231.53], [326.66, 177.89], [269.47, 177.87], [327.32, 177.07], [375.45, 177.08]]}, {\"category\": \"car\", \"corners_3d\": [[386.86, 229.57], [346.85, 229.32], [384.49, 222.68], [419.93, 222.87], [386.86, 188.24], [346.85, 188.17], [384.49, 186.37], [419.93, 186.42]]}, {\"category\": \"car\", \"corners_3d\": [[428.78, 221.93], [387.76, 221.82], [419.38, 215.63], [455.26, 215.71], [428.78, 179.41], [387.76, 179.39], [419.38, 178.56], [455.26, 178.58]]}, {\"category\": \"car\", \"corners_3d\": [[486.23, 209.94], [456.25, 209.86], [473.61, 206.46], [500.86, 206.52], [486.23, 181.98], [456.25, 181.96], [473.61, 181.13], [500.86, 181.14]]}, {\"category\": \"car\", \"corners_3d\": [[505.66, 205.94], [480.22, 205.86], [494.55, 203.22], [517.98, 203.29], [505.66, 182.05], [480.22, 182.03], [494.55, 181.29], [517.98, 181.31]]}, {\"category\": \"car\", \"corners_3d\": [[164.05, 278.24], [75.85, 277.99], [206.04, 253.0], [273.47, 253.14], [164.05, 197.54], [75.85, 197.48], [206.04, 191.63], [273.47, 191.66]]}, {\"category\": \"tram\", \"corners_3d\": [[694.54, 185.11], [713.41, 185.13], [733.95, 189.34], [708.58, 189.3], [694.54, 158.77], [713.41, 158.75], [733.95, 153.92], [708.58, 153.96]]}]\n```" - } - ] - }, - { - "id": 76, - "data_source": "Mono3DRefer", - "scene_name": "000046", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[582.0, 215.24], [623.13, 215.22], [626.1, 221.46], [578.92, 221.49], [582.0, 176.38], [623.13, 176.38], [626.1, 176.89], [578.92, 176.9]]}, {\"category\": \"car\", \"corners_3d\": [[599.49, 203.83], [632.01, 203.83], [634.86, 207.87], [598.1, 207.86], [599.49, 169.77], [632.01, 169.77], [634.86, 169.37], [598.1, 169.37]]}, {\"category\": \"car\", \"corners_3d\": [[771.47, 292.2], [892.98, 292.81], [1087.12, 381.98], [874.38, 380.11], [771.47, 163.19], [892.98, 163.14], [1087.12, 155.92], [874.38, 156.07]]}, {\"category\": \"car\", \"corners_3d\": [[706.86, 239.28], [767.6, 239.17], [807.51, 254.57], [732.72, 254.73], [706.86, 175.35], [767.6, 175.34], [807.51, 175.92], [732.72, 175.93]]}, {\"category\": \"car\", \"corners_3d\": [[660.69, 206.56], [693.77, 206.56], [703.2, 210.37], [666.39, 210.36], [660.69, 172.63], [693.77, 172.63], [703.2, 172.61], [666.39, 172.61]]}, {\"category\": \"car\", \"corners_3d\": [[646.63, 201.92], [673.97, 201.88], [684.39, 205.14], [653.99, 205.19], [646.63, 174.64], [673.97, 174.64], [684.39, 174.84], [653.99, 174.84]]}]\n```" - } - ] - }, - { - "id": 77, - "data_source": "Mono3DRefer", - "scene_name": "000047", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[500.55, 203.76], [537.35, 203.75], [525.65, 207.19], [482.03, 207.2], [500.55, 157.89], [537.35, 157.9], [525.65, 152.85], [482.03, 152.83]]}]\n```" - } - ] - }, - { - "id": 78, - "data_source": "Mono3DRefer", - "scene_name": "000047", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[576.43, 213.73], [620.88, 213.72], [624.05, 218.16], [572.66, 218.18], [576.43, 173.31], [620.88, 173.32], [624.05, 171.47], [572.66, 171.46]]}, {\"category\": \"car\", \"corners_3d\": [[82.55, 277.03], [171.5, 277.04], [-0.01, 313.14], [-123.91, 313.13], [82.55, 192.08], [171.5, 192.09], [-0.01, 194.79], [-123.91, 194.79]]}, {\"category\": \"car\", \"corners_3d\": [[333.24, 259.47], [423.48, 259.47], [344.61, 291.26], [215.75, 291.24], [333.24, 170.7], [423.48, 170.7], [344.61, 164.48], [215.75, 164.48]]}, {\"category\": \"car\", \"corners_3d\": [[348.2, 223.32], [393.13, 223.3], [351.73, 230.9], [297.79, 230.93], [348.2, 182.96], [393.13, 182.96], [351.73, 182.51], [297.79, 182.51]]}, {\"category\": \"car\", \"corners_3d\": [[453.34, 219.71], [497.75, 219.69], [480.56, 225.46], [428.7, 225.49], [453.34, 181.2], [497.75, 181.2], [480.56, 180.53], [428.7, 180.53]]}, {\"category\": \"car\", \"corners_3d\": [[528.54, 200.69], [552.73, 200.68], [548.63, 202.01], [522.37, 202.02], [528.54, 177.09], [552.73, 177.09], [548.63, 176.4], [522.37, 176.4]]}, {\"category\": \"car\", \"corners_3d\": [[123.72, 238.0], [180.49, 238.0], [64.92, 252.28], [-7.2, 252.27], [123.72, 186.62], [180.49, 186.62], [64.92, 187.0], [-7.2, 187.0]]}, {\"category\": \"car\", \"corners_3d\": [[252.14, 222.58], [294.6, 222.56], [243.97, 228.74], [194.45, 228.76], [252.14, 181.83], [294.6, 181.84], [243.97, 181.28], [194.45, 181.27]]}]\n```" - } - ] - }, - { - "id": 79, - "data_source": "Mono3DRefer", - "scene_name": "000047", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[576.43, 213.73], [620.88, 213.72], [624.05, 218.16], [572.66, 218.18], [576.43, 173.31], [620.88, 173.32], [624.05, 171.47], [572.66, 171.46]]}, {\"category\": \"car\", \"corners_3d\": [[82.55, 277.03], [171.5, 277.04], [-0.01, 313.14], [-123.91, 313.13], [82.55, 192.08], [171.5, 192.09], [-0.01, 194.79], [-123.91, 194.79]]}, {\"category\": \"car\", \"corners_3d\": [[333.24, 259.47], [423.48, 259.47], [344.61, 291.26], [215.75, 291.24], [333.24, 170.7], [423.48, 170.7], [344.61, 164.48], [215.75, 164.48]]}, {\"category\": \"car\", \"corners_3d\": [[348.2, 223.32], [393.13, 223.3], [351.73, 230.9], [297.79, 230.93], [348.2, 182.96], [393.13, 182.96], [351.73, 182.51], [297.79, 182.51]]}, {\"category\": \"car\", \"corners_3d\": [[453.34, 219.71], [497.75, 219.69], [480.56, 225.46], [428.7, 225.49], [453.34, 181.2], [497.75, 181.2], [480.56, 180.53], [428.7, 180.53]]}, {\"category\": \"van\", \"corners_3d\": [[500.55, 203.76], [537.35, 203.75], [525.65, 207.19], [482.03, 207.2], [500.55, 157.89], [537.35, 157.9], [525.65, 152.85], [482.03, 152.83]]}, {\"category\": \"car\", \"corners_3d\": [[528.54, 200.69], [552.73, 200.68], [548.63, 202.01], [522.37, 202.02], [528.54, 177.09], [552.73, 177.09], [548.63, 176.4], [522.37, 176.4]]}, {\"category\": \"car\", \"corners_3d\": [[123.72, 238.0], [180.49, 238.0], [64.92, 252.28], [-7.2, 252.27], [123.72, 186.62], [180.49, 186.62], [64.92, 187.0], [-7.2, 187.0]]}, {\"category\": \"car\", \"corners_3d\": [[252.14, 222.58], [294.6, 222.56], [243.97, 228.74], [194.45, 228.76], [252.14, 181.83], [294.6, 181.84], [243.97, 181.28], [194.45, 181.27]]}]\n```" - } - ] - }, - { - "id": 80, - "data_source": "Mono3DRefer", - "scene_name": "000048", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[616.59, 206.53], [600.97, 206.44], [607.12, 205.11], [621.92, 205.18], [616.59, 162.5], [600.97, 162.57], [607.12, 163.58], [621.92, 163.53]]}]\n```" - } - ] - }, - { - "id": 81, - "data_source": "Mono3DRefer", - "scene_name": "000048", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[650.92, 233.81], [686.05, 234.02], [687.49, 236.77], [650.52, 236.54], [650.92, 161.92], [686.05, 161.84], [687.49, 160.81], [650.52, 160.9]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[672.96, 233.12], [709.98, 233.18], [715.41, 236.14], [676.27, 236.08], [672.96, 158.9], [709.98, 158.87], [715.41, 157.57], [676.27, 157.6]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[444.29, 237.63], [483.97, 237.51], [478.32, 241.09], [436.09, 241.23], [444.29, 161.29], [483.97, 161.33], [478.32, 160.04], [436.09, 159.99]]}]\n```" - } - ] - }, - { - "id": 82, - "data_source": "Mono3DRefer", - "scene_name": "000048", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[650.92, 233.81], [686.05, 234.02], [687.49, 236.77], [650.52, 236.54], [650.92, 161.92], [686.05, 161.84], [687.49, 160.81], [650.52, 160.9]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[672.96, 233.12], [709.98, 233.18], [715.41, 236.14], [676.27, 236.08], [672.96, 158.9], [709.98, 158.87], [715.41, 157.57], [676.27, 157.6]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[444.29, 237.63], [483.97, 237.51], [478.32, 241.09], [436.09, 241.23], [444.29, 161.29], [483.97, 161.33], [478.32, 160.04], [436.09, 159.99]]}, {\"category\": \"cyclist\", \"corners_3d\": [[616.59, 206.53], [600.97, 206.44], [607.12, 205.11], [621.92, 205.18], [616.59, 162.5], [600.97, 162.57], [607.12, 163.58], [621.92, 163.53]]}]\n```" - } - ] - }, - { - "id": 83, - "data_source": "Mono3DRefer", - "scene_name": "000049", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1007.46, 211.56], [1040.84, 214.64], [905.99, 214.88], [882.42, 211.77], [1007.46, 157.27], [1040.84, 156.03], [905.99, 155.94], [882.42, 157.19]]}, {\"category\": \"van\", \"corners_3d\": [[515.55, 192.62], [537.35, 192.09], [585.75, 193.51], [563.73, 194.13], [515.55, 158.48], [537.35, 158.87], [585.75, 157.84], [563.73, 157.39]]}]\n```" - } - ] - }, - { - "id": 84, - "data_source": "Mono3DRefer", - "scene_name": "000049", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[951.86, 406.41], [1072.34, 407.21], [1149.9, 447.8], [1008.32, 446.7], [951.86, 188.38], [1072.34, 188.43], [1149.9, 191.13], [1008.32, 191.05]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[687.25, 208.69], [702.75, 208.54], [710.39, 209.72], [694.42, 209.88], [687.25, 173.5], [702.75, 173.5], [710.39, 173.52], [694.42, 173.52]]}]\n```" - } - ] - }, - { - "id": 85, - "data_source": "Mono3DRefer", - "scene_name": "000049", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[341.58, 229.27], [287.93, 229.35], [333.97, 220.87], [379.52, 220.81], [341.58, 176.82], [287.93, 176.83], [333.97, 176.23], [379.52, 176.23]]}, {\"category\": \"car\", \"corners_3d\": [[3.7, 241.7], [-67.26, 241.78], [53.27, 229.36], [111.35, 229.31], [3.7, 171.46], [-67.26, 171.46], [53.27, 171.71], [111.35, 171.71]]}, {\"category\": \"car\", \"corners_3d\": [[600.26, 219.68], [652.74, 219.75], [657.53, 227.74], [596.13, 227.64], [600.26, 177.5], [652.74, 177.5], [657.53, 178.29], [596.13, 178.29]]}, {\"category\": \"car\", \"corners_3d\": [[305.91, 251.95], [375.04, 245.29], [494.54, 256.49], [426.43, 265.5], [305.91, 180.49], [375.04, 179.85], [494.54, 180.93], [426.43, 181.8]]}, {\"category\": \"car\", \"corners_3d\": [[410.71, 223.45], [452.24, 221.23], [514.84, 226.82], [471.49, 229.61], [410.71, 176.69], [452.24, 176.52], [514.84, 176.94], [471.49, 177.15]]}, {\"category\": \"car\", \"corners_3d\": [[436.15, 214.97], [472.87, 213.02], [538.12, 216.77], [501.26, 219.1], [436.15, 174.27], [472.87, 174.2], [538.12, 174.33], [501.26, 174.41]]}, {\"category\": \"car\", \"corners_3d\": [[474.0, 201.87], [501.02, 200.93], [556.23, 203.38], [528.77, 204.49], [474.0, 172.2], [501.02, 172.22], [556.23, 172.16], [528.77, 172.14]]}, {\"category\": \"car\", \"corners_3d\": [[824.99, 278.08], [926.65, 277.94], [1082.92, 328.67], [932.4, 328.97], [824.99, 188.68], [926.65, 188.66], [1082.92, 196.29], [932.4, 196.34]]}, {\"category\": \"car\", \"corners_3d\": [[753.98, 243.2], [822.21, 243.2], [885.18, 264.09], [796.69, 264.08], [753.98, 184.85], [822.21, 184.85], [885.18, 188.41], [796.69, 188.41]]}, {\"category\": \"car\", \"corners_3d\": [[720.46, 226.88], [778.61, 226.84], [816.55, 238.5], [745.87, 238.56], [720.46, 170.89], [778.61, 170.89], [816.55, 170.47], [745.87, 170.46]]}, {\"category\": \"car\", \"corners_3d\": [[864.03, 209.04], [882.13, 211.09], [790.54, 211.54], [777.2, 209.44], [864.03, 171.75], [882.13, 171.69], [790.54, 171.67], [777.2, 171.74]]}, {\"category\": \"car\", \"corners_3d\": [[81.51, 282.36], [-19.73, 280.35], [177.2, 250.77], [252.88, 251.82], [81.51, 180.76], [-19.73, 180.61], [177.2, 178.48], [252.88, 178.55]]}, {\"category\": \"car\", \"corners_3d\": [[259.69, 242.7], [194.82, 241.45], [286.76, 229.89], [341.91, 230.75], [259.69, 175.63], [194.82, 175.58], [286.76, 175.12], [341.91, 175.16]]}, {\"category\": \"car\", \"corners_3d\": [[606.3, 201.51], [638.46, 201.55], [639.17, 204.61], [603.59, 204.57], [606.3, 172.27], [638.46, 172.27], [639.17, 172.21], [603.59, 172.21]]}]\n```" - } - ] - }, - { - "id": 86, - "data_source": "Mono3DRefer", - "scene_name": "000049", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1007.46, 211.56], [1040.84, 214.64], [905.99, 214.88], [882.42, 211.77], [1007.46, 157.27], [1040.84, 156.03], [905.99, 155.94], [882.42, 157.19]]}, {\"category\": \"van\", \"corners_3d\": [[515.55, 192.62], [537.35, 192.09], [585.75, 193.51], [563.73, 194.13], [515.55, 158.48], [537.35, 158.87], [585.75, 157.84], [563.73, 157.39]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[951.86, 406.41], [1072.34, 407.21], [1149.9, 447.8], [1008.32, 446.7], [951.86, 188.38], [1072.34, 188.43], [1149.9, 191.13], [1008.32, 191.05]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[687.25, 208.69], [702.75, 208.54], [710.39, 209.72], [694.42, 209.88], [687.25, 173.5], [702.75, 173.5], [710.39, 173.52], [694.42, 173.52]]}]\n```" - } - ] - }, - { - "id": 87, - "data_source": "Mono3DRefer", - "scene_name": "000049", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[341.58, 229.27], [287.93, 229.35], [333.97, 220.87], [379.52, 220.81], [341.58, 176.82], [287.93, 176.83], [333.97, 176.23], [379.52, 176.23]]}, {\"category\": \"van\", \"corners_3d\": [[1007.46, 211.56], [1040.84, 214.64], [905.99, 214.88], [882.42, 211.77], [1007.46, 157.27], [1040.84, 156.03], [905.99, 155.94], [882.42, 157.19]]}, {\"category\": \"car\", \"corners_3d\": [[3.7, 241.7], [-67.26, 241.78], [53.27, 229.36], [111.35, 229.31], [3.7, 171.46], [-67.26, 171.46], [53.27, 171.71], [111.35, 171.71]]}, {\"category\": \"car\", \"corners_3d\": [[600.26, 219.68], [652.74, 219.75], [657.53, 227.74], [596.13, 227.64], [600.26, 177.5], [652.74, 177.5], [657.53, 178.29], [596.13, 178.29]]}, {\"category\": \"car\", \"corners_3d\": [[305.91, 251.95], [375.04, 245.29], [494.54, 256.49], [426.43, 265.5], [305.91, 180.49], [375.04, 179.85], [494.54, 180.93], [426.43, 181.8]]}, {\"category\": \"car\", \"corners_3d\": [[410.71, 223.45], [452.24, 221.23], [514.84, 226.82], [471.49, 229.61], [410.71, 176.69], [452.24, 176.52], [514.84, 176.94], [471.49, 177.15]]}, {\"category\": \"car\", \"corners_3d\": [[436.15, 214.97], [472.87, 213.02], [538.12, 216.77], [501.26, 219.1], [436.15, 174.27], [472.87, 174.2], [538.12, 174.33], [501.26, 174.41]]}, {\"category\": \"car\", \"corners_3d\": [[474.0, 201.87], [501.02, 200.93], [556.23, 203.38], [528.77, 204.49], [474.0, 172.2], [501.02, 172.22], [556.23, 172.16], [528.77, 172.14]]}, {\"category\": \"van\", \"corners_3d\": [[515.55, 192.62], [537.35, 192.09], [585.75, 193.51], [563.73, 194.13], [515.55, 158.48], [537.35, 158.87], [585.75, 157.84], [563.73, 157.39]]}, {\"category\": \"car\", \"corners_3d\": [[824.99, 278.08], [926.65, 277.94], [1082.92, 328.67], [932.4, 328.97], [824.99, 188.68], [926.65, 188.66], [1082.92, 196.29], [932.4, 196.34]]}, {\"category\": \"car\", \"corners_3d\": [[753.98, 243.2], [822.21, 243.2], [885.18, 264.09], [796.69, 264.08], [753.98, 184.85], [822.21, 184.85], [885.18, 188.41], [796.69, 188.41]]}, {\"category\": \"car\", \"corners_3d\": [[720.46, 226.88], [778.61, 226.84], [816.55, 238.5], [745.87, 238.56], [720.46, 170.89], [778.61, 170.89], [816.55, 170.47], [745.87, 170.46]]}, {\"category\": \"car\", \"corners_3d\": [[864.03, 209.04], [882.13, 211.09], [790.54, 211.54], [777.2, 209.44], [864.03, 171.75], [882.13, 171.69], [790.54, 171.67], [777.2, 171.74]]}, {\"category\": \"car\", \"corners_3d\": [[81.51, 282.36], [-19.73, 280.35], [177.2, 250.77], [252.88, 251.82], [81.51, 180.76], [-19.73, 180.61], [177.2, 178.48], [252.88, 178.55]]}, {\"category\": \"car\", \"corners_3d\": [[259.69, 242.7], [194.82, 241.45], [286.76, 229.89], [341.91, 230.75], [259.69, 175.63], [194.82, 175.58], [286.76, 175.12], [341.91, 175.16]]}, {\"category\": \"car\", \"corners_3d\": [[606.3, 201.51], [638.46, 201.55], [639.17, 204.61], [603.59, 204.57], [606.3, 172.27], [638.46, 172.27], [639.17, 172.21], [603.59, 172.21]]}]\n```" - } - ] - }, - { - "id": 88, - "data_source": "Mono3DRefer", - "scene_name": "000049", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[341.58, 229.27], [287.93, 229.35], [333.97, 220.87], [379.52, 220.81], [341.58, 176.82], [287.93, 176.83], [333.97, 176.23], [379.52, 176.23]]}, {\"category\": \"car\", \"corners_3d\": [[3.7, 241.7], [-67.26, 241.78], [53.27, 229.36], [111.35, 229.31], [3.7, 171.46], [-67.26, 171.46], [53.27, 171.71], [111.35, 171.71]]}, {\"category\": \"car\", \"corners_3d\": [[600.26, 219.68], [652.74, 219.75], [657.53, 227.74], [596.13, 227.64], [600.26, 177.5], [652.74, 177.5], [657.53, 178.29], [596.13, 178.29]]}, {\"category\": \"car\", \"corners_3d\": [[305.91, 251.95], [375.04, 245.29], [494.54, 256.49], [426.43, 265.5], [305.91, 180.49], [375.04, 179.85], [494.54, 180.93], [426.43, 181.8]]}, {\"category\": \"car\", \"corners_3d\": [[410.71, 223.45], [452.24, 221.23], [514.84, 226.82], [471.49, 229.61], [410.71, 176.69], [452.24, 176.52], [514.84, 176.94], [471.49, 177.15]]}, {\"category\": \"car\", \"corners_3d\": [[436.15, 214.97], [472.87, 213.02], [538.12, 216.77], [501.26, 219.1], [436.15, 174.27], [472.87, 174.2], [538.12, 174.33], [501.26, 174.41]]}, {\"category\": \"car\", \"corners_3d\": [[474.0, 201.87], [501.02, 200.93], [556.23, 203.38], [528.77, 204.49], [474.0, 172.2], [501.02, 172.22], [556.23, 172.16], [528.77, 172.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[951.86, 406.41], [1072.34, 407.21], [1149.9, 447.8], [1008.32, 446.7], [951.86, 188.38], [1072.34, 188.43], [1149.9, 191.13], [1008.32, 191.05]]}, {\"category\": \"car\", \"corners_3d\": [[824.99, 278.08], [926.65, 277.94], [1082.92, 328.67], [932.4, 328.97], [824.99, 188.68], [926.65, 188.66], [1082.92, 196.29], [932.4, 196.34]]}, {\"category\": \"car\", \"corners_3d\": [[753.98, 243.2], [822.21, 243.2], [885.18, 264.09], [796.69, 264.08], [753.98, 184.85], [822.21, 184.85], [885.18, 188.41], [796.69, 188.41]]}, {\"category\": \"car\", \"corners_3d\": [[720.46, 226.88], [778.61, 226.84], [816.55, 238.5], [745.87, 238.56], [720.46, 170.89], [778.61, 170.89], [816.55, 170.47], [745.87, 170.46]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[687.25, 208.69], [702.75, 208.54], [710.39, 209.72], [694.42, 209.88], [687.25, 173.5], [702.75, 173.5], [710.39, 173.52], [694.42, 173.52]]}, {\"category\": \"car\", \"corners_3d\": [[864.03, 209.04], [882.13, 211.09], [790.54, 211.54], [777.2, 209.44], [864.03, 171.75], [882.13, 171.69], [790.54, 171.67], [777.2, 171.74]]}, {\"category\": \"car\", \"corners_3d\": [[81.51, 282.36], [-19.73, 280.35], [177.2, 250.77], [252.88, 251.82], [81.51, 180.76], [-19.73, 180.61], [177.2, 178.48], [252.88, 178.55]]}, {\"category\": \"car\", \"corners_3d\": [[259.69, 242.7], [194.82, 241.45], [286.76, 229.89], [341.91, 230.75], [259.69, 175.63], [194.82, 175.58], [286.76, 175.12], [341.91, 175.16]]}, {\"category\": \"car\", \"corners_3d\": [[606.3, 201.51], [638.46, 201.55], [639.17, 204.61], [603.59, 204.57], [606.3, 172.27], [638.46, 172.27], [639.17, 172.21], [603.59, 172.21]]}]\n```" - } - ] - }, - { - "id": 89, - "data_source": "Mono3DRefer", - "scene_name": "000049", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[341.58, 229.27], [287.93, 229.35], [333.97, 220.87], [379.52, 220.81], [341.58, 176.82], [287.93, 176.83], [333.97, 176.23], [379.52, 176.23]]}, {\"category\": \"van\", \"corners_3d\": [[1007.46, 211.56], [1040.84, 214.64], [905.99, 214.88], [882.42, 211.77], [1007.46, 157.27], [1040.84, 156.03], [905.99, 155.94], [882.42, 157.19]]}, {\"category\": \"car\", \"corners_3d\": [[3.7, 241.7], [-67.26, 241.78], [53.27, 229.36], [111.35, 229.31], [3.7, 171.46], [-67.26, 171.46], [53.27, 171.71], [111.35, 171.71]]}, {\"category\": \"car\", \"corners_3d\": [[600.26, 219.68], [652.74, 219.75], [657.53, 227.74], [596.13, 227.64], [600.26, 177.5], [652.74, 177.5], [657.53, 178.29], [596.13, 178.29]]}, {\"category\": \"car\", \"corners_3d\": [[305.91, 251.95], [375.04, 245.29], [494.54, 256.49], [426.43, 265.5], [305.91, 180.49], [375.04, 179.85], [494.54, 180.93], [426.43, 181.8]]}, {\"category\": \"car\", \"corners_3d\": [[410.71, 223.45], [452.24, 221.23], [514.84, 226.82], [471.49, 229.61], [410.71, 176.69], [452.24, 176.52], [514.84, 176.94], [471.49, 177.15]]}, {\"category\": \"car\", \"corners_3d\": [[436.15, 214.97], [472.87, 213.02], [538.12, 216.77], [501.26, 219.1], [436.15, 174.27], [472.87, 174.2], [538.12, 174.33], [501.26, 174.41]]}, {\"category\": \"car\", \"corners_3d\": [[474.0, 201.87], [501.02, 200.93], [556.23, 203.38], [528.77, 204.49], [474.0, 172.2], [501.02, 172.22], [556.23, 172.16], [528.77, 172.14]]}, {\"category\": \"van\", \"corners_3d\": [[515.55, 192.62], [537.35, 192.09], [585.75, 193.51], [563.73, 194.13], [515.55, 158.48], [537.35, 158.87], [585.75, 157.84], [563.73, 157.39]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[951.86, 406.41], [1072.34, 407.21], [1149.9, 447.8], [1008.32, 446.7], [951.86, 188.38], [1072.34, 188.43], [1149.9, 191.13], [1008.32, 191.05]]}, {\"category\": \"car\", \"corners_3d\": [[824.99, 278.08], [926.65, 277.94], [1082.92, 328.67], [932.4, 328.97], [824.99, 188.68], [926.65, 188.66], [1082.92, 196.29], [932.4, 196.34]]}, {\"category\": \"car\", \"corners_3d\": [[753.98, 243.2], [822.21, 243.2], [885.18, 264.09], [796.69, 264.08], [753.98, 184.85], [822.21, 184.85], [885.18, 188.41], [796.69, 188.41]]}, {\"category\": \"car\", \"corners_3d\": [[720.46, 226.88], [778.61, 226.84], [816.55, 238.5], [745.87, 238.56], [720.46, 170.89], [778.61, 170.89], [816.55, 170.47], [745.87, 170.46]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[687.25, 208.69], [702.75, 208.54], [710.39, 209.72], [694.42, 209.88], [687.25, 173.5], [702.75, 173.5], [710.39, 173.52], [694.42, 173.52]]}, {\"category\": \"car\", \"corners_3d\": [[864.03, 209.04], [882.13, 211.09], [790.54, 211.54], [777.2, 209.44], [864.03, 171.75], [882.13, 171.69], [790.54, 171.67], [777.2, 171.74]]}, {\"category\": \"car\", \"corners_3d\": [[81.51, 282.36], [-19.73, 280.35], [177.2, 250.77], [252.88, 251.82], [81.51, 180.76], [-19.73, 180.61], [177.2, 178.48], [252.88, 178.55]]}, {\"category\": \"car\", \"corners_3d\": [[259.69, 242.7], [194.82, 241.45], [286.76, 229.89], [341.91, 230.75], [259.69, 175.63], [194.82, 175.58], [286.76, 175.12], [341.91, 175.16]]}, {\"category\": \"car\", \"corners_3d\": [[606.3, 201.51], [638.46, 201.55], [639.17, 204.61], [603.59, 204.57], [606.3, 172.27], [638.46, 172.27], [639.17, 172.21], [603.59, 172.21]]}]\n```" - } - ] - }, - { - "id": 90, - "data_source": "Mono3DRefer", - "scene_name": "000051", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[714.1, 201.93], [741.71, 202.09], [744.56, 204.87], [714.31, 204.69], [714.1, 175.1], [741.71, 175.11], [744.56, 175.33], [714.31, 175.31]]}]\n```" - } - ] - }, - { - "id": 91, - "data_source": "Mono3DRefer", - "scene_name": "000051", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[920.09, 225.86], [940.55, 225.98], [967.99, 231.33], [945.42, 231.19], [920.09, 157.8], [940.55, 157.77], [967.99, 156.25], [945.42, 156.29]]}]\n```" - } - ] - }, - { - "id": 92, - "data_source": "Mono3DRefer", - "scene_name": "000051", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[920.09, 225.86], [940.55, 225.98], [967.99, 231.33], [945.42, 231.19], [920.09, 157.8], [940.55, 157.77], [967.99, 156.25], [945.42, 156.29]]}, {\"category\": \"car\", \"corners_3d\": [[714.1, 201.93], [741.71, 202.09], [744.56, 204.87], [714.31, 204.69], [714.1, 175.1], [741.71, 175.11], [744.56, 175.33], [714.31, 175.31]]}]\n```" - } - ] - }, - { - "id": 93, - "data_source": "Mono3DRefer", - "scene_name": "000055", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[581.09, 204.04], [633.8, 204.04], [642.7, 215.77], [570.17, 215.76], [581.09, 139.83], [633.8, 139.83], [642.7, 127.41], [570.17, 127.42]]}]\n```" - } - ] - }, - { - "id": 94, - "data_source": "Mono3DRefer", - "scene_name": "000055", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[62.72, 231.28], [31.7, 231.15], [102.27, 224.45], [129.85, 224.55], [62.72, 195.84], [31.7, 195.79], [102.27, 193.15], [129.85, 193.19]]}, {\"category\": \"car\", \"corners_3d\": [[137.95, 220.84], [104.36, 220.82], [158.36, 215.76], [188.43, 215.78], [137.95, 189.54], [104.36, 189.53], [158.36, 187.77], [188.43, 187.78]]}, {\"category\": \"car\", \"corners_3d\": [[211.16, 217.76], [183.44, 217.74], [220.6, 213.89], [245.95, 213.9], [211.16, 188.47], [183.44, 188.46], [220.6, 187.12], [245.95, 187.13]]}, {\"category\": \"car\", \"corners_3d\": [[248.82, 212.25], [227.44, 212.23], [262.68, 208.73], [282.18, 208.75], [248.82, 190.49], [227.44, 190.48], [262.68, 188.91], [282.18, 188.92]]}]\n```" - } - ] - }, - { - "id": 95, - "data_source": "Mono3DRefer", - "scene_name": "000055", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[581.09, 204.04], [633.8, 204.04], [642.7, 215.77], [570.17, 215.76], [581.09, 139.83], [633.8, 139.83], [642.7, 127.41], [570.17, 127.42]]}, {\"category\": \"car\", \"corners_3d\": [[62.72, 231.28], [31.7, 231.15], [102.27, 224.45], [129.85, 224.55], [62.72, 195.84], [31.7, 195.79], [102.27, 193.15], [129.85, 193.19]]}, {\"category\": \"car\", \"corners_3d\": [[137.95, 220.84], [104.36, 220.82], [158.36, 215.76], [188.43, 215.78], [137.95, 189.54], [104.36, 189.53], [158.36, 187.77], [188.43, 187.78]]}, {\"category\": \"car\", \"corners_3d\": [[211.16, 217.76], [183.44, 217.74], [220.6, 213.89], [245.95, 213.9], [211.16, 188.47], [183.44, 188.46], [220.6, 187.12], [245.95, 187.13]]}, {\"category\": \"car\", \"corners_3d\": [[248.82, 212.25], [227.44, 212.23], [262.68, 208.73], [282.18, 208.75], [248.82, 190.49], [227.44, 190.48], [262.68, 188.91], [282.18, 188.92]]}]\n```" - } - ] - }, - { - "id": 96, - "data_source": "Mono3DRefer", - "scene_name": "000057", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[556.93, 232.34], [514.9, 232.35], [526.76, 224.85], [563.49, 224.85], [556.93, 196.39], [514.9, 196.39], [526.76, 193.43], [563.49, 193.42]]}, {\"category\": \"car\", \"corners_3d\": [[591.09, 196.42], [572.78, 196.44], [574.04, 195.07], [591.29, 195.06], [591.09, 177.15], [572.78, 177.16], [574.04, 176.91], [591.29, 176.91]]}]\n```" - } - ] - }, - { - "id": 97, - "data_source": "Mono3DRefer", - "scene_name": "000059", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[567.41, 193.81], [588.11, 193.77], [590.96, 195.85], [568.19, 195.9], [567.41, 165.27], [588.11, 165.29], [590.96, 164.53], [568.19, 164.52]]}]\n```" - } - ] - }, - { - "id": 98, - "data_source": "Mono3DRefer", - "scene_name": "000059", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[560.96, 209.2], [597.3, 209.12], [599.32, 213.69], [558.4, 213.78], [560.96, 180.57], [597.3, 180.55], [599.32, 181.52], [558.4, 181.54]]}, {\"category\": \"car\", \"corners_3d\": [[516.46, 203.06], [543.55, 203.03], [538.89, 205.73], [509.37, 205.76], [516.46, 180.32], [543.55, 180.31], [538.89, 180.98], [509.37, 180.98]]}, {\"category\": \"car\", \"corners_3d\": [[1004.04, 246.44], [970.25, 241.98], [1062.13, 239.74], [1100.49, 243.91], [1004.04, 197.61], [970.25, 196.11], [1062.13, 195.36], [1100.49, 196.76]]}, {\"category\": \"car\", \"corners_3d\": [[1189.92, 249.45], [1130.27, 244.25], [1205.32, 239.88], [1266.18, 244.44], [1189.92, 197.17], [1130.27, 195.52], [1205.32, 194.13], [1266.18, 195.58]]}]\n```" - } - ] - }, - { - "id": 99, - "data_source": "Mono3DRefer", - "scene_name": "000059", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[560.96, 209.2], [597.3, 209.12], [599.32, 213.69], [558.4, 213.78], [560.96, 180.57], [597.3, 180.55], [599.32, 181.52], [558.4, 181.54]]}, {\"category\": \"car\", \"corners_3d\": [[516.46, 203.06], [543.55, 203.03], [538.89, 205.73], [509.37, 205.76], [516.46, 180.32], [543.55, 180.31], [538.89, 180.98], [509.37, 180.98]]}, {\"category\": \"car\", \"corners_3d\": [[1004.04, 246.44], [970.25, 241.98], [1062.13, 239.74], [1100.49, 243.91], [1004.04, 197.61], [970.25, 196.11], [1062.13, 195.36], [1100.49, 196.76]]}, {\"category\": \"car\", \"corners_3d\": [[1189.92, 249.45], [1130.27, 244.25], [1205.32, 239.88], [1266.18, 244.44], [1189.92, 197.17], [1130.27, 195.52], [1205.32, 194.13], [1266.18, 195.58]]}, {\"category\": \"van\", \"corners_3d\": [[567.41, 193.81], [588.11, 193.77], [590.96, 195.85], [568.19, 195.9], [567.41, 165.27], [588.11, 165.29], [590.96, 164.53], [568.19, 164.52]]}]\n```" - } - ] - }, - { - "id": 100, - "data_source": "Mono3DRefer", - "scene_name": "000060", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[561.31, 224.64], [611.86, 224.6], [614.0, 237.0], [551.33, 237.05], [561.31, 174.39], [611.86, 174.39], [614.0, 174.76], [551.33, 174.76]]}, {\"category\": \"car\", \"corners_3d\": [[321.9, 196.46], [305.6, 196.46], [326.42, 194.88], [341.63, 194.88], [321.9, 178.22], [305.6, 178.22], [326.42, 177.86], [341.63, 177.86]]}]\n```" - } - ] - }, - { - "id": 101, - "data_source": "Mono3DRefer", - "scene_name": "000060", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[790.23, 210.12], [796.71, 210.11], [805.17, 211.74], [798.41, 211.75], [790.23, 178.77], [796.71, 178.77], [805.17, 179.03], [798.41, 179.03]]}]\n```" - } - ] - }, - { - "id": 102, - "data_source": "Mono3DRefer", - "scene_name": "000060", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[545.55, 195.09], [520.05, 195.16], [522.38, 192.98], [545.38, 192.92], [545.55, 162.7], [520.05, 162.67], [522.38, 163.66], [545.38, 163.69]]}]\n```" - } - ] - }, - { - "id": 103, - "data_source": "Mono3DRefer", - "scene_name": "000060", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[790.23, 210.12], [796.71, 210.11], [805.17, 211.74], [798.41, 211.75], [790.23, 178.77], [796.71, 178.77], [805.17, 179.03], [798.41, 179.03]]}, {\"category\": \"car\", \"corners_3d\": [[561.31, 224.64], [611.86, 224.6], [614.0, 237.0], [551.33, 237.05], [561.31, 174.39], [611.86, 174.39], [614.0, 174.76], [551.33, 174.76]]}, {\"category\": \"car\", \"corners_3d\": [[321.9, 196.46], [305.6, 196.46], [326.42, 194.88], [341.63, 194.88], [321.9, 178.22], [305.6, 178.22], [326.42, 177.86], [341.63, 177.86]]}]\n```" - } - ] - }, - { - "id": 104, - "data_source": "Mono3DRefer", - "scene_name": "000060", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[545.55, 195.09], [520.05, 195.16], [522.38, 192.98], [545.38, 192.92], [545.55, 162.7], [520.05, 162.67], [522.38, 163.66], [545.38, 163.69]]}, {\"category\": \"car\", \"corners_3d\": [[561.31, 224.64], [611.86, 224.6], [614.0, 237.0], [551.33, 237.05], [561.31, 174.39], [611.86, 174.39], [614.0, 174.76], [551.33, 174.76]]}, {\"category\": \"car\", \"corners_3d\": [[321.9, 196.46], [305.6, 196.46], [326.42, 194.88], [341.63, 194.88], [321.9, 178.22], [305.6, 178.22], [326.42, 177.86], [341.63, 177.86]]}]\n```" - } - ] - }, - { - "id": 105, - "data_source": "Mono3DRefer", - "scene_name": "000060", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[790.23, 210.12], [796.71, 210.11], [805.17, 211.74], [798.41, 211.75], [790.23, 178.77], [796.71, 178.77], [805.17, 179.03], [798.41, 179.03]]}, {\"category\": \"van\", \"corners_3d\": [[545.55, 195.09], [520.05, 195.16], [522.38, 192.98], [545.38, 192.92], [545.55, 162.7], [520.05, 162.67], [522.38, 163.66], [545.38, 163.69]]}]\n```" - } - ] - }, - { - "id": 106, - "data_source": "Mono3DRefer", - "scene_name": "000060", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[790.23, 210.12], [796.71, 210.11], [805.17, 211.74], [798.41, 211.75], [790.23, 178.77], [796.71, 178.77], [805.17, 179.03], [798.41, 179.03]]}, {\"category\": \"van\", \"corners_3d\": [[545.55, 195.09], [520.05, 195.16], [522.38, 192.98], [545.38, 192.92], [545.55, 162.7], [520.05, 162.67], [522.38, 163.66], [545.38, 163.69]]}, {\"category\": \"car\", \"corners_3d\": [[561.31, 224.64], [611.86, 224.6], [614.0, 237.0], [551.33, 237.05], [561.31, 174.39], [611.86, 174.39], [614.0, 174.76], [551.33, 174.76]]}, {\"category\": \"car\", \"corners_3d\": [[321.9, 196.46], [305.6, 196.46], [326.42, 194.88], [341.63, 194.88], [321.9, 178.22], [305.6, 178.22], [326.42, 177.86], [341.63, 177.86]]}]\n```" - } - ] - }, - { - "id": 107, - "data_source": "Mono3DRefer", - "scene_name": "000061", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[28.84, 327.39], [-134.22, 343.46], [-79.8, 300.64], [41.36, 291.41], [28.84, 222.4], [-134.22, 227.56], [-79.8, 213.83], [41.36, 210.87]]}, {\"category\": \"car\", \"corners_3d\": [[112.25, 252.42], [69.72, 254.63], [72.68, 247.67], [111.6, 245.81], [112.25, 221.26], [69.72, 222.61], [72.68, 218.37], [111.6, 217.24]]}, {\"category\": \"car\", \"corners_3d\": [[81.87, 293.27], [-13.98, 300.15], [13.44, 279.67], [93.31, 274.78], [81.87, 225.9], [-13.98, 228.93], [13.44, 219.91], [93.31, 217.76]]}]\n```" - } - ] - }, - { - "id": 108, - "data_source": "Mono3DRefer", - "scene_name": "000061", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1056.23, 226.62], [1163.98, 229.49], [1251.34, 251.69], [1098.22, 246.24], [1056.23, 127.83], [1163.98, 125.42], [1251.34, 106.82], [1098.22, 111.4]]}, {\"category\": \"van\", \"corners_3d\": [[888.27, 169.31], [923.52, 169.25], [923.48, 168.86], [884.54, 168.94], [888.27, 129.41], [923.52, 128.62], [923.48, 123.9], [884.54, 124.86]]}]\n```" - } - ] - }, - { - "id": 109, - "data_source": "Mono3DRefer", - "scene_name": "000061", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1056.23, 226.62], [1163.98, 229.49], [1251.34, 251.69], [1098.22, 246.24], [1056.23, 127.83], [1163.98, 125.42], [1251.34, 106.82], [1098.22, 111.4]]}, {\"category\": \"car\", \"corners_3d\": [[28.84, 327.39], [-134.22, 343.46], [-79.8, 300.64], [41.36, 291.41], [28.84, 222.4], [-134.22, 227.56], [-79.8, 213.83], [41.36, 210.87]]}, {\"category\": \"car\", \"corners_3d\": [[112.25, 252.42], [69.72, 254.63], [72.68, 247.67], [111.6, 245.81], [112.25, 221.26], [69.72, 222.61], [72.68, 218.37], [111.6, 217.24]]}, {\"category\": \"car\", \"corners_3d\": [[81.87, 293.27], [-13.98, 300.15], [13.44, 279.67], [93.31, 274.78], [81.87, 225.9], [-13.98, 228.93], [13.44, 219.91], [93.31, 217.76]]}, {\"category\": \"van\", \"corners_3d\": [[888.27, 169.31], [923.52, 169.25], [923.48, 168.86], [884.54, 168.94], [888.27, 129.41], [923.52, 128.62], [923.48, 123.9], [884.54, 124.86]]}]\n```" - } - ] - }, - { - "id": 110, - "data_source": "Mono3DRefer", - "scene_name": "000064", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[818.85, 230.88], [908.31, 230.67], [1112.2, 267.52], [966.59, 268.09], [818.85, 105.25], [908.31, 105.5], [1112.2, 62.55], [966.59, 61.89]]}]\n```" - } - ] - }, - { - "id": 111, - "data_source": "Mono3DRefer", - "scene_name": "000064", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[715.56, 201.12], [742.47, 201.04], [762.62, 204.02], [732.91, 204.13], [715.56, 175.83], [742.47, 175.82], [762.62, 176.14], [732.91, 176.15]]}, {\"category\": \"car\", \"corners_3d\": [[658.15, 214.68], [695.78, 214.61], [709.92, 220.12], [667.35, 220.2], [658.15, 179.69], [695.78, 179.68], [709.92, 180.58], [667.35, 180.59]]}, {\"category\": \"car\", \"corners_3d\": [[732.61, 250.23], [807.95, 249.99], [881.88, 275.98], [781.36, 276.4], [732.61, 179.87], [807.95, 179.85], [881.88, 182.21], [781.36, 182.25]]}]\n```" - } - ] - }, - { - "id": 112, - "data_source": "Mono3DRefer", - "scene_name": "000064", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[715.56, 201.12], [742.47, 201.04], [762.62, 204.02], [732.91, 204.13], [715.56, 175.83], [742.47, 175.82], [762.62, 176.14], [732.91, 176.15]]}, {\"category\": \"truck\", \"corners_3d\": [[818.85, 230.88], [908.31, 230.67], [1112.2, 267.52], [966.59, 268.09], [818.85, 105.25], [908.31, 105.5], [1112.2, 62.55], [966.59, 61.89]]}, {\"category\": \"car\", \"corners_3d\": [[658.15, 214.68], [695.78, 214.61], [709.92, 220.12], [667.35, 220.2], [658.15, 179.69], [695.78, 179.68], [709.92, 180.58], [667.35, 180.59]]}, {\"category\": \"car\", \"corners_3d\": [[732.61, 250.23], [807.95, 249.99], [881.88, 275.98], [781.36, 276.4], [732.61, 179.87], [807.95, 179.85], [881.88, 182.21], [781.36, 182.25]]}]\n```" - } - ] - }, - { - "id": 113, - "data_source": "Mono3DRefer", - "scene_name": "000066", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[675.72, 203.9], [710.16, 204.12], [709.39, 206.81], [672.02, 206.54], [675.72, 169.57], [710.16, 169.54], [709.39, 169.26], [672.02, 169.29]]}, {\"category\": \"car\", \"corners_3d\": [[622.64, 200.01], [592.02, 199.8], [606.16, 197.23], [633.94, 197.4], [622.64, 173.41], [592.02, 173.4], [606.16, 173.35], [633.94, 173.35]]}]\n```" - } - ] - }, - { - "id": 114, - "data_source": "Mono3DRefer", - "scene_name": "000068", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[580.76, 212.79], [547.94, 212.79], [555.13, 208.09], [584.08, 208.09], [580.76, 171.85], [547.94, 171.85], [555.13, 171.97], [584.08, 171.97]]}, {\"category\": \"van\", \"corners_3d\": [[594.92, 200.21], [572.78, 200.21], [576.15, 197.66], [596.23, 197.66], [594.92, 175.52], [572.78, 175.52], [576.15, 175.27], [596.23, 175.27]]}, {\"category\": \"van\", \"corners_3d\": [[489.49, 289.62], [381.9, 289.63], [446.14, 256.6], [523.29, 256.59], [489.49, 169.76], [381.9, 169.76], [446.14, 170.63], [523.29, 170.63]]}]\n```" - } - ] - }, - { - "id": 115, - "data_source": "Mono3DRefer", - "scene_name": "000068", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[306.99, 432.15], [68.67, 433.91], [305.06, 316.56], [435.77, 316.03], [306.99, 220.61], [68.67, 220.94], [305.06, 199.32], [435.77, 199.23]]}, {\"category\": \"car\", \"corners_3d\": [[-144.78, 478.15], [-409.54, 479.35], [56.38, 338.15], [198.43, 337.8], [-144.78, 237.17], [-409.54, 237.42], [56.38, 207.67], [198.43, 207.6]]}, {\"category\": \"car\", \"corners_3d\": [[540.05, 249.99], [479.03, 250.13], [499.12, 236.69], [549.52, 236.6], [540.05, 190.32], [479.03, 190.35], [499.12, 187.31], [549.52, 187.29]]}, {\"category\": \"car\", \"corners_3d\": [[567.78, 226.37], [530.35, 226.35], [539.82, 220.44], [573.13, 220.46], [567.78, 189.39], [530.35, 189.38], [539.82, 187.56], [573.13, 187.56]]}, {\"category\": \"car\", \"corners_3d\": [[779.85, 226.47], [791.03, 229.99], [672.15, 229.99], [668.3, 226.47], [779.85, 182.81], [791.03, 183.47], [672.15, 183.47], [668.3, 182.81]]}, {\"category\": \"car\", \"corners_3d\": [[756.14, 219.06], [763.19, 221.52], [659.45, 221.38], [657.66, 218.93], [756.14, 184.95], [763.19, 185.59], [659.45, 185.56], [657.66, 184.92]]}, {\"category\": \"car\", \"corners_3d\": [[712.41, 210.61], [714.44, 212.32], [637.74, 211.99], [639.01, 210.31], [712.41, 180.51], [714.44, 180.85], [637.74, 180.79], [639.01, 180.45]]}]\n```" - } - ] - }, - { - "id": 116, - "data_source": "Mono3DRefer", - "scene_name": "000068", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1036.9, 254.94], [1009.54, 254.94], [960.27, 244.84], [984.26, 244.84], [1036.9, 168.19], [1009.54, 168.19], [960.27, 168.76], [984.26, 168.76]]}]\n```" - } - ] - }, - { - "id": 117, - "data_source": "Mono3DRefer", - "scene_name": "000068", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[306.99, 432.15], [68.67, 433.91], [305.06, 316.56], [435.77, 316.03], [306.99, 220.61], [68.67, 220.94], [305.06, 199.32], [435.77, 199.23]]}, {\"category\": \"car\", \"corners_3d\": [[-144.78, 478.15], [-409.54, 479.35], [56.38, 338.15], [198.43, 337.8], [-144.78, 237.17], [-409.54, 237.42], [56.38, 207.67], [198.43, 207.6]]}, {\"category\": \"car\", \"corners_3d\": [[540.05, 249.99], [479.03, 250.13], [499.12, 236.69], [549.52, 236.6], [540.05, 190.32], [479.03, 190.35], [499.12, 187.31], [549.52, 187.29]]}, {\"category\": \"car\", \"corners_3d\": [[567.78, 226.37], [530.35, 226.35], [539.82, 220.44], [573.13, 220.46], [567.78, 189.39], [530.35, 189.38], [539.82, 187.56], [573.13, 187.56]]}, {\"category\": \"van\", \"corners_3d\": [[580.76, 212.79], [547.94, 212.79], [555.13, 208.09], [584.08, 208.09], [580.76, 171.85], [547.94, 171.85], [555.13, 171.97], [584.08, 171.97]]}, {\"category\": \"car\", \"corners_3d\": [[779.85, 226.47], [791.03, 229.99], [672.15, 229.99], [668.3, 226.47], [779.85, 182.81], [791.03, 183.47], [672.15, 183.47], [668.3, 182.81]]}, {\"category\": \"car\", \"corners_3d\": [[756.14, 219.06], [763.19, 221.52], [659.45, 221.38], [657.66, 218.93], [756.14, 184.95], [763.19, 185.59], [659.45, 185.56], [657.66, 184.92]]}, {\"category\": \"car\", \"corners_3d\": [[712.41, 210.61], [714.44, 212.32], [637.74, 211.99], [639.01, 210.31], [712.41, 180.51], [714.44, 180.85], [637.74, 180.79], [639.01, 180.45]]}, {\"category\": \"van\", \"corners_3d\": [[594.92, 200.21], [572.78, 200.21], [576.15, 197.66], [596.23, 197.66], [594.92, 175.52], [572.78, 175.52], [576.15, 175.27], [596.23, 175.27]]}, {\"category\": \"van\", \"corners_3d\": [[489.49, 289.62], [381.9, 289.63], [446.14, 256.6], [523.29, 256.59], [489.49, 169.76], [381.9, 169.76], [446.14, 170.63], [523.29, 170.63]]}]\n```" - } - ] - }, - { - "id": 118, - "data_source": "Mono3DRefer", - "scene_name": "000068", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1036.9, 254.94], [1009.54, 254.94], [960.27, 244.84], [984.26, 244.84], [1036.9, 168.19], [1009.54, 168.19], [960.27, 168.76], [984.26, 168.76]]}, {\"category\": \"van\", \"corners_3d\": [[580.76, 212.79], [547.94, 212.79], [555.13, 208.09], [584.08, 208.09], [580.76, 171.85], [547.94, 171.85], [555.13, 171.97], [584.08, 171.97]]}, {\"category\": \"van\", \"corners_3d\": [[594.92, 200.21], [572.78, 200.21], [576.15, 197.66], [596.23, 197.66], [594.92, 175.52], [572.78, 175.52], [576.15, 175.27], [596.23, 175.27]]}, {\"category\": \"van\", \"corners_3d\": [[489.49, 289.62], [381.9, 289.63], [446.14, 256.6], [523.29, 256.59], [489.49, 169.76], [381.9, 169.76], [446.14, 170.63], [523.29, 170.63]]}]\n```" - } - ] - }, - { - "id": 119, - "data_source": "Mono3DRefer", - "scene_name": "000068", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1036.9, 254.94], [1009.54, 254.94], [960.27, 244.84], [984.26, 244.84], [1036.9, 168.19], [1009.54, 168.19], [960.27, 168.76], [984.26, 168.76]]}, {\"category\": \"car\", \"corners_3d\": [[306.99, 432.15], [68.67, 433.91], [305.06, 316.56], [435.77, 316.03], [306.99, 220.61], [68.67, 220.94], [305.06, 199.32], [435.77, 199.23]]}, {\"category\": \"car\", \"corners_3d\": [[-144.78, 478.15], [-409.54, 479.35], [56.38, 338.15], [198.43, 337.8], [-144.78, 237.17], [-409.54, 237.42], [56.38, 207.67], [198.43, 207.6]]}, {\"category\": \"car\", \"corners_3d\": [[540.05, 249.99], [479.03, 250.13], [499.12, 236.69], [549.52, 236.6], [540.05, 190.32], [479.03, 190.35], [499.12, 187.31], [549.52, 187.29]]}, {\"category\": \"car\", \"corners_3d\": [[567.78, 226.37], [530.35, 226.35], [539.82, 220.44], [573.13, 220.46], [567.78, 189.39], [530.35, 189.38], [539.82, 187.56], [573.13, 187.56]]}, {\"category\": \"car\", \"corners_3d\": [[779.85, 226.47], [791.03, 229.99], [672.15, 229.99], [668.3, 226.47], [779.85, 182.81], [791.03, 183.47], [672.15, 183.47], [668.3, 182.81]]}, {\"category\": \"car\", \"corners_3d\": [[756.14, 219.06], [763.19, 221.52], [659.45, 221.38], [657.66, 218.93], [756.14, 184.95], [763.19, 185.59], [659.45, 185.56], [657.66, 184.92]]}, {\"category\": \"car\", \"corners_3d\": [[712.41, 210.61], [714.44, 212.32], [637.74, 211.99], [639.01, 210.31], [712.41, 180.51], [714.44, 180.85], [637.74, 180.79], [639.01, 180.45]]}]\n```" - } - ] - }, - { - "id": 120, - "data_source": "Mono3DRefer", - "scene_name": "000068", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1036.9, 254.94], [1009.54, 254.94], [960.27, 244.84], [984.26, 244.84], [1036.9, 168.19], [1009.54, 168.19], [960.27, 168.76], [984.26, 168.76]]}, {\"category\": \"car\", \"corners_3d\": [[306.99, 432.15], [68.67, 433.91], [305.06, 316.56], [435.77, 316.03], [306.99, 220.61], [68.67, 220.94], [305.06, 199.32], [435.77, 199.23]]}, {\"category\": \"car\", \"corners_3d\": [[-144.78, 478.15], [-409.54, 479.35], [56.38, 338.15], [198.43, 337.8], [-144.78, 237.17], [-409.54, 237.42], [56.38, 207.67], [198.43, 207.6]]}, {\"category\": \"car\", \"corners_3d\": [[540.05, 249.99], [479.03, 250.13], [499.12, 236.69], [549.52, 236.6], [540.05, 190.32], [479.03, 190.35], [499.12, 187.31], [549.52, 187.29]]}, {\"category\": \"car\", \"corners_3d\": [[567.78, 226.37], [530.35, 226.35], [539.82, 220.44], [573.13, 220.46], [567.78, 189.39], [530.35, 189.38], [539.82, 187.56], [573.13, 187.56]]}, {\"category\": \"van\", \"corners_3d\": [[580.76, 212.79], [547.94, 212.79], [555.13, 208.09], [584.08, 208.09], [580.76, 171.85], [547.94, 171.85], [555.13, 171.97], [584.08, 171.97]]}, {\"category\": \"car\", \"corners_3d\": [[779.85, 226.47], [791.03, 229.99], [672.15, 229.99], [668.3, 226.47], [779.85, 182.81], [791.03, 183.47], [672.15, 183.47], [668.3, 182.81]]}, {\"category\": \"car\", \"corners_3d\": [[756.14, 219.06], [763.19, 221.52], [659.45, 221.38], [657.66, 218.93], [756.14, 184.95], [763.19, 185.59], [659.45, 185.56], [657.66, 184.92]]}, {\"category\": \"car\", \"corners_3d\": [[712.41, 210.61], [714.44, 212.32], [637.74, 211.99], [639.01, 210.31], [712.41, 180.51], [714.44, 180.85], [637.74, 180.79], [639.01, 180.45]]}, {\"category\": \"van\", \"corners_3d\": [[594.92, 200.21], [572.78, 200.21], [576.15, 197.66], [596.23, 197.66], [594.92, 175.52], [572.78, 175.52], [576.15, 175.27], [596.23, 175.27]]}, {\"category\": \"van\", \"corners_3d\": [[489.49, 289.62], [381.9, 289.63], [446.14, 256.6], [523.29, 256.59], [489.49, 169.76], [381.9, 169.76], [446.14, 170.63], [523.29, 170.63]]}]\n```" - } - ] - }, - { - "id": 121, - "data_source": "Mono3DRefer", - "scene_name": "000069", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[130.44, 257.15], [175.54, 248.69], [355.52, 247.92], [330.79, 256.2], [130.44, 189.05], [175.54, 187.43], [355.52, 187.28], [330.79, 188.87]]}, {\"category\": \"car\", \"corners_3d\": [[556.49, 193.65], [553.08, 194.22], [501.41, 194.06], [506.16, 193.51], [556.49, 178.11], [553.08, 178.25], [501.41, 178.21], [506.16, 178.07]]}]\n```" - } - ] - }, - { - "id": 122, - "data_source": "Mono3DRefer", - "scene_name": "000071", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[725.51, 185.07], [704.62, 185.03], [702.02, 183.04], [719.49, 183.07], [725.51, 150.42], [704.62, 150.49], [702.02, 154.14], [719.49, 154.09]]}]\n```" - } - ] - }, - { - "id": 123, - "data_source": "Mono3DRefer", - "scene_name": "000071", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[601.84, 229.68], [658.41, 229.77], [664.44, 239.91], [597.8, 239.78], [601.84, 174.88], [658.41, 174.89], [664.44, 175.25], [597.8, 175.24]]}]\n```" - } - ] - }, - { - "id": 124, - "data_source": "Mono3DRefer", - "scene_name": "000071", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[725.51, 185.07], [704.62, 185.03], [702.02, 183.04], [719.49, 183.07], [725.51, 150.42], [704.62, 150.49], [702.02, 154.14], [719.49, 154.09]]}, {\"category\": \"car\", \"corners_3d\": [[601.84, 229.68], [658.41, 229.77], [664.44, 239.91], [597.8, 239.78], [601.84, 174.88], [658.41, 174.89], [664.44, 175.25], [597.8, 175.24]]}]\n```" - } - ] - }, - { - "id": 125, - "data_source": "Mono3DRefer", - "scene_name": "000074", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1015.98, 302.78], [985.29, 305.66], [933.65, 296.04], [963.13, 293.59], [1015.98, 155.3], [985.29, 154.71], [933.65, 156.69], [963.13, 157.19]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[798.25, 359.23], [726.18, 362.74], [698.04, 342.34], [762.66, 339.56], [798.25, 125.05], [726.18, 123.97], [698.04, 130.3], [762.66, 131.16]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[869.44, 351.16], [805.21, 355.15], [776.02, 341.27], [835.87, 337.88], [869.44, 149.66], [805.21, 148.94], [776.02, 151.45], [835.87, 152.06]]}]\n```" - } - ] - }, - { - "id": 126, - "data_source": "Mono3DRefer", - "scene_name": "000078", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[640.14, 189.48], [616.11, 189.34], [629.56, 188.09], [651.85, 188.21], [640.14, 168.43], [616.11, 168.47], [629.56, 168.8], [651.85, 168.77]]}, {\"category\": \"car\", \"corners_3d\": [[372.28, 310.03], [222.6, 309.77], [355.32, 263.6], [454.64, 263.72], [372.28, 169.52], [222.6, 169.52], [355.32, 170.65], [454.64, 170.64]]}, {\"category\": \"car\", \"corners_3d\": [[665.49, 199.38], [696.99, 199.6], [692.95, 201.68], [659.05, 201.43], [665.49, 167.82], [696.99, 167.78], [692.95, 167.38], [659.05, 167.43]]}]\n```" - } - ] - }, - { - "id": 127, - "data_source": "Mono3DRefer", - "scene_name": "000078", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[104.56, 508.72], [-266.34, 525.02], [176.27, 330.89], [337.63, 327.52], [104.56, 147.39], [-266.34, 146.16], [176.27, 160.87], [337.63, 161.13]]}]\n```" - } - ] - }, - { - "id": 128, - "data_source": "Mono3DRefer", - "scene_name": "000078", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[640.14, 189.48], [616.11, 189.34], [629.56, 188.09], [651.85, 188.21], [640.14, 168.43], [616.11, 168.47], [629.56, 168.8], [651.85, 168.77]]}, {\"category\": \"van\", \"corners_3d\": [[104.56, 508.72], [-266.34, 525.02], [176.27, 330.89], [337.63, 327.52], [104.56, 147.39], [-266.34, 146.16], [176.27, 160.87], [337.63, 161.13]]}, {\"category\": \"car\", \"corners_3d\": [[372.28, 310.03], [222.6, 309.77], [355.32, 263.6], [454.64, 263.72], [372.28, 169.52], [222.6, 169.52], [355.32, 170.65], [454.64, 170.64]]}, {\"category\": \"car\", \"corners_3d\": [[665.49, 199.38], [696.99, 199.6], [692.95, 201.68], [659.05, 201.43], [665.49, 167.82], [696.99, 167.78], [692.95, 167.38], [659.05, 167.43]]}]\n```" - } - ] - }, - { - "id": 129, - "data_source": "Mono3DRefer", - "scene_name": "000079", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[121.63, 269.38], [23.47, 266.12], [397.82, 218.89], [451.77, 219.67], [121.63, 81.89], [23.47, 84.96], [397.82, 129.47], [451.77, 128.73]]}]\n```" - } - ] - }, - { - "id": 130, - "data_source": "Mono3DRefer", - "scene_name": "000079", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[650.73, 189.25], [630.04, 189.42], [590.5, 187.05], [608.6, 186.92], [650.73, 159.39], [630.04, 159.25], [590.5, 161.2], [608.6, 161.3]]}]\n```" - } - ] - }, - { - "id": 131, - "data_source": "Mono3DRefer", - "scene_name": "000079", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[678.67, 194.15], [700.49, 194.29], [693.57, 196.38], [669.69, 196.21], [678.67, 174.76], [700.49, 174.77], [693.57, 174.96], [669.69, 174.94]]}, {\"category\": \"car\", \"corners_3d\": [[713.29, 198.43], [742.34, 198.71], [733.69, 201.68], [701.45, 201.33], [713.29, 172.3], [742.34, 172.29], [733.69, 172.22], [701.45, 172.23]]}, {\"category\": \"car\", \"corners_3d\": [[483.18, 264.86], [568.11, 268.54], [467.14, 303.91], [357.7, 297.09], [483.18, 181.98], [568.11, 182.34], [467.14, 185.85], [357.7, 185.18]]}, {\"category\": \"car\", \"corners_3d\": [[248.17, 272.23], [326.6, 276.71], [39.02, 332.99], [-60.45, 322.58], [248.17, 182.36], [326.6, 182.78], [39.02, 188.17], [-60.45, 187.17]]}, {\"category\": \"car\", \"corners_3d\": [[505.87, 207.27], [483.55, 207.15], [497.9, 204.75], [518.7, 204.85], [505.87, 182.85], [483.55, 182.82], [497.9, 182.12], [518.7, 182.15]]}]\n```" - } - ] - }, - { - "id": 132, - "data_source": "Mono3DRefer", - "scene_name": "000079", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[121.63, 269.38], [23.47, 266.12], [397.82, 218.89], [451.77, 219.67], [121.63, 81.89], [23.47, 84.96], [397.82, 129.47], [451.77, 128.73]]}, {\"category\": \"tram\", \"corners_3d\": [[650.73, 189.25], [630.04, 189.42], [590.5, 187.05], [608.6, 186.92], [650.73, 159.39], [630.04, 159.25], [590.5, 161.2], [608.6, 161.3]]}]\n```" - } - ] - }, - { - "id": 133, - "data_source": "Mono3DRefer", - "scene_name": "000079", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[678.67, 194.15], [700.49, 194.29], [693.57, 196.38], [669.69, 196.21], [678.67, 174.76], [700.49, 174.77], [693.57, 174.96], [669.69, 174.94]]}, {\"category\": \"truck\", \"corners_3d\": [[121.63, 269.38], [23.47, 266.12], [397.82, 218.89], [451.77, 219.67], [121.63, 81.89], [23.47, 84.96], [397.82, 129.47], [451.77, 128.73]]}, {\"category\": \"car\", \"corners_3d\": [[713.29, 198.43], [742.34, 198.71], [733.69, 201.68], [701.45, 201.33], [713.29, 172.3], [742.34, 172.29], [733.69, 172.22], [701.45, 172.23]]}, {\"category\": \"car\", \"corners_3d\": [[483.18, 264.86], [568.11, 268.54], [467.14, 303.91], [357.7, 297.09], [483.18, 181.98], [568.11, 182.34], [467.14, 185.85], [357.7, 185.18]]}, {\"category\": \"car\", \"corners_3d\": [[248.17, 272.23], [326.6, 276.71], [39.02, 332.99], [-60.45, 322.58], [248.17, 182.36], [326.6, 182.78], [39.02, 188.17], [-60.45, 187.17]]}, {\"category\": \"car\", \"corners_3d\": [[505.87, 207.27], [483.55, 207.15], [497.9, 204.75], [518.7, 204.85], [505.87, 182.85], [483.55, 182.82], [497.9, 182.12], [518.7, 182.15]]}]\n```" - } - ] - }, - { - "id": 134, - "data_source": "Mono3DRefer", - "scene_name": "000079", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[678.67, 194.15], [700.49, 194.29], [693.57, 196.38], [669.69, 196.21], [678.67, 174.76], [700.49, 174.77], [693.57, 174.96], [669.69, 174.94]]}, {\"category\": \"car\", \"corners_3d\": [[713.29, 198.43], [742.34, 198.71], [733.69, 201.68], [701.45, 201.33], [713.29, 172.3], [742.34, 172.29], [733.69, 172.22], [701.45, 172.23]]}, {\"category\": \"car\", \"corners_3d\": [[483.18, 264.86], [568.11, 268.54], [467.14, 303.91], [357.7, 297.09], [483.18, 181.98], [568.11, 182.34], [467.14, 185.85], [357.7, 185.18]]}, {\"category\": \"car\", \"corners_3d\": [[248.17, 272.23], [326.6, 276.71], [39.02, 332.99], [-60.45, 322.58], [248.17, 182.36], [326.6, 182.78], [39.02, 188.17], [-60.45, 187.17]]}, {\"category\": \"car\", \"corners_3d\": [[505.87, 207.27], [483.55, 207.15], [497.9, 204.75], [518.7, 204.85], [505.87, 182.85], [483.55, 182.82], [497.9, 182.12], [518.7, 182.15]]}, {\"category\": \"tram\", \"corners_3d\": [[650.73, 189.25], [630.04, 189.42], [590.5, 187.05], [608.6, 186.92], [650.73, 159.39], [630.04, 159.25], [590.5, 161.2], [608.6, 161.3]]}]\n```" - } - ] - }, - { - "id": 135, - "data_source": "Mono3DRefer", - "scene_name": "000079", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[678.67, 194.15], [700.49, 194.29], [693.57, 196.38], [669.69, 196.21], [678.67, 174.76], [700.49, 174.77], [693.57, 174.96], [669.69, 174.94]]}, {\"category\": \"truck\", \"corners_3d\": [[121.63, 269.38], [23.47, 266.12], [397.82, 218.89], [451.77, 219.67], [121.63, 81.89], [23.47, 84.96], [397.82, 129.47], [451.77, 128.73]]}, {\"category\": \"car\", \"corners_3d\": [[713.29, 198.43], [742.34, 198.71], [733.69, 201.68], [701.45, 201.33], [713.29, 172.3], [742.34, 172.29], [733.69, 172.22], [701.45, 172.23]]}, {\"category\": \"car\", \"corners_3d\": [[483.18, 264.86], [568.11, 268.54], [467.14, 303.91], [357.7, 297.09], [483.18, 181.98], [568.11, 182.34], [467.14, 185.85], [357.7, 185.18]]}, {\"category\": \"car\", \"corners_3d\": [[248.17, 272.23], [326.6, 276.71], [39.02, 332.99], [-60.45, 322.58], [248.17, 182.36], [326.6, 182.78], [39.02, 188.17], [-60.45, 187.17]]}, {\"category\": \"car\", \"corners_3d\": [[505.87, 207.27], [483.55, 207.15], [497.9, 204.75], [518.7, 204.85], [505.87, 182.85], [483.55, 182.82], [497.9, 182.12], [518.7, 182.15]]}, {\"category\": \"tram\", \"corners_3d\": [[650.73, 189.25], [630.04, 189.42], [590.5, 187.05], [608.6, 186.92], [650.73, 159.39], [630.04, 159.25], [590.5, 161.2], [608.6, 161.3]]}]\n```" - } - ] - }, - { - "id": 136, - "data_source": "Mono3DRefer", - "scene_name": "000080", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[510.92, 229.97], [458.0, 229.89], [484.28, 220.83], [528.81, 220.89], [510.92, 177.49], [458.0, 177.49], [484.28, 176.75], [528.81, 176.75]]}, {\"category\": \"car\", \"corners_3d\": [[562.2, 202.86], [536.19, 202.87], [541.5, 200.44], [565.4, 200.43], [562.2, 177.37], [536.19, 177.38], [541.5, 177.01], [565.4, 177.01]]}, {\"category\": \"car\", \"corners_3d\": [[577.39, 192.18], [559.9, 192.18], [562.15, 191.29], [578.84, 191.29], [577.39, 176.2], [559.9, 176.2], [562.15, 176.04], [578.84, 176.04]]}, {\"category\": \"car\", \"corners_3d\": [[592.84, 210.24], [630.01, 210.24], [633.3, 216.44], [589.97, 216.44], [592.84, 173.3], [630.01, 173.3], [633.3, 173.38], [589.97, 173.38]]}]\n```" - } - ] - }, - { - "id": 137, - "data_source": "Mono3DRefer", - "scene_name": "000083", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[219.67, 286.48], [103.44, 286.32], [253.02, 255.19], [334.84, 255.27], [219.67, 160.39], [103.44, 160.43], [253.02, 166.69], [334.84, 166.67]]}]\n```" - } - ] - }, - { - "id": 138, - "data_source": "Mono3DRefer", - "scene_name": "000083", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[141.22, 409.25], [362.15, 409.3], [-484.29, 1216.71], [-1486.51, 1215.56], [141.22, 204.93], [362.15, 204.94], [-484.29, 287.98], [-1486.51, 287.86]]}, {\"category\": \"car\", \"corners_3d\": [[425.1, 255.13], [503.36, 255.14], [472.92, 278.11], [370.25, 278.1], [425.1, 183.47], [503.36, 183.47], [472.92, 184.08], [370.25, 184.08]]}, {\"category\": \"car\", \"corners_3d\": [[591.61, 213.13], [637.21, 213.15], [643.02, 219.47], [588.31, 219.44], [591.61, 171.65], [637.21, 171.64], [643.02, 169.67], [588.31, 169.68]]}, {\"category\": \"car\", \"corners_3d\": [[532.01, 210.95], [571.69, 210.99], [565.34, 215.28], [519.9, 215.23], [532.01, 171.98], [571.69, 171.97], [565.34, 170.58], [519.9, 170.6]]}, {\"category\": \"car\", \"corners_3d\": [[497.08, 205.36], [463.8, 205.31], [482.01, 202.68], [511.64, 202.72], [497.08, 174.81], [463.8, 174.82], [482.01, 175.56], [511.64, 175.55]]}]\n```" - } - ] - }, - { - "id": 139, - "data_source": "Mono3DRefer", - "scene_name": "000083", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[141.22, 409.25], [362.15, 409.3], [-484.29, 1216.71], [-1486.51, 1215.56], [141.22, 204.93], [362.15, 204.94], [-484.29, 287.98], [-1486.51, 287.86]]}, {\"category\": \"car\", \"corners_3d\": [[425.1, 255.13], [503.36, 255.14], [472.92, 278.11], [370.25, 278.1], [425.1, 183.47], [503.36, 183.47], [472.92, 184.08], [370.25, 184.08]]}, {\"category\": \"car\", \"corners_3d\": [[591.61, 213.13], [637.21, 213.15], [643.02, 219.47], [588.31, 219.44], [591.61, 171.65], [637.21, 171.64], [643.02, 169.67], [588.31, 169.68]]}, {\"category\": \"car\", \"corners_3d\": [[532.01, 210.95], [571.69, 210.99], [565.34, 215.28], [519.9, 215.23], [532.01, 171.98], [571.69, 171.97], [565.34, 170.58], [519.9, 170.6]]}, {\"category\": \"van\", \"corners_3d\": [[219.67, 286.48], [103.44, 286.32], [253.02, 255.19], [334.84, 255.27], [219.67, 160.39], [103.44, 160.43], [253.02, 166.69], [334.84, 166.67]]}, {\"category\": \"car\", \"corners_3d\": [[497.08, 205.36], [463.8, 205.31], [482.01, 202.68], [511.64, 202.72], [497.08, 174.81], [463.8, 174.82], [482.01, 175.56], [511.64, 175.55]]}]\n```" - } - ] - }, - { - "id": 140, - "data_source": "Mono3DRefer", - "scene_name": "000084", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[506.07, 227.04], [460.04, 227.72], [460.39, 220.64], [500.53, 220.12], [506.07, 180.59], [460.04, 180.68], [460.39, 179.67], [500.53, 179.6]]}, {\"category\": \"car\", \"corners_3d\": [[486.66, 208.5], [455.68, 208.97], [448.27, 205.53], [476.43, 205.15], [486.66, 182.16], [455.68, 182.29], [448.27, 181.39], [476.43, 181.29]]}]\n```" - } - ] - }, - { - "id": 141, - "data_source": "Mono3DRefer", - "scene_name": "000085", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[362.54, 298.23], [328.83, 304.45], [297.22, 299.94], [331.28, 294.13], [362.54, 178.18], [328.83, 178.44], [297.22, 178.25], [331.28, 178.01]]}]\n```" - } - ] - }, - { - "id": 142, - "data_source": "Mono3DRefer", - "scene_name": "000085", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[682.51, 236.76], [705.85, 237.05], [705.17, 244.35], [679.19, 243.99], [682.51, 165.39], [705.85, 165.36], [705.17, 164.5], [679.19, 164.54]]}]\n```" - } - ] - }, - { - "id": 143, - "data_source": "Mono3DRefer", - "scene_name": "000085", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[70.17, 321.25], [245.68, 322.77], [-522.08, 615.39], [-1007.59, 602.41], [70.17, 103.91], [245.68, 103.2], [-522.08, -32.75], [-1007.59, -26.72]]}]\n```" - } - ] - }, - { - "id": 144, - "data_source": "Mono3DRefer", - "scene_name": "000085", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-57.08, 244.18], [7.41, 237.78], [156.7, 238.8], [106.07, 245.42], [-57.08, 180.76], [7.41, 180.05], [156.7, 180.17], [106.07, 180.9]]}]\n```" - } - ] - }, - { - "id": 145, - "data_source": "Mono3DRefer", - "scene_name": "000085", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[682.51, 236.76], [705.85, 237.05], [705.17, 244.35], [679.19, 243.99], [682.51, 165.39], [705.85, 165.36], [705.17, 164.5], [679.19, 164.54]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[362.54, 298.23], [328.83, 304.45], [297.22, 299.94], [331.28, 294.13], [362.54, 178.18], [328.83, 178.44], [297.22, 178.25], [331.28, 178.01]]}]\n```" - } - ] - }, - { - "id": 146, - "data_source": "Mono3DRefer", - "scene_name": "000085", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[70.17, 321.25], [245.68, 322.77], [-522.08, 615.39], [-1007.59, 602.41], [70.17, 103.91], [245.68, 103.2], [-522.08, -32.75], [-1007.59, -26.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[362.54, 298.23], [328.83, 304.45], [297.22, 299.94], [331.28, 294.13], [362.54, 178.18], [328.83, 178.44], [297.22, 178.25], [331.28, 178.01]]}]\n```" - } - ] - }, - { - "id": 147, - "data_source": "Mono3DRefer", - "scene_name": "000085", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[362.54, 298.23], [328.83, 304.45], [297.22, 299.94], [331.28, 294.13], [362.54, 178.18], [328.83, 178.44], [297.22, 178.25], [331.28, 178.01]]}, {\"category\": \"car\", \"corners_3d\": [[-57.08, 244.18], [7.41, 237.78], [156.7, 238.8], [106.07, 245.42], [-57.08, 180.76], [7.41, 180.05], [156.7, 180.17], [106.07, 180.9]]}]\n```" - } - ] - }, - { - "id": 148, - "data_source": "Mono3DRefer", - "scene_name": "000085", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[70.17, 321.25], [245.68, 322.77], [-522.08, 615.39], [-1007.59, 602.41], [70.17, 103.91], [245.68, 103.2], [-522.08, -32.75], [-1007.59, -26.72]]}, {\"category\": \"cyclist\", \"corners_3d\": [[682.51, 236.76], [705.85, 237.05], [705.17, 244.35], [679.19, 243.99], [682.51, 165.39], [705.85, 165.36], [705.17, 164.5], [679.19, 164.54]]}]\n```" - } - ] - }, - { - "id": 149, - "data_source": "Mono3DRefer", - "scene_name": "000085", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[682.51, 236.76], [705.85, 237.05], [705.17, 244.35], [679.19, 243.99], [682.51, 165.39], [705.85, 165.36], [705.17, 164.5], [679.19, 164.54]]}, {\"category\": \"car\", \"corners_3d\": [[-57.08, 244.18], [7.41, 237.78], [156.7, 238.8], [106.07, 245.42], [-57.08, 180.76], [7.41, 180.05], [156.7, 180.17], [106.07, 180.9]]}]\n```" - } - ] - }, - { - "id": 150, - "data_source": "Mono3DRefer", - "scene_name": "000085", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[70.17, 321.25], [245.68, 322.77], [-522.08, 615.39], [-1007.59, 602.41], [70.17, 103.91], [245.68, 103.2], [-522.08, -32.75], [-1007.59, -26.72]]}, {\"category\": \"car\", \"corners_3d\": [[-57.08, 244.18], [7.41, 237.78], [156.7, 238.8], [106.07, 245.42], [-57.08, 180.76], [7.41, 180.05], [156.7, 180.17], [106.07, 180.9]]}]\n```" - } - ] - }, - { - "id": 151, - "data_source": "Mono3DRefer", - "scene_name": "000085", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[70.17, 321.25], [245.68, 322.77], [-522.08, 615.39], [-1007.59, 602.41], [70.17, 103.91], [245.68, 103.2], [-522.08, -32.75], [-1007.59, -26.72]]}, {\"category\": \"cyclist\", \"corners_3d\": [[682.51, 236.76], [705.85, 237.05], [705.17, 244.35], [679.19, 243.99], [682.51, 165.39], [705.85, 165.36], [705.17, 164.5], [679.19, 164.54]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[362.54, 298.23], [328.83, 304.45], [297.22, 299.94], [331.28, 294.13], [362.54, 178.18], [328.83, 178.44], [297.22, 178.25], [331.28, 178.01]]}]\n```" - } - ] - }, - { - "id": 152, - "data_source": "Mono3DRefer", - "scene_name": "000085", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[682.51, 236.76], [705.85, 237.05], [705.17, 244.35], [679.19, 243.99], [682.51, 165.39], [705.85, 165.36], [705.17, 164.5], [679.19, 164.54]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[362.54, 298.23], [328.83, 304.45], [297.22, 299.94], [331.28, 294.13], [362.54, 178.18], [328.83, 178.44], [297.22, 178.25], [331.28, 178.01]]}, {\"category\": \"car\", \"corners_3d\": [[-57.08, 244.18], [7.41, 237.78], [156.7, 238.8], [106.07, 245.42], [-57.08, 180.76], [7.41, 180.05], [156.7, 180.17], [106.07, 180.9]]}]\n```" - } - ] - }, - { - "id": 153, - "data_source": "Mono3DRefer", - "scene_name": "000085", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[70.17, 321.25], [245.68, 322.77], [-522.08, 615.39], [-1007.59, 602.41], [70.17, 103.91], [245.68, 103.2], [-522.08, -32.75], [-1007.59, -26.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[362.54, 298.23], [328.83, 304.45], [297.22, 299.94], [331.28, 294.13], [362.54, 178.18], [328.83, 178.44], [297.22, 178.25], [331.28, 178.01]]}, {\"category\": \"car\", \"corners_3d\": [[-57.08, 244.18], [7.41, 237.78], [156.7, 238.8], [106.07, 245.42], [-57.08, 180.76], [7.41, 180.05], [156.7, 180.17], [106.07, 180.9]]}]\n```" - } - ] - }, - { - "id": 154, - "data_source": "Mono3DRefer", - "scene_name": "000085", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[70.17, 321.25], [245.68, 322.77], [-522.08, 615.39], [-1007.59, 602.41], [70.17, 103.91], [245.68, 103.2], [-522.08, -32.75], [-1007.59, -26.72]]}, {\"category\": \"cyclist\", \"corners_3d\": [[682.51, 236.76], [705.85, 237.05], [705.17, 244.35], [679.19, 243.99], [682.51, 165.39], [705.85, 165.36], [705.17, 164.5], [679.19, 164.54]]}, {\"category\": \"car\", \"corners_3d\": [[-57.08, 244.18], [7.41, 237.78], [156.7, 238.8], [106.07, 245.42], [-57.08, 180.76], [7.41, 180.05], [156.7, 180.17], [106.07, 180.9]]}]\n```" - } - ] - }, - { - "id": 155, - "data_source": "Mono3DRefer", - "scene_name": "000085", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist, van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[70.17, 321.25], [245.68, 322.77], [-522.08, 615.39], [-1007.59, 602.41], [70.17, 103.91], [245.68, 103.2], [-522.08, -32.75], [-1007.59, -26.72]]}, {\"category\": \"cyclist\", \"corners_3d\": [[682.51, 236.76], [705.85, 237.05], [705.17, 244.35], [679.19, 243.99], [682.51, 165.39], [705.85, 165.36], [705.17, 164.5], [679.19, 164.54]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[362.54, 298.23], [328.83, 304.45], [297.22, 299.94], [331.28, 294.13], [362.54, 178.18], [328.83, 178.44], [297.22, 178.25], [331.28, 178.01]]}, {\"category\": \"car\", \"corners_3d\": [[-57.08, 244.18], [7.41, 237.78], [156.7, 238.8], [106.07, 245.42], [-57.08, 180.76], [7.41, 180.05], [156.7, 180.17], [106.07, 180.9]]}]\n```" - } - ] - }, - { - "id": 156, - "data_source": "Mono3DRefer", - "scene_name": "000087", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[98.67, 259.38], [22.65, 258.82], [166.47, 239.18], [225.76, 239.52], [98.67, 186.42], [22.65, 186.33], [166.47, 183.25], [225.76, 183.31]]}, {\"category\": \"car\", \"corners_3d\": [[695.13, 203.62], [727.64, 203.73], [735.63, 207.86], [698.75, 207.72], [695.13, 174.67], [727.64, 174.68], [735.63, 174.92], [698.75, 174.91]]}, {\"category\": \"car\", \"corners_3d\": [[797.58, 185.5], [797.52, 185.21], [843.54, 185.4], [844.72, 185.7], [797.58, 169.71], [797.52, 169.79], [843.54, 169.74], [844.72, 169.67]]}, {\"category\": \"car\", \"corners_3d\": [[887.83, 185.41], [886.32, 185.11], [936.24, 185.34], [939.05, 185.65], [887.83, 167.09], [886.32, 167.23], [936.24, 167.12], [939.05, 166.98]]}]\n```" - } - ] - }, - { - "id": 157, - "data_source": "Mono3DRefer", - "scene_name": "000087", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1060.74, 181.14], [1053.73, 180.92], [1106.95, 181.05], [1115.59, 181.29], [1060.74, 158.82], [1053.73, 159.21], [1106.95, 158.97], [1115.59, 158.58]]}]\n```" - } - ] - }, - { - "id": 158, - "data_source": "Mono3DRefer", - "scene_name": "000087", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[567.56, 196.52], [570.58, 196.86], [562.67, 196.95], [559.75, 196.61], [567.56, 171.69], [570.58, 171.68], [562.67, 171.67], [559.75, 171.69]]}]\n```" - } - ] - }, - { - "id": 159, - "data_source": "Mono3DRefer", - "scene_name": "000087", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[98.67, 259.38], [22.65, 258.82], [166.47, 239.18], [225.76, 239.52], [98.67, 186.42], [22.65, 186.33], [166.47, 183.25], [225.76, 183.31]]}, {\"category\": \"car\", \"corners_3d\": [[695.13, 203.62], [727.64, 203.73], [735.63, 207.86], [698.75, 207.72], [695.13, 174.67], [727.64, 174.68], [735.63, 174.92], [698.75, 174.91]]}, {\"category\": \"car\", \"corners_3d\": [[797.58, 185.5], [797.52, 185.21], [843.54, 185.4], [844.72, 185.7], [797.58, 169.71], [797.52, 169.79], [843.54, 169.74], [844.72, 169.67]]}, {\"category\": \"car\", \"corners_3d\": [[887.83, 185.41], [886.32, 185.11], [936.24, 185.34], [939.05, 185.65], [887.83, 167.09], [886.32, 167.23], [936.24, 167.12], [939.05, 166.98]]}, {\"category\": \"van\", \"corners_3d\": [[1060.74, 181.14], [1053.73, 180.92], [1106.95, 181.05], [1115.59, 181.29], [1060.74, 158.82], [1053.73, 159.21], [1106.95, 158.97], [1115.59, 158.58]]}]\n```" - } - ] - }, - { - "id": 160, - "data_source": "Mono3DRefer", - "scene_name": "000087", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[98.67, 259.38], [22.65, 258.82], [166.47, 239.18], [225.76, 239.52], [98.67, 186.42], [22.65, 186.33], [166.47, 183.25], [225.76, 183.31]]}, {\"category\": \"car\", \"corners_3d\": [[695.13, 203.62], [727.64, 203.73], [735.63, 207.86], [698.75, 207.72], [695.13, 174.67], [727.64, 174.68], [735.63, 174.92], [698.75, 174.91]]}, {\"category\": \"car\", \"corners_3d\": [[797.58, 185.5], [797.52, 185.21], [843.54, 185.4], [844.72, 185.7], [797.58, 169.71], [797.52, 169.79], [843.54, 169.74], [844.72, 169.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[567.56, 196.52], [570.58, 196.86], [562.67, 196.95], [559.75, 196.61], [567.56, 171.69], [570.58, 171.68], [562.67, 171.67], [559.75, 171.69]]}, {\"category\": \"car\", \"corners_3d\": [[887.83, 185.41], [886.32, 185.11], [936.24, 185.34], [939.05, 185.65], [887.83, 167.09], [886.32, 167.23], [936.24, 167.12], [939.05, 166.98]]}]\n```" - } - ] - }, - { - "id": 161, - "data_source": "Mono3DRefer", - "scene_name": "000087", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[567.56, 196.52], [570.58, 196.86], [562.67, 196.95], [559.75, 196.61], [567.56, 171.69], [570.58, 171.68], [562.67, 171.67], [559.75, 171.69]]}, {\"category\": \"van\", \"corners_3d\": [[1060.74, 181.14], [1053.73, 180.92], [1106.95, 181.05], [1115.59, 181.29], [1060.74, 158.82], [1053.73, 159.21], [1106.95, 158.97], [1115.59, 158.58]]}]\n```" - } - ] - }, - { - "id": 162, - "data_source": "Mono3DRefer", - "scene_name": "000087", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[98.67, 259.38], [22.65, 258.82], [166.47, 239.18], [225.76, 239.52], [98.67, 186.42], [22.65, 186.33], [166.47, 183.25], [225.76, 183.31]]}, {\"category\": \"car\", \"corners_3d\": [[695.13, 203.62], [727.64, 203.73], [735.63, 207.86], [698.75, 207.72], [695.13, 174.67], [727.64, 174.68], [735.63, 174.92], [698.75, 174.91]]}, {\"category\": \"car\", \"corners_3d\": [[797.58, 185.5], [797.52, 185.21], [843.54, 185.4], [844.72, 185.7], [797.58, 169.71], [797.52, 169.79], [843.54, 169.74], [844.72, 169.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[567.56, 196.52], [570.58, 196.86], [562.67, 196.95], [559.75, 196.61], [567.56, 171.69], [570.58, 171.68], [562.67, 171.67], [559.75, 171.69]]}, {\"category\": \"car\", \"corners_3d\": [[887.83, 185.41], [886.32, 185.11], [936.24, 185.34], [939.05, 185.65], [887.83, 167.09], [886.32, 167.23], [936.24, 167.12], [939.05, 166.98]]}, {\"category\": \"van\", \"corners_3d\": [[1060.74, 181.14], [1053.73, 180.92], [1106.95, 181.05], [1115.59, 181.29], [1060.74, 158.82], [1053.73, 159.21], [1106.95, 158.97], [1115.59, 158.58]]}]\n```" - } - ] - }, - { - "id": 163, - "data_source": "Mono3DRefer", - "scene_name": "000088", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[372.8, 241.38], [347.01, 247.69], [175.93, 246.6], [215.59, 240.47], [372.8, 192.06], [347.01, 193.83], [175.93, 193.52], [215.59, 191.8]]}]\n```" - } - ] - }, - { - "id": 164, - "data_source": "Mono3DRefer", - "scene_name": "000089", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[408.75, 270.34], [504.16, 270.22], [458.65, 315.09], [319.13, 315.34], [408.75, 196.24], [504.16, 196.21], [458.65, 206.97], [319.13, 207.03]]}, {\"category\": \"car\", \"corners_3d\": [[657.95, 216.79], [699.33, 216.72], [720.43, 225.07], [671.2, 225.17], [657.95, 177.11], [699.33, 177.1], [720.43, 177.91], [671.2, 177.92]]}, {\"category\": \"car\", \"corners_3d\": [[488.34, 230.44], [538.64, 230.49], [520.41, 243.84], [458.48, 243.77], [488.34, 183.7], [538.64, 183.71], [520.41, 186.23], [458.48, 186.22]]}, {\"category\": \"car\", \"corners_3d\": [[705.0, 257.39], [784.16, 257.21], [853.55, 288.27], [745.36, 288.6], [705.0, 177.84], [784.16, 177.83], [853.55, 179.66], [745.36, 179.68]]}, {\"category\": \"car\", \"corners_3d\": [[637.96, 200.85], [665.21, 200.93], [664.97, 203.4], [635.34, 203.29], [637.96, 177.98], [665.21, 178.0], [664.97, 178.45], [635.34, 178.43]]}]\n```" - } - ] - }, - { - "id": 165, - "data_source": "Mono3DRefer", - "scene_name": "000089", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[746.62, 284.01], [878.91, 283.61], [1368.95, 475.26], [1010.3, 478.21], [746.62, 100.13], [878.91, 100.39], [1368.95, -24.99], [1010.3, -26.92]]}]\n```" - } - ] - }, - { - "id": 166, - "data_source": "Mono3DRefer", - "scene_name": "000089", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[408.75, 270.34], [504.16, 270.22], [458.65, 315.09], [319.13, 315.34], [408.75, 196.24], [504.16, 196.21], [458.65, 206.97], [319.13, 207.03]]}, {\"category\": \"van\", \"corners_3d\": [[746.62, 284.01], [878.91, 283.61], [1368.95, 475.26], [1010.3, 478.21], [746.62, 100.13], [878.91, 100.39], [1368.95, -24.99], [1010.3, -26.92]]}, {\"category\": \"car\", \"corners_3d\": [[657.95, 216.79], [699.33, 216.72], [720.43, 225.07], [671.2, 225.17], [657.95, 177.11], [699.33, 177.1], [720.43, 177.91], [671.2, 177.92]]}, {\"category\": \"car\", \"corners_3d\": [[488.34, 230.44], [538.64, 230.49], [520.41, 243.84], [458.48, 243.77], [488.34, 183.7], [538.64, 183.71], [520.41, 186.23], [458.48, 186.22]]}, {\"category\": \"car\", \"corners_3d\": [[705.0, 257.39], [784.16, 257.21], [853.55, 288.27], [745.36, 288.6], [705.0, 177.84], [784.16, 177.83], [853.55, 179.66], [745.36, 179.68]]}, {\"category\": \"car\", \"corners_3d\": [[637.96, 200.85], [665.21, 200.93], [664.97, 203.4], [635.34, 203.29], [637.96, 177.98], [665.21, 178.0], [664.97, 178.45], [635.34, 178.43]]}]\n```" - } - ] - }, - { - "id": 167, - "data_source": "Mono3DRefer", - "scene_name": "000090", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[5.56, 259.35], [82.98, 253.72], [126.94, 263.52], [42.86, 270.65], [5.56, 200.3], [82.98, 198.51], [126.94, 201.62], [42.86, 203.88]]}]\n```" - } - ] - }, - { - "id": 168, - "data_source": "Mono3DRefer", - "scene_name": "000091", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[904.0, 223.58], [923.6, 223.7], [947.41, 228.57], [925.87, 228.43], [904.0, 158.45], [923.6, 158.42], [947.41, 157.03], [925.87, 157.08]]}]\n```" - } - ] - }, - { - "id": 169, - "data_source": "Mono3DRefer", - "scene_name": "000091", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[710.48, 201.36], [737.21, 201.52], [738.87, 204.15], [709.68, 203.96], [710.48, 175.36], [737.21, 175.38], [738.87, 175.61], [709.68, 175.59]]}]\n```" - } - ] - }, - { - "id": 170, - "data_source": "Mono3DRefer", - "scene_name": "000091", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[904.0, 223.58], [923.6, 223.7], [947.41, 228.57], [925.87, 228.43], [904.0, 158.45], [923.6, 158.42], [947.41, 157.03], [925.87, 157.08]]}, {\"category\": \"car\", \"corners_3d\": [[710.48, 201.36], [737.21, 201.52], [738.87, 204.15], [709.68, 203.96], [710.48, 175.36], [737.21, 175.38], [738.87, 175.61], [709.68, 175.59]]}]\n```" - } - ] - }, - { - "id": 171, - "data_source": "Mono3DRefer", - "scene_name": "000093", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[115.75, 417.78], [-124.99, 418.65], [191.82, 311.51], [327.18, 311.23], [115.75, 197.44], [-124.99, 197.53], [191.82, 186.77], [327.18, 186.74]]}, {\"category\": \"car\", \"corners_3d\": [[723.09, 224.07], [777.3, 224.26], [813.15, 238.32], [744.01, 238.01], [723.09, 170.42], [777.3, 170.41], [813.15, 169.74], [744.01, 169.76]]}, {\"category\": \"car\", \"corners_3d\": [[353.24, 298.94], [246.88, 298.39], [346.54, 265.79], [425.51, 266.09], [353.24, 202.09], [246.88, 201.96], [346.54, 194.41], [425.51, 194.47]]}, {\"category\": \"car\", \"corners_3d\": [[437.97, 260.03], [359.7, 259.66], [417.68, 241.56], [479.77, 241.8], [437.97, 188.91], [359.7, 188.84], [417.68, 185.51], [479.77, 185.55]]}, {\"category\": \"car\", \"corners_3d\": [[579.29, 202.58], [550.91, 202.48], [561.02, 200.05], [587.1, 200.14], [579.29, 177.17], [550.91, 177.15], [561.02, 176.8], [587.1, 176.81]]}, {\"category\": \"car\", \"corners_3d\": [[622.56, 190.4], [605.9, 190.35], [610.19, 189.58], [626.13, 189.62], [622.56, 175.16], [605.9, 175.16], [610.19, 175.05], [626.13, 175.06]]}]\n```" - } - ] - }, - { - "id": 172, - "data_source": "Mono3DRefer", - "scene_name": "000095", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[601.29, 234.59], [657.27, 234.59], [667.17, 247.55], [599.44, 247.54], [601.29, 181.66], [657.27, 181.66], [667.17, 183.51], [599.44, 183.51]]}, {\"category\": \"car\", \"corners_3d\": [[878.23, 245.89], [963.79, 252.34], [870.39, 271.38], [775.58, 261.66], [878.23, 172.84], [963.79, 172.84], [870.39, 172.83], [775.58, 172.83]]}, {\"category\": \"car\", \"corners_3d\": [[765.17, 221.66], [802.52, 223.72], [745.14, 229.31], [706.44, 226.79], [765.17, 179.26], [802.52, 179.53], [745.14, 180.26], [706.44, 179.93]]}, {\"category\": \"car\", \"corners_3d\": [[727.25, 212.02], [757.08, 213.47], [703.9, 217.0], [673.64, 215.29], [727.25, 181.26], [757.08, 181.57], [703.9, 182.33], [673.64, 181.96]]}, {\"category\": \"car\", \"corners_3d\": [[245.38, 291.12], [146.39, 290.97], [266.25, 260.84], [340.08, 260.92], [245.38, 199.39], [146.39, 199.35], [266.25, 192.59], [340.08, 192.61]]}, {\"category\": \"car\", \"corners_3d\": [[490.35, 222.57], [452.59, 222.39], [475.55, 216.89], [509.18, 217.04], [490.35, 184.14], [452.59, 184.09], [475.55, 182.85], [509.18, 182.88]]}, {\"category\": \"car\", \"corners_3d\": [[553.72, 206.94], [527.57, 206.81], [539.66, 204.14], [563.79, 204.25], [553.72, 181.76], [527.57, 181.73], [539.66, 181.03], [563.79, 181.06]]}, {\"category\": \"car\", \"corners_3d\": [[295.52, 243.25], [241.86, 243.04], [311.21, 230.75], [355.62, 230.89], [295.52, 195.09], [241.86, 195.03], [311.21, 191.14], [355.62, 191.19]]}, {\"category\": \"car\", \"corners_3d\": [[368.31, 228.54], [328.0, 228.48], [364.05, 221.69], [399.47, 221.74], [368.31, 182.3], [328.0, 182.29], [364.05, 181.14], [399.47, 181.14]]}, {\"category\": \"car\", \"corners_3d\": [[600.93, 196.89], [581.24, 196.78], [590.92, 195.25], [609.39, 195.35], [600.93, 181.05], [581.24, 181.01], [590.92, 180.49], [609.39, 180.52]]}, {\"category\": \"car\", \"corners_3d\": [[440.37, 213.76], [411.18, 213.56], [438.86, 209.6], [465.33, 209.77], [440.37, 184.48], [411.18, 184.43], [438.86, 183.3], [465.33, 183.35]]}, {\"category\": \"car\", \"corners_3d\": [[508.47, 205.29], [486.94, 204.99], [513.21, 202.49], [533.28, 202.74], [508.47, 185.17], [486.94, 185.06], [513.21, 184.11], [533.28, 184.21]]}, {\"category\": \"car\", \"corners_3d\": [[117.96, 240.32], [71.65, 239.76], [153.83, 230.95], [194.6, 231.37], [117.96, 189.45], [71.65, 189.31], [153.83, 187.15], [194.6, 187.25]]}, {\"category\": \"car\", \"corners_3d\": [[118.41, 221.65], [87.67, 220.8], [175.83, 215.18], [204.29, 215.84], [118.41, 186.99], [87.67, 186.75], [175.83, 185.12], [204.29, 185.31]]}, {\"category\": \"car\", \"corners_3d\": [[655.64, 193.39], [673.86, 193.5], [668.29, 194.68], [649.06, 194.55], [655.64, 178.86], [673.86, 178.89], [668.29, 179.24], [649.06, 179.2]]}]\n```" - } - ] - }, - { - "id": 173, - "data_source": "Mono3DRefer", - "scene_name": "000096", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1053.42, 297.61], [1177.21, 324.12], [859.2, 339.23], [785.62, 307.71], [1053.42, 185.54], [1177.21, 188.24], [859.2, 189.78], [785.62, 186.57]]}, {\"category\": \"car\", \"corners_3d\": [[915.71, 270.44], [1002.73, 288.91], [705.64, 300.43], [662.62, 278.47], [915.71, 164.45], [1002.73, 162.86], [705.64, 161.87], [662.62, 163.76]]}, {\"category\": \"car\", \"corners_3d\": [[556.71, 219.52], [518.21, 220.14], [506.21, 214.07], [539.96, 213.6], [556.71, 186.56], [518.21, 186.74], [506.21, 184.96], [539.96, 184.82]]}, {\"category\": \"car\", \"corners_3d\": [[835.02, 258.32], [879.59, 268.17], [664.19, 275.12], [640.43, 263.87], [835.02, 187.51], [879.59, 189.19], [664.19, 190.38], [640.43, 188.46]]}]\n```" - } - ] - }, - { - "id": 174, - "data_source": "Mono3DRefer", - "scene_name": "000096", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[346.12, 242.15], [280.8, 243.23], [314.22, 231.28], [368.17, 230.54], [346.12, 179.46], [280.8, 179.56], [314.22, 178.42], [368.17, 178.35]]}]\n```" - } - ] - }, - { - "id": 175, - "data_source": "Mono3DRefer", - "scene_name": "000096", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1053.42, 297.61], [1177.21, 324.12], [859.2, 339.23], [785.62, 307.71], [1053.42, 185.54], [1177.21, 188.24], [859.2, 189.78], [785.62, 186.57]]}, {\"category\": \"car\", \"corners_3d\": [[915.71, 270.44], [1002.73, 288.91], [705.64, 300.43], [662.62, 278.47], [915.71, 164.45], [1002.73, 162.86], [705.64, 161.87], [662.62, 163.76]]}, {\"category\": \"van\", \"corners_3d\": [[346.12, 242.15], [280.8, 243.23], [314.22, 231.28], [368.17, 230.54], [346.12, 179.46], [280.8, 179.56], [314.22, 178.42], [368.17, 178.35]]}, {\"category\": \"car\", \"corners_3d\": [[556.71, 219.52], [518.21, 220.14], [506.21, 214.07], [539.96, 213.6], [556.71, 186.56], [518.21, 186.74], [506.21, 184.96], [539.96, 184.82]]}, {\"category\": \"car\", \"corners_3d\": [[835.02, 258.32], [879.59, 268.17], [664.19, 275.12], [640.43, 263.87], [835.02, 187.51], [879.59, 189.19], [664.19, 190.38], [640.43, 188.46]]}]\n```" - } - ] - }, - { - "id": 176, - "data_source": "Mono3DRefer", - "scene_name": "000097", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[555.85, 195.05], [536.26, 194.93], [551.39, 193.4], [569.7, 193.51], [555.85, 176.44], [536.26, 176.42], [551.39, 176.18], [569.7, 176.19]]}, {\"category\": \"car\", \"corners_3d\": [[718.48, 192.33], [738.39, 192.42], [738.73, 193.92], [717.29, 193.82], [718.48, 174.7], [738.39, 174.71], [738.73, 174.85], [717.29, 174.84]]}]\n```" - } - ] - }, - { - "id": 177, - "data_source": "Mono3DRefer", - "scene_name": "000100", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[409.17, 281.54], [383.44, 285.96], [355.84, 282.76], [381.92, 278.59], [409.17, 182.31], [383.44, 182.7], [355.84, 182.42], [381.92, 182.05]]}]\n```" - } - ] - }, - { - "id": 178, - "data_source": "Mono3DRefer", - "scene_name": "000100", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[685.06, 235.77], [707.62, 235.97], [710.37, 242.91], [685.33, 242.67], [685.06, 166.77], [707.62, 166.76], [710.37, 166.08], [685.33, 166.11]]}]\n```" - } - ] - }, - { - "id": 179, - "data_source": "Mono3DRefer", - "scene_name": "000100", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[19.86, 237.31], [72.86, 232.1], [206.33, 233.18], [164.25, 238.59], [19.86, 180.7], [72.86, 180.07], [206.33, 180.2], [164.25, 180.86]]}]\n```" - } - ] - }, - { - "id": 180, - "data_source": "Mono3DRefer", - "scene_name": "000100", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[177.65, 295.08], [320.65, 296.61], [-73.36, 443.11], [-370.65, 435.92], [177.65, 117.13], [320.65, 116.43], [-73.36, 49.64], [-370.65, 52.91]]}]\n```" - } - ] - }, - { - "id": 181, - "data_source": "Mono3DRefer", - "scene_name": "000100", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[685.06, 235.77], [707.62, 235.97], [710.37, 242.91], [685.33, 242.67], [685.06, 166.77], [707.62, 166.76], [710.37, 166.08], [685.33, 166.11]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[409.17, 281.54], [383.44, 285.96], [355.84, 282.76], [381.92, 278.59], [409.17, 182.31], [383.44, 182.7], [355.84, 182.42], [381.92, 182.05]]}]\n```" - } - ] - }, - { - "id": 182, - "data_source": "Mono3DRefer", - "scene_name": "000100", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[409.17, 281.54], [383.44, 285.96], [355.84, 282.76], [381.92, 278.59], [409.17, 182.31], [383.44, 182.7], [355.84, 182.42], [381.92, 182.05]]}, {\"category\": \"car\", \"corners_3d\": [[19.86, 237.31], [72.86, 232.1], [206.33, 233.18], [164.25, 238.59], [19.86, 180.7], [72.86, 180.07], [206.33, 180.2], [164.25, 180.86]]}]\n```" - } - ] - }, - { - "id": 183, - "data_source": "Mono3DRefer", - "scene_name": "000100", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[177.65, 295.08], [320.65, 296.61], [-73.36, 443.11], [-370.65, 435.92], [177.65, 117.13], [320.65, 116.43], [-73.36, 49.64], [-370.65, 52.91]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[409.17, 281.54], [383.44, 285.96], [355.84, 282.76], [381.92, 278.59], [409.17, 182.31], [383.44, 182.7], [355.84, 182.42], [381.92, 182.05]]}]\n```" - } - ] - }, - { - "id": 184, - "data_source": "Mono3DRefer", - "scene_name": "000100", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[685.06, 235.77], [707.62, 235.97], [710.37, 242.91], [685.33, 242.67], [685.06, 166.77], [707.62, 166.76], [710.37, 166.08], [685.33, 166.11]]}, {\"category\": \"car\", \"corners_3d\": [[19.86, 237.31], [72.86, 232.1], [206.33, 233.18], [164.25, 238.59], [19.86, 180.7], [72.86, 180.07], [206.33, 180.2], [164.25, 180.86]]}]\n```" - } - ] - }, - { - "id": 185, - "data_source": "Mono3DRefer", - "scene_name": "000100", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[177.65, 295.08], [320.65, 296.61], [-73.36, 443.11], [-370.65, 435.92], [177.65, 117.13], [320.65, 116.43], [-73.36, 49.64], [-370.65, 52.91]]}, {\"category\": \"cyclist\", \"corners_3d\": [[685.06, 235.77], [707.62, 235.97], [710.37, 242.91], [685.33, 242.67], [685.06, 166.77], [707.62, 166.76], [710.37, 166.08], [685.33, 166.11]]}]\n```" - } - ] - }, - { - "id": 186, - "data_source": "Mono3DRefer", - "scene_name": "000100", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[177.65, 295.08], [320.65, 296.61], [-73.36, 443.11], [-370.65, 435.92], [177.65, 117.13], [320.65, 116.43], [-73.36, 49.64], [-370.65, 52.91]]}, {\"category\": \"car\", \"corners_3d\": [[19.86, 237.31], [72.86, 232.1], [206.33, 233.18], [164.25, 238.59], [19.86, 180.7], [72.86, 180.07], [206.33, 180.2], [164.25, 180.86]]}]\n```" - } - ] - }, - { - "id": 187, - "data_source": "Mono3DRefer", - "scene_name": "000100", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[685.06, 235.77], [707.62, 235.97], [710.37, 242.91], [685.33, 242.67], [685.06, 166.77], [707.62, 166.76], [710.37, 166.08], [685.33, 166.11]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[409.17, 281.54], [383.44, 285.96], [355.84, 282.76], [381.92, 278.59], [409.17, 182.31], [383.44, 182.7], [355.84, 182.42], [381.92, 182.05]]}, {\"category\": \"car\", \"corners_3d\": [[19.86, 237.31], [72.86, 232.1], [206.33, 233.18], [164.25, 238.59], [19.86, 180.7], [72.86, 180.07], [206.33, 180.2], [164.25, 180.86]]}]\n```" - } - ] - }, - { - "id": 188, - "data_source": "Mono3DRefer", - "scene_name": "000100", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[177.65, 295.08], [320.65, 296.61], [-73.36, 443.11], [-370.65, 435.92], [177.65, 117.13], [320.65, 116.43], [-73.36, 49.64], [-370.65, 52.91]]}, {\"category\": \"cyclist\", \"corners_3d\": [[685.06, 235.77], [707.62, 235.97], [710.37, 242.91], [685.33, 242.67], [685.06, 166.77], [707.62, 166.76], [710.37, 166.08], [685.33, 166.11]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[409.17, 281.54], [383.44, 285.96], [355.84, 282.76], [381.92, 278.59], [409.17, 182.31], [383.44, 182.7], [355.84, 182.42], [381.92, 182.05]]}]\n```" - } - ] - }, - { - "id": 189, - "data_source": "Mono3DRefer", - "scene_name": "000100", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[177.65, 295.08], [320.65, 296.61], [-73.36, 443.11], [-370.65, 435.92], [177.65, 117.13], [320.65, 116.43], [-73.36, 49.64], [-370.65, 52.91]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[409.17, 281.54], [383.44, 285.96], [355.84, 282.76], [381.92, 278.59], [409.17, 182.31], [383.44, 182.7], [355.84, 182.42], [381.92, 182.05]]}, {\"category\": \"car\", \"corners_3d\": [[19.86, 237.31], [72.86, 232.1], [206.33, 233.18], [164.25, 238.59], [19.86, 180.7], [72.86, 180.07], [206.33, 180.2], [164.25, 180.86]]}]\n```" - } - ] - }, - { - "id": 190, - "data_source": "Mono3DRefer", - "scene_name": "000100", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[177.65, 295.08], [320.65, 296.61], [-73.36, 443.11], [-370.65, 435.92], [177.65, 117.13], [320.65, 116.43], [-73.36, 49.64], [-370.65, 52.91]]}, {\"category\": \"cyclist\", \"corners_3d\": [[685.06, 235.77], [707.62, 235.97], [710.37, 242.91], [685.33, 242.67], [685.06, 166.77], [707.62, 166.76], [710.37, 166.08], [685.33, 166.11]]}, {\"category\": \"car\", \"corners_3d\": [[19.86, 237.31], [72.86, 232.1], [206.33, 233.18], [164.25, 238.59], [19.86, 180.7], [72.86, 180.07], [206.33, 180.2], [164.25, 180.86]]}]\n```" - } - ] - }, - { - "id": 191, - "data_source": "Mono3DRefer", - "scene_name": "000100", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[177.65, 295.08], [320.65, 296.61], [-73.36, 443.11], [-370.65, 435.92], [177.65, 117.13], [320.65, 116.43], [-73.36, 49.64], [-370.65, 52.91]]}, {\"category\": \"cyclist\", \"corners_3d\": [[685.06, 235.77], [707.62, 235.97], [710.37, 242.91], [685.33, 242.67], [685.06, 166.77], [707.62, 166.76], [710.37, 166.08], [685.33, 166.11]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[409.17, 281.54], [383.44, 285.96], [355.84, 282.76], [381.92, 278.59], [409.17, 182.31], [383.44, 182.7], [355.84, 182.42], [381.92, 182.05]]}, {\"category\": \"car\", \"corners_3d\": [[19.86, 237.31], [72.86, 232.1], [206.33, 233.18], [164.25, 238.59], [19.86, 180.7], [72.86, 180.07], [206.33, 180.2], [164.25, 180.86]]}]\n```" - } - ] - }, - { - "id": 192, - "data_source": "Mono3DRefer", - "scene_name": "000101", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[576.31, 230.2], [631.96, 230.2], [636.03, 240.89], [570.01, 240.89], [576.31, 177.28], [631.96, 177.28], [636.03, 178.11], [570.01, 178.11]]}, {\"category\": \"car\", \"corners_3d\": [[16.89, 265.94], [-44.13, 266.03], [57.27, 251.4], [108.64, 251.34], [16.89, 220.21], [-44.13, 220.25], [57.27, 212.81], [108.64, 212.78]]}, {\"category\": \"car\", \"corners_3d\": [[164.28, 253.6], [106.28, 253.8], [169.83, 243.11], [220.06, 242.96], [164.28, 203.86], [106.28, 203.93], [169.83, 199.83], [220.06, 199.77]]}]\n```" - } - ] - }, - { - "id": 193, - "data_source": "Mono3DRefer", - "scene_name": "000104", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[830.25, 228.81], [882.56, 230.26], [879.96, 239.67], [819.4, 237.72], [830.25, 180.46], [882.56, 180.66], [879.96, 181.94], [819.4, 181.67]]}, {\"category\": \"car\", \"corners_3d\": [[431.73, 263.08], [362.91, 259.52], [477.34, 240.36], [534.09, 242.5], [431.73, 184.12], [362.91, 183.67], [477.34, 181.28], [534.09, 181.55]]}, {\"category\": \"car\", \"corners_3d\": [[653.95, 222.9], [610.67, 221.8], [646.94, 215.22], [685.0, 216.04], [653.95, 184.4], [610.67, 184.14], [646.94, 182.62], [685.0, 182.81]]}, {\"category\": \"car\", \"corners_3d\": [[716.13, 213.43], [683.28, 212.83], [699.11, 209.17], [729.13, 209.66], [716.13, 183.41], [683.28, 183.25], [699.11, 182.3], [729.13, 182.43]]}, {\"category\": \"car\", \"corners_3d\": [[738.8, 208.09], [714.21, 207.68], [727.47, 204.76], [750.12, 205.1], [738.8, 182.04], [714.21, 181.93], [727.47, 181.17], [750.12, 181.26]]}, {\"category\": \"car\", \"corners_3d\": [[762.46, 204.18], [740.07, 203.85], [747.81, 202.03], [768.95, 202.32], [762.46, 181.54], [740.07, 181.44], [747.81, 180.94], [768.95, 181.02]]}, {\"category\": \"car\", \"corners_3d\": [[859.92, 208.95], [886.87, 209.43], [885.07, 212.73], [855.76, 212.15], [859.92, 185.51], [886.87, 185.68], [885.07, 186.83], [855.76, 186.63]]}]\n```" - } - ] - }, - { - "id": 194, - "data_source": "Mono3DRefer", - "scene_name": "000106", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.36, 202.05], [628.49, 202.04], [630.59, 204.44], [603.4, 204.45], [603.36, 177.21], [628.49, 177.21], [630.59, 177.57], [603.4, 177.57]]}, {\"category\": \"car\", \"corners_3d\": [[70.71, 292.67], [-17.39, 292.54], [138.34, 263.12], [204.9, 263.2], [70.71, 211.76], [-17.39, 211.72], [138.34, 202.17], [204.9, 202.19]]}, {\"category\": \"car\", \"corners_3d\": [[218.12, 245.12], [167.36, 245.02], [237.43, 233.93], [280.47, 234.0], [218.12, 189.28], [167.36, 189.26], [237.43, 186.74], [280.47, 186.76]]}, {\"category\": \"car\", \"corners_3d\": [[471.64, 205.09], [449.4, 205.08], [462.97, 202.56], [483.49, 202.58], [471.64, 184.45], [449.4, 184.44], [462.97, 183.54], [483.49, 183.54]]}]\n```" - } - ] - }, - { - "id": 195, - "data_source": "Mono3DRefer", - "scene_name": "000107", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[813.86, 348.65], [989.96, 348.25], [1692.25, 666.51], [1199.06, 669.65], [813.86, 183.02], [989.96, 183.0], [1692.25, 201.41], [1199.06, 201.59]]}, {\"category\": \"car\", \"corners_3d\": [[734.8, 271.67], [826.58, 271.55], [933.81, 318.87], [798.13, 319.12], [734.8, 182.1], [826.58, 182.09], [933.81, 186.51], [798.13, 186.54]]}, {\"category\": \"car\", \"corners_3d\": [[684.8, 239.35], [746.61, 239.24], [792.11, 259.25], [711.72, 259.44], [684.8, 178.52], [746.61, 178.51], [792.11, 180.21], [711.72, 180.23]]}, {\"category\": \"car\", \"corners_3d\": [[491.54, 224.26], [539.51, 224.23], [527.67, 233.82], [470.73, 233.86], [491.54, 178.36], [539.51, 178.36], [527.67, 179.39], [470.73, 179.39]]}, {\"category\": \"car\", \"corners_3d\": [[672.39, 219.88], [716.22, 219.94], [733.41, 228.77], [681.35, 228.69], [672.39, 178.97], [716.22, 178.98], [733.41, 180.12], [681.35, 180.11]]}, {\"category\": \"car\", \"corners_3d\": [[650.76, 210.42], [684.72, 210.44], [692.63, 214.85], [654.68, 214.83], [650.76, 179.29], [684.72, 179.29], [692.63, 180.05], [654.68, 180.04]]}, {\"category\": \"car\", \"corners_3d\": [[635.25, 200.96], [661.95, 200.97], [666.76, 204.01], [637.19, 203.99], [635.25, 177.64], [661.95, 177.64], [666.76, 178.16], [637.19, 178.16]]}]\n```" - } - ] - }, - { - "id": 196, - "data_source": "Mono3DRefer", - "scene_name": "000108", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[577.66, 253.55], [660.78, 253.56], [673.69, 274.13], [569.39, 274.11], [577.66, 173.8], [660.78, 173.8], [673.69, 174.04], [569.39, 174.04]]}]\n```" - } - ] - }, - { - "id": 197, - "data_source": "Mono3DRefer", - "scene_name": "000108", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[864.23, 238.44], [889.05, 238.4], [933.16, 248.25], [904.63, 248.31], [864.23, 152.54], [889.05, 152.56], [933.16, 149.5], [904.63, 149.49]]}]\n```" - } - ] - }, - { - "id": 198, - "data_source": "Mono3DRefer", - "scene_name": "000108", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[386.13, 242.4], [493.97, 242.3], [290.21, 372.19], [-20.9, 372.98], [386.13, 91.49], [493.97, 91.6], [290.21, -60.35], [-20.9, -61.28]]}]\n```" - } - ] - }, - { - "id": 199, - "data_source": "Mono3DRefer", - "scene_name": "000108", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[577.66, 253.55], [660.78, 253.56], [673.69, 274.13], [569.39, 274.11], [577.66, 173.8], [660.78, 173.8], [673.69, 174.04], [569.39, 174.04]]}, {\"category\": \"cyclist\", \"corners_3d\": [[864.23, 238.44], [889.05, 238.4], [933.16, 248.25], [904.63, 248.31], [864.23, 152.54], [889.05, 152.56], [933.16, 149.5], [904.63, 149.49]]}]\n```" - } - ] - }, - { - "id": 200, - "data_source": "Mono3DRefer", - "scene_name": "000108", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[577.66, 253.55], [660.78, 253.56], [673.69, 274.13], [569.39, 274.11], [577.66, 173.8], [660.78, 173.8], [673.69, 174.04], [569.39, 174.04]]}, {\"category\": \"tram\", \"corners_3d\": [[386.13, 242.4], [493.97, 242.3], [290.21, 372.19], [-20.9, 372.98], [386.13, 91.49], [493.97, 91.6], [290.21, -60.35], [-20.9, -61.28]]}]\n```" - } - ] - }, - { - "id": 201, - "data_source": "Mono3DRefer", - "scene_name": "000108", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[864.23, 238.44], [889.05, 238.4], [933.16, 248.25], [904.63, 248.31], [864.23, 152.54], [889.05, 152.56], [933.16, 149.5], [904.63, 149.49]]}, {\"category\": \"tram\", \"corners_3d\": [[386.13, 242.4], [493.97, 242.3], [290.21, 372.19], [-20.9, 372.98], [386.13, 91.49], [493.97, 91.6], [290.21, -60.35], [-20.9, -61.28]]}]\n```" - } - ] - }, - { - "id": 202, - "data_source": "Mono3DRefer", - "scene_name": "000108", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[577.66, 253.55], [660.78, 253.56], [673.69, 274.13], [569.39, 274.11], [577.66, 173.8], [660.78, 173.8], [673.69, 174.04], [569.39, 174.04]]}, {\"category\": \"cyclist\", \"corners_3d\": [[864.23, 238.44], [889.05, 238.4], [933.16, 248.25], [904.63, 248.31], [864.23, 152.54], [889.05, 152.56], [933.16, 149.5], [904.63, 149.49]]}, {\"category\": \"tram\", \"corners_3d\": [[386.13, 242.4], [493.97, 242.3], [290.21, 372.19], [-20.9, 372.98], [386.13, 91.49], [493.97, 91.6], [290.21, -60.35], [-20.9, -61.28]]}]\n```" - } - ] - }, - { - "id": 203, - "data_source": "Mono3DRefer", - "scene_name": "000110", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1084.97, 177.33], [1120.18, 177.35], [1173.16, 177.91], [1133.24, 177.88], [1084.97, 137.94], [1120.18, 137.75], [1173.16, 133.36], [1133.24, 133.6]]}]\n```" - } - ] - }, - { - "id": 204, - "data_source": "Mono3DRefer", - "scene_name": "000110", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[939.56, 303.89], [865.05, 278.31], [1223.02, 274.46], [1377.09, 298.0], [939.56, 129.46], [865.05, 137.93], [1223.02, 139.21], [1377.09, 131.42]]}, {\"category\": \"car\", \"corners_3d\": [[1308.3, 186.46], [1356.63, 187.37], [1226.34, 187.45], [1185.99, 186.53], [1308.3, 143.35], [1356.63, 141.37], [1226.34, 141.19], [1185.99, 143.19]]}, {\"category\": \"car\", \"corners_3d\": [[429.46, 248.7], [346.78, 249.64], [384.99, 233.57], [450.16, 232.98], [429.46, 176.93], [346.78, 176.98], [384.99, 176.11], [450.16, 176.08]]}, {\"category\": \"car\", \"corners_3d\": [[517.21, 256.94], [421.03, 256.84], [467.88, 236.68], [541.0, 236.74], [517.21, 172.32], [421.03, 172.32], [467.88, 172.45], [541.0, 172.45]]}, {\"category\": \"car\", \"corners_3d\": [[940.41, 196.44], [962.79, 198.07], [830.08, 198.03], [816.31, 196.4], [940.41, 146.63], [962.79, 144.81], [830.08, 144.86], [816.31, 146.67]]}, {\"category\": \"car\", \"corners_3d\": [[966.79, 182.21], [998.0, 182.2], [1039.71, 183.17], [1005.3, 183.18], [966.79, 147.81], [998.0, 147.83], [1039.71, 145.24], [1005.3, 145.21]]}, {\"category\": \"car\", \"corners_3d\": [[1005.59, 178.48], [1037.15, 178.48], [1087.12, 179.11], [1052.03, 179.12], [1005.59, 147.68], [1037.15, 147.7], [1087.12, 144.85], [1052.03, 144.82]]}, {\"category\": \"car\", \"corners_3d\": [[464.98, 230.83], [408.33, 230.92], [440.1, 221.01], [487.06, 220.95], [464.98, 179.76], [408.33, 179.77], [440.1, 178.59], [487.06, 178.58]]}, {\"category\": \"car\", \"corners_3d\": [[702.54, 198.92], [666.07, 198.95], [657.75, 195.91], [690.0, 195.89], [702.54, 166.86], [666.07, 166.86], [657.75, 167.55], [690.0, 167.56]]}, {\"category\": \"car\", \"corners_3d\": [[607.03, 195.4], [631.98, 195.49], [626.63, 197.35], [599.66, 197.24], [607.03, 172.26], [631.98, 172.26], [626.63, 172.21], [599.66, 172.21]]}, {\"category\": \"car\", \"corners_3d\": [[579.06, 180.81], [591.55, 180.89], [570.83, 181.29], [557.93, 181.21], [579.06, 166.29], [591.55, 166.23], [570.83, 165.89], [557.93, 165.96]]}]\n```" - } - ] - }, - { - "id": 205, - "data_source": "Mono3DRefer", - "scene_name": "000110", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[939.56, 303.89], [865.05, 278.31], [1223.02, 274.46], [1377.09, 298.0], [939.56, 129.46], [865.05, 137.93], [1223.02, 139.21], [1377.09, 131.42]]}, {\"category\": \"car\", \"corners_3d\": [[1308.3, 186.46], [1356.63, 187.37], [1226.34, 187.45], [1185.99, 186.53], [1308.3, 143.35], [1356.63, 141.37], [1226.34, 141.19], [1185.99, 143.19]]}, {\"category\": \"car\", \"corners_3d\": [[429.46, 248.7], [346.78, 249.64], [384.99, 233.57], [450.16, 232.98], [429.46, 176.93], [346.78, 176.98], [384.99, 176.11], [450.16, 176.08]]}, {\"category\": \"car\", \"corners_3d\": [[517.21, 256.94], [421.03, 256.84], [467.88, 236.68], [541.0, 236.74], [517.21, 172.32], [421.03, 172.32], [467.88, 172.45], [541.0, 172.45]]}, {\"category\": \"car\", \"corners_3d\": [[940.41, 196.44], [962.79, 198.07], [830.08, 198.03], [816.31, 196.4], [940.41, 146.63], [962.79, 144.81], [830.08, 144.86], [816.31, 146.67]]}, {\"category\": \"car\", \"corners_3d\": [[966.79, 182.21], [998.0, 182.2], [1039.71, 183.17], [1005.3, 183.18], [966.79, 147.81], [998.0, 147.83], [1039.71, 145.24], [1005.3, 145.21]]}, {\"category\": \"van\", \"corners_3d\": [[1084.97, 177.33], [1120.18, 177.35], [1173.16, 177.91], [1133.24, 177.88], [1084.97, 137.94], [1120.18, 137.75], [1173.16, 133.36], [1133.24, 133.6]]}, {\"category\": \"car\", \"corners_3d\": [[1005.59, 178.48], [1037.15, 178.48], [1087.12, 179.11], [1052.03, 179.12], [1005.59, 147.68], [1037.15, 147.7], [1087.12, 144.85], [1052.03, 144.82]]}, {\"category\": \"car\", \"corners_3d\": [[464.98, 230.83], [408.33, 230.92], [440.1, 221.01], [487.06, 220.95], [464.98, 179.76], [408.33, 179.77], [440.1, 178.59], [487.06, 178.58]]}, {\"category\": \"car\", \"corners_3d\": [[702.54, 198.92], [666.07, 198.95], [657.75, 195.91], [690.0, 195.89], [702.54, 166.86], [666.07, 166.86], [657.75, 167.55], [690.0, 167.56]]}, {\"category\": \"car\", \"corners_3d\": [[607.03, 195.4], [631.98, 195.49], [626.63, 197.35], [599.66, 197.24], [607.03, 172.26], [631.98, 172.26], [626.63, 172.21], [599.66, 172.21]]}, {\"category\": \"car\", \"corners_3d\": [[579.06, 180.81], [591.55, 180.89], [570.83, 181.29], [557.93, 181.21], [579.06, 166.29], [591.55, 166.23], [570.83, 165.89], [557.93, 165.96]]}]\n```" - } - ] - }, - { - "id": 206, - "data_source": "Mono3DRefer", - "scene_name": "000111", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[392.37, 199.83], [372.96, 199.84], [386.18, 198.28], [404.46, 198.27], [392.37, 177.8], [372.96, 177.8], [386.18, 177.51], [404.46, 177.51]]}, {\"category\": \"car\", \"corners_3d\": [[175.08, 203.42], [144.24, 204.15], [135.46, 202.24], [164.64, 201.59], [175.08, 179.09], [144.24, 179.24], [135.46, 178.85], [164.64, 178.72]]}, {\"category\": \"car\", \"corners_3d\": [[593.83, 214.57], [635.54, 214.55], [640.38, 220.74], [592.48, 220.77], [593.83, 175.12], [635.54, 175.12], [640.38, 175.46], [592.48, 175.46]]}, {\"category\": \"car\", \"corners_3d\": [[433.6, 225.72], [386.68, 225.79], [414.52, 218.7], [455.13, 218.65], [433.6, 186.77], [386.68, 186.79], [414.52, 184.93], [455.13, 184.91]]}, {\"category\": \"car\", \"corners_3d\": [[851.23, 234.84], [918.39, 239.6], [810.69, 253.59], [739.8, 246.72], [851.23, 170.69], [918.39, 170.52], [810.69, 170.03], [739.8, 170.27]]}, {\"category\": \"car\", \"corners_3d\": [[764.24, 218.08], [809.39, 220.43], [731.8, 227.4], [684.77, 224.33], [764.24, 175.0], [809.39, 175.11], [731.8, 175.44], [684.77, 175.29]]}, {\"category\": \"car\", \"corners_3d\": [[736.89, 207.4], [770.26, 208.77], [716.9, 212.15], [682.75, 210.51], [736.89, 170.06], [770.26, 169.95], [716.9, 169.68], [682.75, 169.81]]}, {\"category\": \"car\", \"corners_3d\": [[691.42, 197.24], [712.32, 197.88], [667.81, 199.66], [646.68, 198.92], [691.42, 173.49], [712.32, 173.5], [667.81, 173.55], [646.68, 173.53]]}, {\"category\": \"car\", \"corners_3d\": [[424.28, 206.52], [394.04, 206.5], [417.25, 203.09], [444.44, 203.11], [424.28, 176.29], [394.04, 176.28], [417.25, 175.94], [444.44, 175.94]]}, {\"category\": \"car\", \"corners_3d\": [[450.47, 201.12], [426.12, 201.09], [443.06, 198.75], [465.4, 198.78], [450.47, 175.8], [426.12, 175.8], [443.06, 175.55], [465.4, 175.56]]}, {\"category\": \"car\", \"corners_3d\": [[507.39, 192.41], [490.38, 192.39], [500.04, 191.17], [516.0, 191.19], [507.39, 177.32], [490.38, 177.32], [500.04, 177.04], [516.0, 177.04]]}]\n```" - } - ] - }, - { - "id": 207, - "data_source": "Mono3DRefer", - "scene_name": "000111", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[707.22, 200.81], [736.48, 201.73], [682.48, 204.66], [652.26, 203.54], [707.22, 165.53], [736.48, 165.29], [682.48, 164.52], [652.26, 164.81]]}, {\"category\": \"van\", \"corners_3d\": [[556.69, 188.77], [536.42, 188.73], [547.75, 187.42], [566.37, 187.46], [556.69, 161.36], [536.42, 161.39], [547.75, 162.33], [566.37, 162.31]]}]\n```" - } - ] - }, - { - "id": 208, - "data_source": "Mono3DRefer", - "scene_name": "000111", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[392.37, 199.83], [372.96, 199.84], [386.18, 198.28], [404.46, 198.27], [392.37, 177.8], [372.96, 177.8], [386.18, 177.51], [404.46, 177.51]]}, {\"category\": \"car\", \"corners_3d\": [[175.08, 203.42], [144.24, 204.15], [135.46, 202.24], [164.64, 201.59], [175.08, 179.09], [144.24, 179.24], [135.46, 178.85], [164.64, 178.72]]}, {\"category\": \"car\", \"corners_3d\": [[593.83, 214.57], [635.54, 214.55], [640.38, 220.74], [592.48, 220.77], [593.83, 175.12], [635.54, 175.12], [640.38, 175.46], [592.48, 175.46]]}, {\"category\": \"car\", \"corners_3d\": [[433.6, 225.72], [386.68, 225.79], [414.52, 218.7], [455.13, 218.65], [433.6, 186.77], [386.68, 186.79], [414.52, 184.93], [455.13, 184.91]]}, {\"category\": \"car\", \"corners_3d\": [[851.23, 234.84], [918.39, 239.6], [810.69, 253.59], [739.8, 246.72], [851.23, 170.69], [918.39, 170.52], [810.69, 170.03], [739.8, 170.27]]}, {\"category\": \"car\", \"corners_3d\": [[764.24, 218.08], [809.39, 220.43], [731.8, 227.4], [684.77, 224.33], [764.24, 175.0], [809.39, 175.11], [731.8, 175.44], [684.77, 175.29]]}, {\"category\": \"car\", \"corners_3d\": [[736.89, 207.4], [770.26, 208.77], [716.9, 212.15], [682.75, 210.51], [736.89, 170.06], [770.26, 169.95], [716.9, 169.68], [682.75, 169.81]]}, {\"category\": \"van\", \"corners_3d\": [[707.22, 200.81], [736.48, 201.73], [682.48, 204.66], [652.26, 203.54], [707.22, 165.53], [736.48, 165.29], [682.48, 164.52], [652.26, 164.81]]}, {\"category\": \"car\", \"corners_3d\": [[691.42, 197.24], [712.32, 197.88], [667.81, 199.66], [646.68, 198.92], [691.42, 173.49], [712.32, 173.5], [667.81, 173.55], [646.68, 173.53]]}, {\"category\": \"car\", \"corners_3d\": [[424.28, 206.52], [394.04, 206.5], [417.25, 203.09], [444.44, 203.11], [424.28, 176.29], [394.04, 176.28], [417.25, 175.94], [444.44, 175.94]]}, {\"category\": \"van\", \"corners_3d\": [[556.69, 188.77], [536.42, 188.73], [547.75, 187.42], [566.37, 187.46], [556.69, 161.36], [536.42, 161.39], [547.75, 162.33], [566.37, 162.31]]}, {\"category\": \"car\", \"corners_3d\": [[450.47, 201.12], [426.12, 201.09], [443.06, 198.75], [465.4, 198.78], [450.47, 175.8], [426.12, 175.8], [443.06, 175.55], [465.4, 175.56]]}, {\"category\": \"car\", \"corners_3d\": [[507.39, 192.41], [490.38, 192.39], [500.04, 191.17], [516.0, 191.19], [507.39, 177.32], [490.38, 177.32], [500.04, 177.04], [516.0, 177.04]]}]\n```" - } - ] - }, - { - "id": 209, - "data_source": "Mono3DRefer", - "scene_name": "000113", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[534.25, 215.52], [580.25, 215.31], [585.56, 223.36], [530.81, 223.67], [534.25, 164.92], [580.25, 164.96], [585.56, 163.47], [530.81, 163.41]]}, {\"category\": \"van\", \"corners_3d\": [[1036.34, 216.54], [1087.12, 220.98], [866.49, 222.56], [835.18, 217.84], [1036.34, 132.92], [1087.12, 128.86], [866.49, 127.41], [835.18, 131.73]]}, {\"category\": \"van\", \"corners_3d\": [[646.26, 187.81], [618.15, 187.86], [610.43, 186.3], [635.65, 186.26], [646.26, 149.72], [618.15, 149.64], [610.43, 152.05], [635.65, 152.12]]}, {\"category\": \"van\", \"corners_3d\": [[1296.66, 203.37], [1237.97, 203.56], [1120.38, 198.25], [1169.57, 198.12], [1296.66, 133.16], [1237.97, 132.91], [1120.38, 139.82], [1169.57, 139.99]]}]\n```" - } - ] - }, - { - "id": 210, - "data_source": "Mono3DRefer", - "scene_name": "000113", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[602.84, 190.72], [569.02, 190.79], [566.47, 188.23], [595.51, 188.18], [602.84, 137.11], [569.02, 136.98], [566.47, 142.08], [595.51, 142.18]]}]\n```" - } - ] - }, - { - "id": 211, - "data_source": "Mono3DRefer", - "scene_name": "000113", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[926.16, 289.18], [1089.12, 297.24], [1287.93, 407.8], [973.87, 380.6], [926.16, 156.1], [1089.12, 154.93], [1287.93, 139.01], [973.87, 142.93]]}, {\"category\": \"car\", \"corners_3d\": [[356.02, 220.65], [401.41, 220.39], [372.86, 229.84], [318.19, 230.21], [356.02, 182.2], [401.41, 182.15], [372.86, 183.99], [318.19, 184.07]]}, {\"category\": \"car\", \"corners_3d\": [[274.76, 221.56], [318.8, 221.41], [267.97, 231.09], [214.93, 231.29], [274.76, 168.42], [318.8, 168.43], [267.97, 167.55], [214.93, 167.53]]}, {\"category\": \"car\", \"corners_3d\": [[464.03, 201.91], [498.96, 201.75], [495.85, 205.46], [456.39, 205.67], [464.03, 172.0], [498.96, 172.01], [495.85, 171.9], [456.39, 171.89]]}, {\"category\": \"car\", \"corners_3d\": [[693.02, 209.5], [733.62, 209.25], [770.56, 215.64], [723.08, 215.98], [693.02, 172.85], [733.62, 172.85], [770.56, 172.84], [723.08, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[574.35, 191.58], [602.39, 191.52], [607.75, 193.52], [576.73, 193.59], [574.35, 165.04], [602.39, 165.07], [607.75, 164.23], [576.73, 164.2]]}, {\"category\": \"car\", \"corners_3d\": [[746.48, 207.48], [786.46, 207.31], [827.2, 213.12], [780.67, 213.36], [746.48, 168.57], [786.46, 168.6], [827.2, 167.88], [780.67, 167.85]]}, {\"category\": \"car\", \"corners_3d\": [[129.14, 226.63], [172.37, 226.48], [89.84, 237.49], [37.4, 237.72], [129.14, 186.68], [172.37, 186.64], [89.84, 189.48], [37.4, 189.53]]}, {\"category\": \"car\", \"corners_3d\": [[-13.34, 230.41], [32.82, 229.98], [-38.27, 238.39], [-91.89, 238.95], [-13.34, 192.57], [32.82, 192.43], [-38.27, 195.31], [-91.89, 195.5]]}]\n```" - } - ] - }, - { - "id": 212, - "data_source": "Mono3DRefer", - "scene_name": "000113", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[534.25, 215.52], [580.25, 215.31], [585.56, 223.36], [530.81, 223.67], [534.25, 164.92], [580.25, 164.96], [585.56, 163.47], [530.81, 163.41]]}, {\"category\": \"van\", \"corners_3d\": [[1036.34, 216.54], [1087.12, 220.98], [866.49, 222.56], [835.18, 217.84], [1036.34, 132.92], [1087.12, 128.86], [866.49, 127.41], [835.18, 131.73]]}, {\"category\": \"van\", \"corners_3d\": [[646.26, 187.81], [618.15, 187.86], [610.43, 186.3], [635.65, 186.26], [646.26, 149.72], [618.15, 149.64], [610.43, 152.05], [635.65, 152.12]]}, {\"category\": \"van\", \"corners_3d\": [[1296.66, 203.37], [1237.97, 203.56], [1120.38, 198.25], [1169.57, 198.12], [1296.66, 133.16], [1237.97, 132.91], [1120.38, 139.82], [1169.57, 139.99]]}, {\"category\": \"truck\", \"corners_3d\": [[602.84, 190.72], [569.02, 190.79], [566.47, 188.23], [595.51, 188.18], [602.84, 137.11], [569.02, 136.98], [566.47, 142.08], [595.51, 142.18]]}]\n```" - } - ] - }, - { - "id": 213, - "data_source": "Mono3DRefer", - "scene_name": "000113", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[926.16, 289.18], [1089.12, 297.24], [1287.93, 407.8], [973.87, 380.6], [926.16, 156.1], [1089.12, 154.93], [1287.93, 139.01], [973.87, 142.93]]}, {\"category\": \"van\", \"corners_3d\": [[534.25, 215.52], [580.25, 215.31], [585.56, 223.36], [530.81, 223.67], [534.25, 164.92], [580.25, 164.96], [585.56, 163.47], [530.81, 163.41]]}, {\"category\": \"van\", \"corners_3d\": [[1036.34, 216.54], [1087.12, 220.98], [866.49, 222.56], [835.18, 217.84], [1036.34, 132.92], [1087.12, 128.86], [866.49, 127.41], [835.18, 131.73]]}, {\"category\": \"car\", \"corners_3d\": [[356.02, 220.65], [401.41, 220.39], [372.86, 229.84], [318.19, 230.21], [356.02, 182.2], [401.41, 182.15], [372.86, 183.99], [318.19, 184.07]]}, {\"category\": \"car\", \"corners_3d\": [[274.76, 221.56], [318.8, 221.41], [267.97, 231.09], [214.93, 231.29], [274.76, 168.42], [318.8, 168.43], [267.97, 167.55], [214.93, 167.53]]}, {\"category\": \"car\", \"corners_3d\": [[464.03, 201.91], [498.96, 201.75], [495.85, 205.46], [456.39, 205.67], [464.03, 172.0], [498.96, 172.01], [495.85, 171.9], [456.39, 171.89]]}, {\"category\": \"car\", \"corners_3d\": [[693.02, 209.5], [733.62, 209.25], [770.56, 215.64], [723.08, 215.98], [693.02, 172.85], [733.62, 172.85], [770.56, 172.84], [723.08, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[574.35, 191.58], [602.39, 191.52], [607.75, 193.52], [576.73, 193.59], [574.35, 165.04], [602.39, 165.07], [607.75, 164.23], [576.73, 164.2]]}, {\"category\": \"van\", \"corners_3d\": [[646.26, 187.81], [618.15, 187.86], [610.43, 186.3], [635.65, 186.26], [646.26, 149.72], [618.15, 149.64], [610.43, 152.05], [635.65, 152.12]]}, {\"category\": \"car\", \"corners_3d\": [[746.48, 207.48], [786.46, 207.31], [827.2, 213.12], [780.67, 213.36], [746.48, 168.57], [786.46, 168.6], [827.2, 167.88], [780.67, 167.85]]}, {\"category\": \"van\", \"corners_3d\": [[1296.66, 203.37], [1237.97, 203.56], [1120.38, 198.25], [1169.57, 198.12], [1296.66, 133.16], [1237.97, 132.91], [1120.38, 139.82], [1169.57, 139.99]]}, {\"category\": \"car\", \"corners_3d\": [[129.14, 226.63], [172.37, 226.48], [89.84, 237.49], [37.4, 237.72], [129.14, 186.68], [172.37, 186.64], [89.84, 189.48], [37.4, 189.53]]}, {\"category\": \"car\", \"corners_3d\": [[-13.34, 230.41], [32.82, 229.98], [-38.27, 238.39], [-91.89, 238.95], [-13.34, 192.57], [32.82, 192.43], [-38.27, 195.31], [-91.89, 195.5]]}]\n```" - } - ] - }, - { - "id": 214, - "data_source": "Mono3DRefer", - "scene_name": "000113", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[926.16, 289.18], [1089.12, 297.24], [1287.93, 407.8], [973.87, 380.6], [926.16, 156.1], [1089.12, 154.93], [1287.93, 139.01], [973.87, 142.93]]}, {\"category\": \"car\", \"corners_3d\": [[356.02, 220.65], [401.41, 220.39], [372.86, 229.84], [318.19, 230.21], [356.02, 182.2], [401.41, 182.15], [372.86, 183.99], [318.19, 184.07]]}, {\"category\": \"car\", \"corners_3d\": [[274.76, 221.56], [318.8, 221.41], [267.97, 231.09], [214.93, 231.29], [274.76, 168.42], [318.8, 168.43], [267.97, 167.55], [214.93, 167.53]]}, {\"category\": \"car\", \"corners_3d\": [[464.03, 201.91], [498.96, 201.75], [495.85, 205.46], [456.39, 205.67], [464.03, 172.0], [498.96, 172.01], [495.85, 171.9], [456.39, 171.89]]}, {\"category\": \"car\", \"corners_3d\": [[693.02, 209.5], [733.62, 209.25], [770.56, 215.64], [723.08, 215.98], [693.02, 172.85], [733.62, 172.85], [770.56, 172.84], [723.08, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[574.35, 191.58], [602.39, 191.52], [607.75, 193.52], [576.73, 193.59], [574.35, 165.04], [602.39, 165.07], [607.75, 164.23], [576.73, 164.2]]}, {\"category\": \"car\", \"corners_3d\": [[746.48, 207.48], [786.46, 207.31], [827.2, 213.12], [780.67, 213.36], [746.48, 168.57], [786.46, 168.6], [827.2, 167.88], [780.67, 167.85]]}, {\"category\": \"truck\", \"corners_3d\": [[602.84, 190.72], [569.02, 190.79], [566.47, 188.23], [595.51, 188.18], [602.84, 137.11], [569.02, 136.98], [566.47, 142.08], [595.51, 142.18]]}, {\"category\": \"car\", \"corners_3d\": [[129.14, 226.63], [172.37, 226.48], [89.84, 237.49], [37.4, 237.72], [129.14, 186.68], [172.37, 186.64], [89.84, 189.48], [37.4, 189.53]]}, {\"category\": \"car\", \"corners_3d\": [[-13.34, 230.41], [32.82, 229.98], [-38.27, 238.39], [-91.89, 238.95], [-13.34, 192.57], [32.82, 192.43], [-38.27, 195.31], [-91.89, 195.5]]}]\n```" - } - ] - }, - { - "id": 215, - "data_source": "Mono3DRefer", - "scene_name": "000113", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[926.16, 289.18], [1089.12, 297.24], [1287.93, 407.8], [973.87, 380.6], [926.16, 156.1], [1089.12, 154.93], [1287.93, 139.01], [973.87, 142.93]]}, {\"category\": \"van\", \"corners_3d\": [[534.25, 215.52], [580.25, 215.31], [585.56, 223.36], [530.81, 223.67], [534.25, 164.92], [580.25, 164.96], [585.56, 163.47], [530.81, 163.41]]}, {\"category\": \"van\", \"corners_3d\": [[1036.34, 216.54], [1087.12, 220.98], [866.49, 222.56], [835.18, 217.84], [1036.34, 132.92], [1087.12, 128.86], [866.49, 127.41], [835.18, 131.73]]}, {\"category\": \"car\", \"corners_3d\": [[356.02, 220.65], [401.41, 220.39], [372.86, 229.84], [318.19, 230.21], [356.02, 182.2], [401.41, 182.15], [372.86, 183.99], [318.19, 184.07]]}, {\"category\": \"car\", \"corners_3d\": [[274.76, 221.56], [318.8, 221.41], [267.97, 231.09], [214.93, 231.29], [274.76, 168.42], [318.8, 168.43], [267.97, 167.55], [214.93, 167.53]]}, {\"category\": \"car\", \"corners_3d\": [[464.03, 201.91], [498.96, 201.75], [495.85, 205.46], [456.39, 205.67], [464.03, 172.0], [498.96, 172.01], [495.85, 171.9], [456.39, 171.89]]}, {\"category\": \"car\", \"corners_3d\": [[693.02, 209.5], [733.62, 209.25], [770.56, 215.64], [723.08, 215.98], [693.02, 172.85], [733.62, 172.85], [770.56, 172.84], [723.08, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[574.35, 191.58], [602.39, 191.52], [607.75, 193.52], [576.73, 193.59], [574.35, 165.04], [602.39, 165.07], [607.75, 164.23], [576.73, 164.2]]}, {\"category\": \"van\", \"corners_3d\": [[646.26, 187.81], [618.15, 187.86], [610.43, 186.3], [635.65, 186.26], [646.26, 149.72], [618.15, 149.64], [610.43, 152.05], [635.65, 152.12]]}, {\"category\": \"car\", \"corners_3d\": [[746.48, 207.48], [786.46, 207.31], [827.2, 213.12], [780.67, 213.36], [746.48, 168.57], [786.46, 168.6], [827.2, 167.88], [780.67, 167.85]]}, {\"category\": \"van\", \"corners_3d\": [[1296.66, 203.37], [1237.97, 203.56], [1120.38, 198.25], [1169.57, 198.12], [1296.66, 133.16], [1237.97, 132.91], [1120.38, 139.82], [1169.57, 139.99]]}, {\"category\": \"truck\", \"corners_3d\": [[602.84, 190.72], [569.02, 190.79], [566.47, 188.23], [595.51, 188.18], [602.84, 137.11], [569.02, 136.98], [566.47, 142.08], [595.51, 142.18]]}, {\"category\": \"car\", \"corners_3d\": [[129.14, 226.63], [172.37, 226.48], [89.84, 237.49], [37.4, 237.72], [129.14, 186.68], [172.37, 186.64], [89.84, 189.48], [37.4, 189.53]]}, {\"category\": \"car\", \"corners_3d\": [[-13.34, 230.41], [32.82, 229.98], [-38.27, 238.39], [-91.89, 238.95], [-13.34, 192.57], [32.82, 192.43], [-38.27, 195.31], [-91.89, 195.5]]}]\n```" - } - ] - }, - { - "id": 216, - "data_source": "Mono3DRefer", - "scene_name": "000116", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[539.0, 223.19], [495.83, 223.16], [510.16, 217.17], [548.19, 217.19], [539.0, 182.33], [495.83, 182.33], [510.16, 181.2], [548.19, 181.2]]}, {\"category\": \"car\", \"corners_3d\": [[575.68, 205.27], [606.27, 205.4], [598.02, 209.29], [563.83, 209.13], [575.68, 177.6], [606.27, 177.62], [598.02, 178.19], [563.83, 178.17]]}, {\"category\": \"car\", \"corners_3d\": [[1130.07, 370.49], [927.48, 370.54], [795.61, 288.69], [914.34, 288.67], [1130.07, 150.43], [927.48, 150.43], [795.61, 159.71], [914.34, 159.71]]}, {\"category\": \"car\", \"corners_3d\": [[682.65, 221.47], [728.68, 221.44], [753.33, 230.97], [698.3, 231.01], [682.65, 173.97], [728.68, 173.97], [753.33, 174.19], [698.3, 174.19]]}]\n```" - } - ] - }, - { - "id": 217, - "data_source": "Mono3DRefer", - "scene_name": "000119", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[855.25, 247.28], [887.53, 247.15], [933.69, 258.35], [896.64, 258.52], [855.25, 155.13], [887.53, 155.17], [933.69, 152.5], [896.64, 152.46]]}]\n```" - } - ] - }, - { - "id": 218, - "data_source": "Mono3DRefer", - "scene_name": "000119", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[452.42, 257.04], [382.02, 257.12], [421.87, 241.84], [479.48, 241.78], [452.42, 189.29], [382.02, 189.31], [421.87, 186.32], [479.48, 186.31]]}, {\"category\": \"car\", \"corners_3d\": [[574.48, 252.86], [496.91, 252.52], [525.73, 236.23], [587.49, 236.45], [574.48, 185.78], [496.91, 185.72], [525.73, 183.09], [587.49, 183.12]]}, {\"category\": \"car\", \"corners_3d\": [[544.81, 206.91], [519.61, 206.92], [527.97, 203.45], [550.6, 203.44], [544.81, 180.79], [519.61, 180.8], [527.97, 179.99], [550.6, 179.99]]}]\n```" - } - ] - }, - { - "id": 219, - "data_source": "Mono3DRefer", - "scene_name": "000119", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[512.16, 231.12], [447.21, 231.28], [474.8, 219.78], [526.93, 219.67], [512.16, 152.43], [447.21, 152.37], [474.8, 156.4], [526.93, 156.44]]}]\n```" - } - ] - }, - { - "id": 220, - "data_source": "Mono3DRefer", - "scene_name": "000119", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[452.42, 257.04], [382.02, 257.12], [421.87, 241.84], [479.48, 241.78], [452.42, 189.29], [382.02, 189.31], [421.87, 186.32], [479.48, 186.31]]}, {\"category\": \"car\", \"corners_3d\": [[574.48, 252.86], [496.91, 252.52], [525.73, 236.23], [587.49, 236.45], [574.48, 185.78], [496.91, 185.72], [525.73, 183.09], [587.49, 183.12]]}, {\"category\": \"cyclist\", \"corners_3d\": [[855.25, 247.28], [887.53, 247.15], [933.69, 258.35], [896.64, 258.52], [855.25, 155.13], [887.53, 155.17], [933.69, 152.5], [896.64, 152.46]]}, {\"category\": \"car\", \"corners_3d\": [[544.81, 206.91], [519.61, 206.92], [527.97, 203.45], [550.6, 203.44], [544.81, 180.79], [519.61, 180.8], [527.97, 179.99], [550.6, 179.99]]}]\n```" - } - ] - }, - { - "id": 221, - "data_source": "Mono3DRefer", - "scene_name": "000119", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[512.16, 231.12], [447.21, 231.28], [474.8, 219.78], [526.93, 219.67], [512.16, 152.43], [447.21, 152.37], [474.8, 156.4], [526.93, 156.44]]}, {\"category\": \"cyclist\", \"corners_3d\": [[855.25, 247.28], [887.53, 247.15], [933.69, 258.35], [896.64, 258.52], [855.25, 155.13], [887.53, 155.17], [933.69, 152.5], [896.64, 152.46]]}]\n```" - } - ] - }, - { - "id": 222, - "data_source": "Mono3DRefer", - "scene_name": "000119", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[452.42, 257.04], [382.02, 257.12], [421.87, 241.84], [479.48, 241.78], [452.42, 189.29], [382.02, 189.31], [421.87, 186.32], [479.48, 186.31]]}, {\"category\": \"van\", \"corners_3d\": [[512.16, 231.12], [447.21, 231.28], [474.8, 219.78], [526.93, 219.67], [512.16, 152.43], [447.21, 152.37], [474.8, 156.4], [526.93, 156.44]]}, {\"category\": \"car\", \"corners_3d\": [[574.48, 252.86], [496.91, 252.52], [525.73, 236.23], [587.49, 236.45], [574.48, 185.78], [496.91, 185.72], [525.73, 183.09], [587.49, 183.12]]}, {\"category\": \"car\", \"corners_3d\": [[544.81, 206.91], [519.61, 206.92], [527.97, 203.45], [550.6, 203.44], [544.81, 180.79], [519.61, 180.8], [527.97, 179.99], [550.6, 179.99]]}]\n```" - } - ] - }, - { - "id": 223, - "data_source": "Mono3DRefer", - "scene_name": "000119", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[452.42, 257.04], [382.02, 257.12], [421.87, 241.84], [479.48, 241.78], [452.42, 189.29], [382.02, 189.31], [421.87, 186.32], [479.48, 186.31]]}, {\"category\": \"van\", \"corners_3d\": [[512.16, 231.12], [447.21, 231.28], [474.8, 219.78], [526.93, 219.67], [512.16, 152.43], [447.21, 152.37], [474.8, 156.4], [526.93, 156.44]]}, {\"category\": \"car\", \"corners_3d\": [[574.48, 252.86], [496.91, 252.52], [525.73, 236.23], [587.49, 236.45], [574.48, 185.78], [496.91, 185.72], [525.73, 183.09], [587.49, 183.12]]}, {\"category\": \"cyclist\", \"corners_3d\": [[855.25, 247.28], [887.53, 247.15], [933.69, 258.35], [896.64, 258.52], [855.25, 155.13], [887.53, 155.17], [933.69, 152.5], [896.64, 152.46]]}, {\"category\": \"car\", \"corners_3d\": [[544.81, 206.91], [519.61, 206.92], [527.97, 203.45], [550.6, 203.44], [544.81, 180.79], [519.61, 180.8], [527.97, 179.99], [550.6, 179.99]]}]\n```" - } - ] - }, - { - "id": 224, - "data_source": "Mono3DRefer", - "scene_name": "000121", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[659.6, 188.87], [638.23, 189.04], [596.21, 186.65], [614.84, 186.53], [659.6, 157.9], [638.23, 157.74], [596.21, 159.97], [614.84, 160.09]]}]\n```" - } - ] - }, - { - "id": 225, - "data_source": "Mono3DRefer", - "scene_name": "000121", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[374.28, 224.74], [340.68, 224.22], [384.86, 218.75], [415.25, 219.16], [374.28, 185.2], [340.68, 185.08], [384.86, 183.77], [415.25, 183.87]]}, {\"category\": \"car\", \"corners_3d\": [[509.96, 205.76], [487.72, 205.62], [504.07, 203.21], [524.75, 203.34], [509.96, 181.24], [487.72, 181.21], [504.07, 180.59], [524.75, 180.63]]}, {\"category\": \"car\", \"corners_3d\": [[726.77, 194.46], [752.88, 194.64], [750.12, 196.42], [721.93, 196.2], [726.77, 171.07], [752.88, 171.06], [750.12, 170.91], [721.93, 170.93]]}, {\"category\": \"car\", \"corners_3d\": [[656.03, 199.22], [684.03, 199.51], [668.85, 202.53], [637.89, 202.17], [656.03, 172.67], [684.03, 172.67], [668.85, 172.65], [637.89, 172.65]]}]\n```" - } - ] - }, - { - "id": 226, - "data_source": "Mono3DRefer", - "scene_name": "000121", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[374.28, 224.74], [340.68, 224.22], [384.86, 218.75], [415.25, 219.16], [374.28, 185.2], [340.68, 185.08], [384.86, 183.77], [415.25, 183.87]]}, {\"category\": \"car\", \"corners_3d\": [[509.96, 205.76], [487.72, 205.62], [504.07, 203.21], [524.75, 203.34], [509.96, 181.24], [487.72, 181.21], [504.07, 180.59], [524.75, 180.63]]}, {\"category\": \"tram\", \"corners_3d\": [[659.6, 188.87], [638.23, 189.04], [596.21, 186.65], [614.84, 186.53], [659.6, 157.9], [638.23, 157.74], [596.21, 159.97], [614.84, 160.09]]}, {\"category\": \"car\", \"corners_3d\": [[726.77, 194.46], [752.88, 194.64], [750.12, 196.42], [721.93, 196.2], [726.77, 171.07], [752.88, 171.06], [750.12, 170.91], [721.93, 170.93]]}, {\"category\": \"car\", \"corners_3d\": [[656.03, 199.22], [684.03, 199.51], [668.85, 202.53], [637.89, 202.17], [656.03, 172.67], [684.03, 172.67], [668.85, 172.65], [637.89, 172.65]]}]\n```" - } - ] - }, - { - "id": 227, - "data_source": "Mono3DRefer", - "scene_name": "000122", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[437.56, 227.97], [518.07, 227.97], [344.9, 331.64], [113.04, 331.6], [437.56, 115.86], [518.07, 115.86], [344.9, 8.65], [113.04, 8.69]]}, {\"category\": \"tram\", \"corners_3d\": [[500.75, 208.89], [553.28, 208.89], [526.72, 225.73], [449.67, 225.72], [500.75, 135.27], [553.28, 135.27], [526.72, 117.71], [449.67, 117.72]]}]\n```" - } - ] - }, - { - "id": 228, - "data_source": "Mono3DRefer", - "scene_name": "000122", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[586.01, 234.61], [644.73, 234.62], [650.52, 244.97], [581.96, 244.96], [586.01, 178.27], [644.73, 178.27], [650.52, 179.18], [581.96, 179.18]]}]\n```" - } - ] - }, - { - "id": 229, - "data_source": "Mono3DRefer", - "scene_name": "000122", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[789.62, 225.22], [807.74, 225.13], [834.21, 230.71], [814.22, 230.82], [789.62, 161.67], [807.74, 161.69], [834.21, 160.5], [814.22, 160.47]]}]\n```" - } - ] - }, - { - "id": 230, - "data_source": "Mono3DRefer", - "scene_name": "000122", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[437.56, 227.97], [518.07, 227.97], [344.9, 331.64], [113.04, 331.6], [437.56, 115.86], [518.07, 115.86], [344.9, 8.65], [113.04, 8.69]]}, {\"category\": \"car\", \"corners_3d\": [[586.01, 234.61], [644.73, 234.62], [650.52, 244.97], [581.96, 244.96], [586.01, 178.27], [644.73, 178.27], [650.52, 179.18], [581.96, 179.18]]}, {\"category\": \"tram\", \"corners_3d\": [[500.75, 208.89], [553.28, 208.89], [526.72, 225.73], [449.67, 225.72], [500.75, 135.27], [553.28, 135.27], [526.72, 117.71], [449.67, 117.72]]}]\n```" - } - ] - }, - { - "id": 231, - "data_source": "Mono3DRefer", - "scene_name": "000122", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[437.56, 227.97], [518.07, 227.97], [344.9, 331.64], [113.04, 331.6], [437.56, 115.86], [518.07, 115.86], [344.9, 8.65], [113.04, 8.69]]}, {\"category\": \"cyclist\", \"corners_3d\": [[789.62, 225.22], [807.74, 225.13], [834.21, 230.71], [814.22, 230.82], [789.62, 161.67], [807.74, 161.69], [834.21, 160.5], [814.22, 160.47]]}, {\"category\": \"tram\", \"corners_3d\": [[500.75, 208.89], [553.28, 208.89], [526.72, 225.73], [449.67, 225.72], [500.75, 135.27], [553.28, 135.27], [526.72, 117.71], [449.67, 117.72]]}]\n```" - } - ] - }, - { - "id": 232, - "data_source": "Mono3DRefer", - "scene_name": "000122", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[586.01, 234.61], [644.73, 234.62], [650.52, 244.97], [581.96, 244.96], [586.01, 178.27], [644.73, 178.27], [650.52, 179.18], [581.96, 179.18]]}, {\"category\": \"cyclist\", \"corners_3d\": [[789.62, 225.22], [807.74, 225.13], [834.21, 230.71], [814.22, 230.82], [789.62, 161.67], [807.74, 161.69], [834.21, 160.5], [814.22, 160.47]]}]\n```" - } - ] - }, - { - "id": 233, - "data_source": "Mono3DRefer", - "scene_name": "000122", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[437.56, 227.97], [518.07, 227.97], [344.9, 331.64], [113.04, 331.6], [437.56, 115.86], [518.07, 115.86], [344.9, 8.65], [113.04, 8.69]]}, {\"category\": \"car\", \"corners_3d\": [[586.01, 234.61], [644.73, 234.62], [650.52, 244.97], [581.96, 244.96], [586.01, 178.27], [644.73, 178.27], [650.52, 179.18], [581.96, 179.18]]}, {\"category\": \"cyclist\", \"corners_3d\": [[789.62, 225.22], [807.74, 225.13], [834.21, 230.71], [814.22, 230.82], [789.62, 161.67], [807.74, 161.69], [834.21, 160.5], [814.22, 160.47]]}, {\"category\": \"tram\", \"corners_3d\": [[500.75, 208.89], [553.28, 208.89], [526.72, 225.73], [449.67, 225.72], [500.75, 135.27], [553.28, 135.27], [526.72, 117.71], [449.67, 117.72]]}]\n```" - } - ] - }, - { - "id": 234, - "data_source": "Mono3DRefer", - "scene_name": "000123", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[628.36, 198.79], [602.51, 198.67], [611.11, 196.47], [634.78, 196.57], [628.36, 176.42], [602.51, 176.4], [611.11, 176.1], [634.78, 176.12]]}]\n```" - } - ] - }, - { - "id": 235, - "data_source": "Mono3DRefer", - "scene_name": "000125", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[555.17, 299.27], [690.78, 299.29], [740.8, 377.72], [521.08, 377.67], [555.17, 170.33], [690.78, 170.33], [740.8, 168.76], [521.08, 168.76]]}, {\"category\": \"car\", \"corners_3d\": [[207.76, 223.28], [238.11, 220.8], [313.98, 222.59], [286.51, 225.26], [207.76, 192.6], [238.11, 191.63], [313.98, 192.33], [286.51, 193.38]]}, {\"category\": \"car\", \"corners_3d\": [[402.36, 205.41], [376.87, 205.39], [394.56, 203.06], [418.24, 203.07], [402.36, 186.1], [376.87, 186.09], [394.56, 185.14], [418.24, 185.15]]}, {\"category\": \"car\", \"corners_3d\": [[62.41, 231.8], [95.46, 229.16], [144.32, 230.43], [112.83, 233.18], [62.41, 202.32], [95.46, 201.01], [144.32, 201.64], [112.83, 203.01]]}, {\"category\": \"car\", \"corners_3d\": [[327.84, 214.54], [345.6, 212.9], [399.9, 213.68], [384.05, 215.39], [327.84, 184.22], [345.6, 183.77], [399.9, 183.98], [384.05, 184.45]]}, {\"category\": \"car\", \"corners_3d\": [[360.22, 208.3], [335.13, 208.28], [350.8, 206.35], [374.54, 206.37], [360.22, 188.81], [335.13, 188.8], [350.8, 187.93], [374.54, 187.94]]}, {\"category\": \"car\", \"corners_3d\": [[450.32, 201.68], [425.88, 201.67], [441.93, 199.24], [464.31, 199.25], [450.32, 182.09], [425.88, 182.08], [441.93, 181.31], [464.31, 181.31]]}]\n```" - } - ] - }, - { - "id": 236, - "data_source": "Mono3DRefer", - "scene_name": "000128", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[624.12, 195.52], [657.5, 195.62], [652.59, 201.91], [610.06, 201.76], [624.12, 159.42], [657.5, 159.37], [652.59, 155.64], [610.06, 155.73]]}]\n```" - } - ] - }, - { - "id": 237, - "data_source": "Mono3DRefer", - "scene_name": "000128", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[90.65, 262.99], [6.24, 263.0], [126.08, 245.07], [193.71, 245.07], [90.65, 186.9], [6.24, 186.9], [126.08, 184.11], [193.71, 184.11]]}]\n```" - } - ] - }, - { - "id": 238, - "data_source": "Mono3DRefer", - "scene_name": "000128", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[90.65, 262.99], [6.24, 263.0], [126.08, 245.07], [193.71, 245.07], [90.65, 186.9], [6.24, 186.9], [126.08, 184.11], [193.71, 184.11]]}, {\"category\": \"bus\", \"corners_3d\": [[624.12, 195.52], [657.5, 195.62], [652.59, 201.91], [610.06, 201.76], [624.12, 159.42], [657.5, 159.37], [652.59, 155.64], [610.06, 155.73]]}]\n```" - } - ] - }, - { - "id": 239, - "data_source": "Mono3DRefer", - "scene_name": "000129", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[143.01, 289.45], [53.5, 288.87], [197.48, 260.28], [265.4, 260.61], [143.01, 189.58], [53.5, 189.49], [197.48, 185.39], [265.4, 185.44]]}, {\"category\": \"car\", \"corners_3d\": [[354.01, 236.01], [301.57, 235.78], [355.56, 225.89], [399.9, 226.05], [354.01, 187.65], [301.57, 187.59], [355.56, 185.28], [399.9, 185.32]]}, {\"category\": \"car\", \"corners_3d\": [[426.1, 221.72], [386.41, 221.58], [416.68, 215.88], [451.78, 215.98], [426.1, 176.6], [386.41, 176.59], [416.68, 176.16], [451.78, 176.16]]}, {\"category\": \"car\", \"corners_3d\": [[459.84, 215.14], [429.59, 215.05], [449.06, 211.24], [476.61, 211.31], [459.84, 184.48], [429.59, 184.45], [449.06, 183.41], [476.61, 183.43]]}, {\"category\": \"car\", \"corners_3d\": [[483.51, 211.46], [451.93, 211.36], [472.03, 207.5], [500.48, 207.58], [483.51, 178.66], [451.93, 178.64], [472.03, 178.06], [500.48, 178.07]]}, {\"category\": \"car\", \"corners_3d\": [[521.89, 203.71], [497.35, 203.64], [509.76, 201.29], [532.45, 201.35], [521.89, 180.79], [497.35, 180.77], [509.76, 180.17], [532.45, 180.18]]}]\n```" - } - ] - }, - { - "id": 240, - "data_source": "Mono3DRefer", - "scene_name": "000129", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[703.85, 185.51], [722.32, 185.53], [740.31, 189.73], [715.7, 189.69], [703.85, 159.81], [722.32, 159.78], [740.31, 155.45], [715.7, 155.5]]}]\n```" - } - ] - }, - { - "id": 241, - "data_source": "Mono3DRefer", - "scene_name": "000129", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[143.01, 289.45], [53.5, 288.87], [197.48, 260.28], [265.4, 260.61], [143.01, 189.58], [53.5, 189.49], [197.48, 185.39], [265.4, 185.44]]}, {\"category\": \"car\", \"corners_3d\": [[354.01, 236.01], [301.57, 235.78], [355.56, 225.89], [399.9, 226.05], [354.01, 187.65], [301.57, 187.59], [355.56, 185.28], [399.9, 185.32]]}, {\"category\": \"tram\", \"corners_3d\": [[703.85, 185.51], [722.32, 185.53], [740.31, 189.73], [715.7, 189.69], [703.85, 159.81], [722.32, 159.78], [740.31, 155.45], [715.7, 155.5]]}, {\"category\": \"car\", \"corners_3d\": [[426.1, 221.72], [386.41, 221.58], [416.68, 215.88], [451.78, 215.98], [426.1, 176.6], [386.41, 176.59], [416.68, 176.16], [451.78, 176.16]]}, {\"category\": \"car\", \"corners_3d\": [[459.84, 215.14], [429.59, 215.05], [449.06, 211.24], [476.61, 211.31], [459.84, 184.48], [429.59, 184.45], [449.06, 183.41], [476.61, 183.43]]}, {\"category\": \"car\", \"corners_3d\": [[483.51, 211.46], [451.93, 211.36], [472.03, 207.5], [500.48, 207.58], [483.51, 178.66], [451.93, 178.64], [472.03, 178.06], [500.48, 178.07]]}, {\"category\": \"car\", \"corners_3d\": [[521.89, 203.71], [497.35, 203.64], [509.76, 201.29], [532.45, 201.35], [521.89, 180.79], [497.35, 180.77], [509.76, 180.17], [532.45, 180.18]]}]\n```" - } - ] - }, - { - "id": 242, - "data_source": "Mono3DRefer", - "scene_name": "000131", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[320.7, 228.98], [358.08, 230.78], [245.41, 241.49], [205.5, 238.98], [320.7, 186.4], [358.08, 186.84], [245.41, 189.42], [205.5, 188.82]]}, {\"category\": \"car\", \"corners_3d\": [[501.15, 212.56], [530.34, 213.47], [474.59, 218.44], [443.31, 217.3], [501.15, 178.55], [530.34, 178.68], [474.59, 179.4], [443.31, 179.23]]}, {\"category\": \"car\", \"corners_3d\": [[235.4, 224.02], [209.35, 222.95], [302.91, 216.46], [327.26, 217.26], [235.4, 181.89], [209.35, 181.7], [302.91, 180.56], [327.26, 180.7]]}, {\"category\": \"car\", \"corners_3d\": [[483.94, 206.16], [459.46, 205.75], [497.88, 202.56], [520.4, 202.89], [483.94, 174.85], [459.46, 174.83], [497.88, 174.63], [520.4, 174.65]]}, {\"category\": \"car\", \"corners_3d\": [[605.37, 199.98], [629.61, 200.41], [602.2, 202.77], [576.39, 202.26], [605.37, 174.71], [629.61, 174.74], [602.2, 174.9], [576.39, 174.87]]}, {\"category\": \"car\", \"corners_3d\": [[721.67, 189.33], [737.26, 189.48], [728.33, 190.24], [712.11, 190.08], [721.67, 173.72], [737.26, 173.73], [728.33, 173.77], [712.11, 173.76]]}]\n```" - } - ] - }, - { - "id": 243, - "data_source": "Mono3DRefer", - "scene_name": "000131", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[584.72, 195.91], [565.74, 195.75], [583.88, 193.98], [601.5, 194.12], [584.72, 171.17], [565.74, 171.18], [583.88, 171.31], [601.5, 171.3]]}]\n```" - } - ] - }, - { - "id": 244, - "data_source": "Mono3DRefer", - "scene_name": "000131", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[320.7, 228.98], [358.08, 230.78], [245.41, 241.49], [205.5, 238.98], [320.7, 186.4], [358.08, 186.84], [245.41, 189.42], [205.5, 188.82]]}, {\"category\": \"car\", \"corners_3d\": [[501.15, 212.56], [530.34, 213.47], [474.59, 218.44], [443.31, 217.3], [501.15, 178.55], [530.34, 178.68], [474.59, 179.4], [443.31, 179.23]]}, {\"category\": \"car\", \"corners_3d\": [[235.4, 224.02], [209.35, 222.95], [302.91, 216.46], [327.26, 217.26], [235.4, 181.89], [209.35, 181.7], [302.91, 180.56], [327.26, 180.7]]}, {\"category\": \"car\", \"corners_3d\": [[483.94, 206.16], [459.46, 205.75], [497.88, 202.56], [520.4, 202.89], [483.94, 174.85], [459.46, 174.83], [497.88, 174.63], [520.4, 174.65]]}, {\"category\": \"van\", \"corners_3d\": [[584.72, 195.91], [565.74, 195.75], [583.88, 193.98], [601.5, 194.12], [584.72, 171.17], [565.74, 171.18], [583.88, 171.31], [601.5, 171.3]]}, {\"category\": \"car\", \"corners_3d\": [[605.37, 199.98], [629.61, 200.41], [602.2, 202.77], [576.39, 202.26], [605.37, 174.71], [629.61, 174.74], [602.2, 174.9], [576.39, 174.87]]}, {\"category\": \"car\", \"corners_3d\": [[721.67, 189.33], [737.26, 189.48], [728.33, 190.24], [712.11, 190.08], [721.67, 173.72], [737.26, 173.73], [728.33, 173.77], [712.11, 173.76]]}]\n```" - } - ] - }, - { - "id": 245, - "data_source": "Mono3DRefer", - "scene_name": "000137", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[84.53, 269.81], [-43.65, 269.99], [191.13, 234.66], [272.45, 234.58], [84.53, 105.85], [-43.65, 105.72], [191.13, 130.14], [272.45, 130.19]]}]\n```" - } - ] - }, - { - "id": 246, - "data_source": "Mono3DRefer", - "scene_name": "000137", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[666.49, 191.82], [689.93, 191.73], [724.37, 195.17], [696.83, 195.3], [666.49, 165.7], [689.93, 165.74], [724.37, 164.44], [696.83, 164.39]]}]\n```" - } - ] - }, - { - "id": 247, - "data_source": "Mono3DRefer", - "scene_name": "000137", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[597.11, 204.23], [626.46, 204.23], [627.72, 206.65], [596.1, 206.65], [597.11, 176.07], [626.46, 176.07], [627.72, 176.32], [596.1, 176.32]]}, {\"category\": \"car\", \"corners_3d\": [[354.92, 218.26], [317.96, 218.31], [342.59, 214.26], [376.23, 214.22], [354.92, 187.71], [317.96, 187.73], [342.59, 186.4], [376.23, 186.39]]}, {\"category\": \"car\", \"corners_3d\": [[448.31, 202.42], [426.92, 202.41], [438.73, 200.57], [458.79, 200.58], [448.31, 182.53], [426.92, 182.53], [438.73, 181.93], [458.79, 181.93]]}]\n```" - } - ] - }, - { - "id": 248, - "data_source": "Mono3DRefer", - "scene_name": "000137", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[666.49, 191.82], [689.93, 191.73], [724.37, 195.17], [696.83, 195.3], [666.49, 165.7], [689.93, 165.74], [724.37, 164.44], [696.83, 164.39]]}, {\"category\": \"truck\", \"corners_3d\": [[84.53, 269.81], [-43.65, 269.99], [191.13, 234.66], [272.45, 234.58], [84.53, 105.85], [-43.65, 105.72], [191.13, 130.14], [272.45, 130.19]]}]\n```" - } - ] - }, - { - "id": 249, - "data_source": "Mono3DRefer", - "scene_name": "000137", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[597.11, 204.23], [626.46, 204.23], [627.72, 206.65], [596.1, 206.65], [597.11, 176.07], [626.46, 176.07], [627.72, 176.32], [596.1, 176.32]]}, {\"category\": \"truck\", \"corners_3d\": [[84.53, 269.81], [-43.65, 269.99], [191.13, 234.66], [272.45, 234.58], [84.53, 105.85], [-43.65, 105.72], [191.13, 130.14], [272.45, 130.19]]}, {\"category\": \"car\", \"corners_3d\": [[354.92, 218.26], [317.96, 218.31], [342.59, 214.26], [376.23, 214.22], [354.92, 187.71], [317.96, 187.73], [342.59, 186.4], [376.23, 186.39]]}, {\"category\": \"car\", \"corners_3d\": [[448.31, 202.42], [426.92, 202.41], [438.73, 200.57], [458.79, 200.58], [448.31, 182.53], [426.92, 182.53], [438.73, 181.93], [458.79, 181.93]]}]\n```" - } - ] - }, - { - "id": 250, - "data_source": "Mono3DRefer", - "scene_name": "000137", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[597.11, 204.23], [626.46, 204.23], [627.72, 206.65], [596.1, 206.65], [597.11, 176.07], [626.46, 176.07], [627.72, 176.32], [596.1, 176.32]]}, {\"category\": \"bus\", \"corners_3d\": [[666.49, 191.82], [689.93, 191.73], [724.37, 195.17], [696.83, 195.3], [666.49, 165.7], [689.93, 165.74], [724.37, 164.44], [696.83, 164.39]]}, {\"category\": \"car\", \"corners_3d\": [[354.92, 218.26], [317.96, 218.31], [342.59, 214.26], [376.23, 214.22], [354.92, 187.71], [317.96, 187.73], [342.59, 186.4], [376.23, 186.39]]}, {\"category\": \"car\", \"corners_3d\": [[448.31, 202.42], [426.92, 202.41], [438.73, 200.57], [458.79, 200.58], [448.31, 182.53], [426.92, 182.53], [438.73, 181.93], [458.79, 181.93]]}]\n```" - } - ] - }, - { - "id": 251, - "data_source": "Mono3DRefer", - "scene_name": "000137", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[597.11, 204.23], [626.46, 204.23], [627.72, 206.65], [596.1, 206.65], [597.11, 176.07], [626.46, 176.07], [627.72, 176.32], [596.1, 176.32]]}, {\"category\": \"bus\", \"corners_3d\": [[666.49, 191.82], [689.93, 191.73], [724.37, 195.17], [696.83, 195.3], [666.49, 165.7], [689.93, 165.74], [724.37, 164.44], [696.83, 164.39]]}, {\"category\": \"truck\", \"corners_3d\": [[84.53, 269.81], [-43.65, 269.99], [191.13, 234.66], [272.45, 234.58], [84.53, 105.85], [-43.65, 105.72], [191.13, 130.14], [272.45, 130.19]]}, {\"category\": \"car\", \"corners_3d\": [[354.92, 218.26], [317.96, 218.31], [342.59, 214.26], [376.23, 214.22], [354.92, 187.71], [317.96, 187.73], [342.59, 186.4], [376.23, 186.39]]}, {\"category\": \"car\", \"corners_3d\": [[448.31, 202.42], [426.92, 202.41], [438.73, 200.57], [458.79, 200.58], [448.31, 182.53], [426.92, 182.53], [438.73, 181.93], [458.79, 181.93]]}]\n```" - } - ] - }, - { - "id": 252, - "data_source": "Mono3DRefer", - "scene_name": "000138", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[98.61, 219.56], [140.66, 219.57], [85.73, 225.03], [38.76, 225.03], [98.61, 181.41], [140.66, 181.41], [85.73, 182.41], [38.76, 182.41]]}, {\"category\": \"car\", \"corners_3d\": [[458.19, 251.98], [383.06, 251.9], [432.67, 235.08], [491.84, 235.13], [458.19, 182.79], [383.06, 182.78], [432.67, 180.67], [491.84, 180.67]]}, {\"category\": \"car\", \"corners_3d\": [[294.36, 200.68], [317.78, 200.7], [294.95, 202.77], [269.81, 202.75], [294.36, 177.06], [317.78, 177.06], [294.95, 177.38], [269.81, 177.37]]}, {\"category\": \"car\", \"corners_3d\": [[324.88, 199.59], [307.36, 199.56], [339.5, 196.97], [355.34, 196.99], [324.88, 176.27], [307.36, 176.26], [339.5, 175.93], [355.34, 175.93]]}, {\"category\": \"car\", \"corners_3d\": [[568.89, 194.97], [547.26, 194.97], [551.17, 193.39], [571.24, 193.38], [568.89, 176.94], [547.26, 176.94], [551.17, 176.65], [571.24, 176.65]]}, {\"category\": \"car\", \"corners_3d\": [[542.81, 210.37], [506.75, 210.39], [517.68, 206.08], [549.59, 206.06], [542.81, 173.06], [506.75, 173.06], [517.68, 173.04], [549.59, 173.04]]}, {\"category\": \"car\", \"corners_3d\": [[590.16, 207.02], [622.74, 207.04], [623.51, 211.91], [586.28, 211.89], [590.16, 174.65], [622.74, 174.65], [623.51, 174.9], [586.28, 174.9]]}]\n```" - } - ] - }, - { - "id": 253, - "data_source": "Mono3DRefer", - "scene_name": "000139", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[197.97, 244.67], [133.16, 244.67], [215.07, 232.31], [268.72, 232.3], [197.97, 178.55], [133.16, 178.55], [215.07, 177.57], [268.72, 177.57]]}]\n```" - } - ] - }, - { - "id": 254, - "data_source": "Mono3DRefer", - "scene_name": "000139", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[595.54, 202.23], [620.7, 202.23], [621.39, 204.15], [594.58, 204.15], [595.54, 178.08], [620.7, 178.09], [621.39, 178.43], [594.58, 178.43]]}]\n```" - } - ] - }, - { - "id": 255, - "data_source": "Mono3DRefer", - "scene_name": "000139", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[624.34, 191.01], [647.91, 190.99], [659.99, 194.27], [632.16, 194.3], [624.34, 165.39], [647.91, 165.4], [659.99, 164.05], [632.16, 164.03]]}]\n```" - } - ] - }, - { - "id": 256, - "data_source": "Mono3DRefer", - "scene_name": "000139", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[595.54, 202.23], [620.7, 202.23], [621.39, 204.15], [594.58, 204.15], [595.54, 178.08], [620.7, 178.09], [621.39, 178.43], [594.58, 178.43]]}, {\"category\": \"van\", \"corners_3d\": [[197.97, 244.67], [133.16, 244.67], [215.07, 232.31], [268.72, 232.3], [197.97, 178.55], [133.16, 178.55], [215.07, 177.57], [268.72, 177.57]]}]\n```" - } - ] - }, - { - "id": 257, - "data_source": "Mono3DRefer", - "scene_name": "000139", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[624.34, 191.01], [647.91, 190.99], [659.99, 194.27], [632.16, 194.3], [624.34, 165.39], [647.91, 165.4], [659.99, 164.05], [632.16, 164.03]]}, {\"category\": \"van\", \"corners_3d\": [[197.97, 244.67], [133.16, 244.67], [215.07, 232.31], [268.72, 232.3], [197.97, 178.55], [133.16, 178.55], [215.07, 177.57], [268.72, 177.57]]}]\n```" - } - ] - }, - { - "id": 258, - "data_source": "Mono3DRefer", - "scene_name": "000139", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[595.54, 202.23], [620.7, 202.23], [621.39, 204.15], [594.58, 204.15], [595.54, 178.08], [620.7, 178.09], [621.39, 178.43], [594.58, 178.43]]}, {\"category\": \"bus\", \"corners_3d\": [[624.34, 191.01], [647.91, 190.99], [659.99, 194.27], [632.16, 194.3], [624.34, 165.39], [647.91, 165.4], [659.99, 164.05], [632.16, 164.03]]}]\n```" - } - ] - }, - { - "id": 259, - "data_source": "Mono3DRefer", - "scene_name": "000139", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[595.54, 202.23], [620.7, 202.23], [621.39, 204.15], [594.58, 204.15], [595.54, 178.08], [620.7, 178.09], [621.39, 178.43], [594.58, 178.43]]}, {\"category\": \"bus\", \"corners_3d\": [[624.34, 191.01], [647.91, 190.99], [659.99, 194.27], [632.16, 194.3], [624.34, 165.39], [647.91, 165.4], [659.99, 164.05], [632.16, 164.03]]}, {\"category\": \"van\", \"corners_3d\": [[197.97, 244.67], [133.16, 244.67], [215.07, 232.31], [268.72, 232.3], [197.97, 178.55], [133.16, 178.55], [215.07, 177.57], [268.72, 177.57]]}]\n```" - } - ] - }, - { - "id": 260, - "data_source": "Mono3DRefer", - "scene_name": "000141", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[495.18, 264.8], [405.9, 264.35], [462.84, 241.89], [530.34, 242.14], [495.18, 187.36], [405.9, 187.28], [462.84, 183.74], [530.34, 183.78]]}]\n```" - } - ] - }, - { - "id": 261, - "data_source": "Mono3DRefer", - "scene_name": "000142", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[118.36, 447.42], [38.56, 447.74], [190.1, 373.78], [248.32, 373.6], [118.36, 166.59], [38.56, 166.58], [190.1, 168.27], [248.32, 168.28]]}]\n```" - } - ] - }, - { - "id": 262, - "data_source": "Mono3DRefer", - "scene_name": "000142", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1203.75, 337.56], [1298.35, 337.58], [1360.12, 352.36], [1257.02, 352.34], [1203.75, 112.04], [1298.35, 112.03], [1360.12, 106.57], [1257.02, 106.58]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[673.47, 199.98], [668.49, 200.23], [651.31, 199.71], [656.36, 199.48], [673.47, 161.22], [668.49, 161.11], [651.31, 161.33], [656.36, 161.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[680.56, 200.61], [674.73, 200.91], [660.12, 200.46], [666.01, 200.17], [680.56, 163.26], [674.73, 163.16], [660.12, 163.31], [666.01, 163.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[617.57, 191.64], [617.94, 191.83], [604.79, 191.85], [604.54, 191.66], [617.57, 167.22], [617.94, 167.17], [604.79, 167.16], [604.54, 167.22]]}]\n```" - } - ] - }, - { - "id": 263, - "data_source": "Mono3DRefer", - "scene_name": "000142", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[837.17, 340.37], [1000.82, 339.64], [1370.53, 491.84], [1059.39, 494.54], [837.17, 171.74], [1000.82, 171.75], [1370.53, 170.74], [1059.39, 170.72]]}, {\"category\": \"car\", \"corners_3d\": [[726.88, 265.24], [821.95, 265.37], [917.05, 308.37], [777.69, 308.09], [726.88, 162.17], [821.95, 162.16], [917.05, 157.18], [777.69, 157.22]]}, {\"category\": \"car\", \"corners_3d\": [[688.86, 233.63], [757.36, 233.64], [806.28, 253.83], [715.01, 253.82], [688.86, 159.47], [757.36, 159.46], [806.28, 155.02], [715.01, 155.02]]}, {\"category\": \"car\", \"corners_3d\": [[679.16, 221.89], [730.07, 222.04], [747.94, 231.68], [687.01, 231.48], [679.16, 171.59], [730.07, 171.59], [747.94, 171.34], [687.01, 171.35]]}, {\"category\": \"car\", \"corners_3d\": [[555.86, 197.54], [531.04, 197.53], [537.64, 195.62], [560.54, 195.62], [555.86, 168.1], [531.04, 168.1], [537.64, 168.47], [560.54, 168.47]]}, {\"category\": \"car\", \"corners_3d\": [[474.05, 225.92], [428.54, 225.83], [456.52, 218.49], [495.76, 218.56], [474.05, 178.81], [428.54, 178.8], [456.52, 177.98], [495.76, 177.98]]}, {\"category\": \"car\", \"corners_3d\": [[16.03, 425.98], [-186.54, 423.1], [153.14, 319.99], [273.97, 320.98], [16.03, 214.08], [-186.54, 213.61], [153.14, 196.82], [273.97, 196.98]]}, {\"category\": \"car\", \"corners_3d\": [[310.66, 301.78], [197.24, 301.8], [318.29, 263.89], [398.36, 263.88], [310.66, 190.54], [197.24, 190.54], [318.29, 185.34], [398.36, 185.34]]}, {\"category\": \"car\", \"corners_3d\": [[410.77, 252.63], [338.2, 252.39], [393.83, 237.26], [452.69, 237.42], [410.77, 177.4], [338.2, 177.38], [393.83, 176.52], [452.69, 176.53]]}]\n```" - } - ] - }, - { - "id": 264, - "data_source": "Mono3DRefer", - "scene_name": "000142", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1203.75, 337.56], [1298.35, 337.58], [1360.12, 352.36], [1257.02, 352.34], [1203.75, 112.04], [1298.35, 112.03], [1360.12, 106.57], [1257.02, 106.58]]}, {\"category\": \"cyclist\", \"corners_3d\": [[118.36, 447.42], [38.56, 447.74], [190.1, 373.78], [248.32, 373.6], [118.36, 166.59], [38.56, 166.58], [190.1, 168.27], [248.32, 168.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[673.47, 199.98], [668.49, 200.23], [651.31, 199.71], [656.36, 199.48], [673.47, 161.22], [668.49, 161.11], [651.31, 161.33], [656.36, 161.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[680.56, 200.61], [674.73, 200.91], [660.12, 200.46], [666.01, 200.17], [680.56, 163.26], [674.73, 163.16], [660.12, 163.31], [666.01, 163.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[617.57, 191.64], [617.94, 191.83], [604.79, 191.85], [604.54, 191.66], [617.57, 167.22], [617.94, 167.17], [604.79, 167.16], [604.54, 167.22]]}]\n```" - } - ] - }, - { - "id": 265, - "data_source": "Mono3DRefer", - "scene_name": "000142", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[118.36, 447.42], [38.56, 447.74], [190.1, 373.78], [248.32, 373.6], [118.36, 166.59], [38.56, 166.58], [190.1, 168.27], [248.32, 168.28]]}, {\"category\": \"car\", \"corners_3d\": [[837.17, 340.37], [1000.82, 339.64], [1370.53, 491.84], [1059.39, 494.54], [837.17, 171.74], [1000.82, 171.75], [1370.53, 170.74], [1059.39, 170.72]]}, {\"category\": \"car\", \"corners_3d\": [[726.88, 265.24], [821.95, 265.37], [917.05, 308.37], [777.69, 308.09], [726.88, 162.17], [821.95, 162.16], [917.05, 157.18], [777.69, 157.22]]}, {\"category\": \"car\", \"corners_3d\": [[688.86, 233.63], [757.36, 233.64], [806.28, 253.83], [715.01, 253.82], [688.86, 159.47], [757.36, 159.46], [806.28, 155.02], [715.01, 155.02]]}, {\"category\": \"car\", \"corners_3d\": [[679.16, 221.89], [730.07, 222.04], [747.94, 231.68], [687.01, 231.48], [679.16, 171.59], [730.07, 171.59], [747.94, 171.34], [687.01, 171.35]]}, {\"category\": \"car\", \"corners_3d\": [[555.86, 197.54], [531.04, 197.53], [537.64, 195.62], [560.54, 195.62], [555.86, 168.1], [531.04, 168.1], [537.64, 168.47], [560.54, 168.47]]}, {\"category\": \"car\", \"corners_3d\": [[474.05, 225.92], [428.54, 225.83], [456.52, 218.49], [495.76, 218.56], [474.05, 178.81], [428.54, 178.8], [456.52, 177.98], [495.76, 177.98]]}, {\"category\": \"car\", \"corners_3d\": [[16.03, 425.98], [-186.54, 423.1], [153.14, 319.99], [273.97, 320.98], [16.03, 214.08], [-186.54, 213.61], [153.14, 196.82], [273.97, 196.98]]}, {\"category\": \"car\", \"corners_3d\": [[310.66, 301.78], [197.24, 301.8], [318.29, 263.89], [398.36, 263.88], [310.66, 190.54], [197.24, 190.54], [318.29, 185.34], [398.36, 185.34]]}, {\"category\": \"car\", \"corners_3d\": [[410.77, 252.63], [338.2, 252.39], [393.83, 237.26], [452.69, 237.42], [410.77, 177.4], [338.2, 177.38], [393.83, 176.52], [452.69, 176.53]]}]\n```" - } - ] - }, - { - "id": 266, - "data_source": "Mono3DRefer", - "scene_name": "000142", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1203.75, 337.56], [1298.35, 337.58], [1360.12, 352.36], [1257.02, 352.34], [1203.75, 112.04], [1298.35, 112.03], [1360.12, 106.57], [1257.02, 106.58]]}, {\"category\": \"car\", \"corners_3d\": [[837.17, 340.37], [1000.82, 339.64], [1370.53, 491.84], [1059.39, 494.54], [837.17, 171.74], [1000.82, 171.75], [1370.53, 170.74], [1059.39, 170.72]]}, {\"category\": \"car\", \"corners_3d\": [[726.88, 265.24], [821.95, 265.37], [917.05, 308.37], [777.69, 308.09], [726.88, 162.17], [821.95, 162.16], [917.05, 157.18], [777.69, 157.22]]}, {\"category\": \"car\", \"corners_3d\": [[688.86, 233.63], [757.36, 233.64], [806.28, 253.83], [715.01, 253.82], [688.86, 159.47], [757.36, 159.46], [806.28, 155.02], [715.01, 155.02]]}, {\"category\": \"car\", \"corners_3d\": [[679.16, 221.89], [730.07, 222.04], [747.94, 231.68], [687.01, 231.48], [679.16, 171.59], [730.07, 171.59], [747.94, 171.34], [687.01, 171.35]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[673.47, 199.98], [668.49, 200.23], [651.31, 199.71], [656.36, 199.48], [673.47, 161.22], [668.49, 161.11], [651.31, 161.33], [656.36, 161.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[680.56, 200.61], [674.73, 200.91], [660.12, 200.46], [666.01, 200.17], [680.56, 163.26], [674.73, 163.16], [660.12, 163.31], [666.01, 163.41]]}, {\"category\": \"car\", \"corners_3d\": [[555.86, 197.54], [531.04, 197.53], [537.64, 195.62], [560.54, 195.62], [555.86, 168.1], [531.04, 168.1], [537.64, 168.47], [560.54, 168.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[617.57, 191.64], [617.94, 191.83], [604.79, 191.85], [604.54, 191.66], [617.57, 167.22], [617.94, 167.17], [604.79, 167.16], [604.54, 167.22]]}, {\"category\": \"car\", \"corners_3d\": [[474.05, 225.92], [428.54, 225.83], [456.52, 218.49], [495.76, 218.56], [474.05, 178.81], [428.54, 178.8], [456.52, 177.98], [495.76, 177.98]]}, {\"category\": \"car\", \"corners_3d\": [[16.03, 425.98], [-186.54, 423.1], [153.14, 319.99], [273.97, 320.98], [16.03, 214.08], [-186.54, 213.61], [153.14, 196.82], [273.97, 196.98]]}, {\"category\": \"car\", \"corners_3d\": [[310.66, 301.78], [197.24, 301.8], [318.29, 263.89], [398.36, 263.88], [310.66, 190.54], [197.24, 190.54], [318.29, 185.34], [398.36, 185.34]]}, {\"category\": \"car\", \"corners_3d\": [[410.77, 252.63], [338.2, 252.39], [393.83, 237.26], [452.69, 237.42], [410.77, 177.4], [338.2, 177.38], [393.83, 176.52], [452.69, 176.53]]}]\n```" - } - ] - }, - { - "id": 267, - "data_source": "Mono3DRefer", - "scene_name": "000142", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1203.75, 337.56], [1298.35, 337.58], [1360.12, 352.36], [1257.02, 352.34], [1203.75, 112.04], [1298.35, 112.03], [1360.12, 106.57], [1257.02, 106.58]]}, {\"category\": \"cyclist\", \"corners_3d\": [[118.36, 447.42], [38.56, 447.74], [190.1, 373.78], [248.32, 373.6], [118.36, 166.59], [38.56, 166.58], [190.1, 168.27], [248.32, 168.28]]}, {\"category\": \"car\", \"corners_3d\": [[837.17, 340.37], [1000.82, 339.64], [1370.53, 491.84], [1059.39, 494.54], [837.17, 171.74], [1000.82, 171.75], [1370.53, 170.74], [1059.39, 170.72]]}, {\"category\": \"car\", \"corners_3d\": [[726.88, 265.24], [821.95, 265.37], [917.05, 308.37], [777.69, 308.09], [726.88, 162.17], [821.95, 162.16], [917.05, 157.18], [777.69, 157.22]]}, {\"category\": \"car\", \"corners_3d\": [[688.86, 233.63], [757.36, 233.64], [806.28, 253.83], [715.01, 253.82], [688.86, 159.47], [757.36, 159.46], [806.28, 155.02], [715.01, 155.02]]}, {\"category\": \"car\", \"corners_3d\": [[679.16, 221.89], [730.07, 222.04], [747.94, 231.68], [687.01, 231.48], [679.16, 171.59], [730.07, 171.59], [747.94, 171.34], [687.01, 171.35]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[673.47, 199.98], [668.49, 200.23], [651.31, 199.71], [656.36, 199.48], [673.47, 161.22], [668.49, 161.11], [651.31, 161.33], [656.36, 161.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[680.56, 200.61], [674.73, 200.91], [660.12, 200.46], [666.01, 200.17], [680.56, 163.26], [674.73, 163.16], [660.12, 163.31], [666.01, 163.41]]}, {\"category\": \"car\", \"corners_3d\": [[555.86, 197.54], [531.04, 197.53], [537.64, 195.62], [560.54, 195.62], [555.86, 168.1], [531.04, 168.1], [537.64, 168.47], [560.54, 168.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[617.57, 191.64], [617.94, 191.83], [604.79, 191.85], [604.54, 191.66], [617.57, 167.22], [617.94, 167.17], [604.79, 167.16], [604.54, 167.22]]}, {\"category\": \"car\", \"corners_3d\": [[474.05, 225.92], [428.54, 225.83], [456.52, 218.49], [495.76, 218.56], [474.05, 178.81], [428.54, 178.8], [456.52, 177.98], [495.76, 177.98]]}, {\"category\": \"car\", \"corners_3d\": [[16.03, 425.98], [-186.54, 423.1], [153.14, 319.99], [273.97, 320.98], [16.03, 214.08], [-186.54, 213.61], [153.14, 196.82], [273.97, 196.98]]}, {\"category\": \"car\", \"corners_3d\": [[310.66, 301.78], [197.24, 301.8], [318.29, 263.89], [398.36, 263.88], [310.66, 190.54], [197.24, 190.54], [318.29, 185.34], [398.36, 185.34]]}, {\"category\": \"car\", \"corners_3d\": [[410.77, 252.63], [338.2, 252.39], [393.83, 237.26], [452.69, 237.42], [410.77, 177.4], [338.2, 177.38], [393.83, 176.52], [452.69, 176.53]]}]\n```" - } - ] - }, - { - "id": 268, - "data_source": "Mono3DRefer", - "scene_name": "000143", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[951.25, 201.94], [940.75, 201.02], [991.08, 201.05], [1003.22, 201.96], [951.25, 180.86], [940.75, 180.61], [991.08, 180.62], [1003.22, 180.87]]}, {\"category\": \"car\", \"corners_3d\": [[60.04, 248.32], [4.7, 248.32], [79.98, 238.92], [128.42, 238.92], [60.04, 202.79], [4.7, 202.79], [79.98, 199.06], [128.42, 199.06]]}, {\"category\": \"car\", \"corners_3d\": [[152.91, 228.29], [121.14, 228.63], [148.83, 224.55], [178.14, 224.26], [152.91, 203.23], [121.14, 203.42], [148.83, 201.19], [178.14, 201.03]]}]\n```" - } - ] - }, - { - "id": 269, - "data_source": "Mono3DRefer", - "scene_name": "000143", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[589.8, 193.19], [612.85, 193.18], [615.88, 196.76], [588.77, 196.77], [589.8, 168.2], [612.85, 168.21], [615.88, 167.39], [588.77, 167.38]]}]\n```" - } - ] - }, - { - "id": 270, - "data_source": "Mono3DRefer", - "scene_name": "000143", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[951.25, 201.94], [940.75, 201.02], [991.08, 201.05], [1003.22, 201.96], [951.25, 180.86], [940.75, 180.61], [991.08, 180.62], [1003.22, 180.87]]}, {\"category\": \"bus\", \"corners_3d\": [[589.8, 193.19], [612.85, 193.18], [615.88, 196.76], [588.77, 196.77], [589.8, 168.2], [612.85, 168.21], [615.88, 167.39], [588.77, 167.38]]}, {\"category\": \"car\", \"corners_3d\": [[60.04, 248.32], [4.7, 248.32], [79.98, 238.92], [128.42, 238.92], [60.04, 202.79], [4.7, 202.79], [79.98, 199.06], [128.42, 199.06]]}, {\"category\": \"car\", \"corners_3d\": [[152.91, 228.29], [121.14, 228.63], [148.83, 224.55], [178.14, 224.26], [152.91, 203.23], [121.14, 203.42], [148.83, 201.19], [178.14, 201.03]]}]\n```" - } - ] - }, - { - "id": 271, - "data_source": "Mono3DRefer", - "scene_name": "000144", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[659.94, 205.2], [663.73, 205.64], [646.08, 205.9], [642.5, 205.44], [659.94, 170.51], [663.73, 170.48], [646.08, 170.46], [642.5, 170.49]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[342.36, 243.82], [321.42, 243.84], [327.13, 242.4], [347.64, 242.38], [342.36, 187.48], [321.42, 187.48], [327.13, 187.18], [347.64, 187.18]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[848.41, 239.8], [840.46, 241.15], [796.18, 238.28], [804.65, 237.04], [848.41, 153.84], [840.46, 153.45], [796.18, 154.27], [804.65, 154.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[850.45, 238.17], [841.33, 239.64], [804.95, 237.3], [814.54, 235.93], [850.45, 158.32], [841.33, 157.99], [804.95, 158.51], [814.54, 158.82]]}]\n```" - } - ] - }, - { - "id": 272, - "data_source": "Mono3DRefer", - "scene_name": "000144", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[45.33, 415.98], [-141.69, 414.75], [164.41, 317.97], [277.32, 318.41], [45.33, 219.87], [-141.69, 219.63], [164.41, 200.92], [277.32, 201.0]]}, {\"category\": \"car\", \"corners_3d\": [[923.83, 384.5], [1151.87, 389.1], [2386.31, 932.61], [1538.16, 878.69], [923.83, 168.79], [1151.87, 168.7], [2386.31, 158.25], [1538.16, 159.29]]}, {\"category\": \"car\", \"corners_3d\": [[513.59, 221.16], [473.16, 221.19], [488.6, 215.38], [524.17, 215.36], [513.59, 173.36], [473.16, 173.36], [488.6, 173.3], [524.17, 173.3]]}]\n```" - } - ] - }, - { - "id": 273, - "data_source": "Mono3DRefer", - "scene_name": "000144", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[659.94, 205.2], [663.73, 205.64], [646.08, 205.9], [642.5, 205.44], [659.94, 170.51], [663.73, 170.48], [646.08, 170.46], [642.5, 170.49]]}, {\"category\": \"car\", \"corners_3d\": [[45.33, 415.98], [-141.69, 414.75], [164.41, 317.97], [277.32, 318.41], [45.33, 219.87], [-141.69, 219.63], [164.41, 200.92], [277.32, 201.0]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[342.36, 243.82], [321.42, 243.84], [327.13, 242.4], [347.64, 242.38], [342.36, 187.48], [321.42, 187.48], [327.13, 187.18], [347.64, 187.18]]}, {\"category\": \"car\", \"corners_3d\": [[923.83, 384.5], [1151.87, 389.1], [2386.31, 932.61], [1538.16, 878.69], [923.83, 168.79], [1151.87, 168.7], [2386.31, 158.25], [1538.16, 159.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[848.41, 239.8], [840.46, 241.15], [796.18, 238.28], [804.65, 237.04], [848.41, 153.84], [840.46, 153.45], [796.18, 154.27], [804.65, 154.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[850.45, 238.17], [841.33, 239.64], [804.95, 237.3], [814.54, 235.93], [850.45, 158.32], [841.33, 157.99], [804.95, 158.51], [814.54, 158.82]]}, {\"category\": \"car\", \"corners_3d\": [[513.59, 221.16], [473.16, 221.19], [488.6, 215.38], [524.17, 215.36], [513.59, 173.36], [473.16, 173.36], [488.6, 173.3], [524.17, 173.3]]}]\n```" - } - ] - }, - { - "id": 274, - "data_source": "Mono3DRefer", - "scene_name": "000145", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[961.23, 222.21], [1002.58, 222.03], [1069.07, 228.2], [1021.87, 228.43], [961.23, 181.29], [1002.58, 181.28], [1069.07, 181.4], [1021.87, 181.4]]}, {\"category\": \"car\", \"corners_3d\": [[1034.48, 222.37], [1076.15, 222.17], [1155.39, 228.42], [1107.89, 228.69], [1034.48, 183.51], [1076.15, 183.5], [1155.39, 183.94], [1107.89, 183.96]]}, {\"category\": \"car\", \"corners_3d\": [[1230.44, 229.55], [1184.73, 229.65], [1097.93, 222.57], [1137.21, 222.5], [1230.44, 184.52], [1184.73, 184.53], [1097.93, 183.95], [1137.21, 183.94]]}, {\"category\": \"car\", \"corners_3d\": [[601.83, 197.3], [630.8, 197.27], [637.72, 199.4], [605.07, 199.44], [601.83, 172.68], [630.8, 172.7], [637.72, 171.7], [605.07, 171.69]]}]\n```" - } - ] - }, - { - "id": 275, - "data_source": "Mono3DRefer", - "scene_name": "000145", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[677.81, 256.85], [709.77, 255.05], [784.02, 267.21], [748.79, 269.65], [677.81, 156.42], [709.77, 156.98], [784.02, 153.15], [748.79, 152.38]]}, {\"category\": \"cyclist\", \"corners_3d\": [[136.2, 334.65], [82.92, 335.0], [182.96, 304.05], [225.41, 303.83], [136.2, 151.71], [82.92, 151.65], [182.96, 157.43], [225.41, 157.47]]}]\n```" - } - ] - }, - { - "id": 276, - "data_source": "Mono3DRefer", - "scene_name": "000145", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[327.79, 201.2], [346.88, 201.1], [343.08, 201.74], [323.37, 201.85], [327.79, 158.8], [346.88, 158.91], [343.08, 158.23], [323.37, 158.12]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[427.82, 194.65], [439.89, 194.65], [434.8, 195.08], [422.36, 195.08], [427.82, 163.85], [439.89, 163.85], [434.8, 163.34], [422.36, 163.34]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[324.14, 189.6], [334.92, 189.57], [333.3, 189.72], [322.33, 189.75], [324.14, 156.37], [334.92, 156.45], [333.3, 156.05], [322.33, 155.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[351.2, 190.68], [364.57, 190.68], [359.44, 190.9], [345.78, 190.9], [351.2, 155.31], [364.57, 155.31], [359.44, 154.77], [345.78, 154.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[388.79, 250.73], [414.2, 248.81], [449.14, 253.5], [422.99, 255.7], [388.79, 167.07], [414.2, 167.44], [449.14, 166.54], [422.99, 166.12]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[498.15, 252.38], [524.59, 249.43], [563.0, 252.56], [537.02, 255.78], [498.15, 158.63], [524.59, 159.52], [563.0, 158.57], [537.02, 157.59]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[315.18, 195.19], [297.54, 195.25], [299.96, 195.04], [317.35, 194.98], [315.18, 156.01], [297.54, 155.9], [299.96, 156.26], [317.35, 156.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[222.3, 197.07], [228.52, 196.79], [236.54, 196.78], [230.46, 197.06], [222.3, 153.26], [228.52, 153.72], [236.54, 153.73], [230.46, 153.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1004.89, 210.12], [995.77, 209.46], [1010.65, 209.45], [1020.11, 210.11], [1004.89, 173.82], [995.77, 173.97], [1010.65, 173.97], [1020.11, 173.82]]}]\n```" - } - ] - }, - { - "id": 277, - "data_source": "Mono3DRefer", - "scene_name": "000145", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[961.23, 222.21], [1002.58, 222.03], [1069.07, 228.2], [1021.87, 228.43], [961.23, 181.29], [1002.58, 181.28], [1069.07, 181.4], [1021.87, 181.4]]}, {\"category\": \"car\", \"corners_3d\": [[1034.48, 222.37], [1076.15, 222.17], [1155.39, 228.42], [1107.89, 228.69], [1034.48, 183.51], [1076.15, 183.5], [1155.39, 183.94], [1107.89, 183.96]]}, {\"category\": \"car\", \"corners_3d\": [[1230.44, 229.55], [1184.73, 229.65], [1097.93, 222.57], [1137.21, 222.5], [1230.44, 184.52], [1184.73, 184.53], [1097.93, 183.95], [1137.21, 183.94]]}, {\"category\": \"cyclist\", \"corners_3d\": [[677.81, 256.85], [709.77, 255.05], [784.02, 267.21], [748.79, 269.65], [677.81, 156.42], [709.77, 156.98], [784.02, 153.15], [748.79, 152.38]]}, {\"category\": \"cyclist\", \"corners_3d\": [[136.2, 334.65], [82.92, 335.0], [182.96, 304.05], [225.41, 303.83], [136.2, 151.71], [82.92, 151.65], [182.96, 157.43], [225.41, 157.47]]}, {\"category\": \"car\", \"corners_3d\": [[601.83, 197.3], [630.8, 197.27], [637.72, 199.4], [605.07, 199.44], [601.83, 172.68], [630.8, 172.7], [637.72, 171.7], [605.07, 171.69]]}]\n```" - } - ] - }, - { - "id": 278, - "data_source": "Mono3DRefer", - "scene_name": "000145", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[327.79, 201.2], [346.88, 201.1], [343.08, 201.74], [323.37, 201.85], [327.79, 158.8], [346.88, 158.91], [343.08, 158.23], [323.37, 158.12]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[427.82, 194.65], [439.89, 194.65], [434.8, 195.08], [422.36, 195.08], [427.82, 163.85], [439.89, 163.85], [434.8, 163.34], [422.36, 163.34]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[324.14, 189.6], [334.92, 189.57], [333.3, 189.72], [322.33, 189.75], [324.14, 156.37], [334.92, 156.45], [333.3, 156.05], [322.33, 155.97]]}, {\"category\": \"car\", \"corners_3d\": [[961.23, 222.21], [1002.58, 222.03], [1069.07, 228.2], [1021.87, 228.43], [961.23, 181.29], [1002.58, 181.28], [1069.07, 181.4], [1021.87, 181.4]]}, {\"category\": \"car\", \"corners_3d\": [[1034.48, 222.37], [1076.15, 222.17], [1155.39, 228.42], [1107.89, 228.69], [1034.48, 183.51], [1076.15, 183.5], [1155.39, 183.94], [1107.89, 183.96]]}, {\"category\": \"car\", \"corners_3d\": [[1230.44, 229.55], [1184.73, 229.65], [1097.93, 222.57], [1137.21, 222.5], [1230.44, 184.52], [1184.73, 184.53], [1097.93, 183.95], [1137.21, 183.94]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[351.2, 190.68], [364.57, 190.68], [359.44, 190.9], [345.78, 190.9], [351.2, 155.31], [364.57, 155.31], [359.44, 154.77], [345.78, 154.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[388.79, 250.73], [414.2, 248.81], [449.14, 253.5], [422.99, 255.7], [388.79, 167.07], [414.2, 167.44], [449.14, 166.54], [422.99, 166.12]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[498.15, 252.38], [524.59, 249.43], [563.0, 252.56], [537.02, 255.78], [498.15, 158.63], [524.59, 159.52], [563.0, 158.57], [537.02, 157.59]]}, {\"category\": \"car\", \"corners_3d\": [[601.83, 197.3], [630.8, 197.27], [637.72, 199.4], [605.07, 199.44], [601.83, 172.68], [630.8, 172.7], [637.72, 171.7], [605.07, 171.69]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[315.18, 195.19], [297.54, 195.25], [299.96, 195.04], [317.35, 194.98], [315.18, 156.01], [297.54, 155.9], [299.96, 156.26], [317.35, 156.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[222.3, 197.07], [228.52, 196.79], [236.54, 196.78], [230.46, 197.06], [222.3, 153.26], [228.52, 153.72], [236.54, 153.73], [230.46, 153.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1004.89, 210.12], [995.77, 209.46], [1010.65, 209.45], [1020.11, 210.11], [1004.89, 173.82], [995.77, 173.97], [1010.65, 173.97], [1020.11, 173.82]]}]\n```" - } - ] - }, - { - "id": 279, - "data_source": "Mono3DRefer", - "scene_name": "000145", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[327.79, 201.2], [346.88, 201.1], [343.08, 201.74], [323.37, 201.85], [327.79, 158.8], [346.88, 158.91], [343.08, 158.23], [323.37, 158.12]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[427.82, 194.65], [439.89, 194.65], [434.8, 195.08], [422.36, 195.08], [427.82, 163.85], [439.89, 163.85], [434.8, 163.34], [422.36, 163.34]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[324.14, 189.6], [334.92, 189.57], [333.3, 189.72], [322.33, 189.75], [324.14, 156.37], [334.92, 156.45], [333.3, 156.05], [322.33, 155.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[351.2, 190.68], [364.57, 190.68], [359.44, 190.9], [345.78, 190.9], [351.2, 155.31], [364.57, 155.31], [359.44, 154.77], [345.78, 154.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[388.79, 250.73], [414.2, 248.81], [449.14, 253.5], [422.99, 255.7], [388.79, 167.07], [414.2, 167.44], [449.14, 166.54], [422.99, 166.12]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[498.15, 252.38], [524.59, 249.43], [563.0, 252.56], [537.02, 255.78], [498.15, 158.63], [524.59, 159.52], [563.0, 158.57], [537.02, 157.59]]}, {\"category\": \"cyclist\", \"corners_3d\": [[677.81, 256.85], [709.77, 255.05], [784.02, 267.21], [748.79, 269.65], [677.81, 156.42], [709.77, 156.98], [784.02, 153.15], [748.79, 152.38]]}, {\"category\": \"cyclist\", \"corners_3d\": [[136.2, 334.65], [82.92, 335.0], [182.96, 304.05], [225.41, 303.83], [136.2, 151.71], [82.92, 151.65], [182.96, 157.43], [225.41, 157.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[315.18, 195.19], [297.54, 195.25], [299.96, 195.04], [317.35, 194.98], [315.18, 156.01], [297.54, 155.9], [299.96, 156.26], [317.35, 156.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[222.3, 197.07], [228.52, 196.79], [236.54, 196.78], [230.46, 197.06], [222.3, 153.26], [228.52, 153.72], [236.54, 153.73], [230.46, 153.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1004.89, 210.12], [995.77, 209.46], [1010.65, 209.45], [1020.11, 210.11], [1004.89, 173.82], [995.77, 173.97], [1010.65, 173.97], [1020.11, 173.82]]}]\n```" - } - ] - }, - { - "id": 280, - "data_source": "Mono3DRefer", - "scene_name": "000145", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[327.79, 201.2], [346.88, 201.1], [343.08, 201.74], [323.37, 201.85], [327.79, 158.8], [346.88, 158.91], [343.08, 158.23], [323.37, 158.12]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[427.82, 194.65], [439.89, 194.65], [434.8, 195.08], [422.36, 195.08], [427.82, 163.85], [439.89, 163.85], [434.8, 163.34], [422.36, 163.34]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[324.14, 189.6], [334.92, 189.57], [333.3, 189.72], [322.33, 189.75], [324.14, 156.37], [334.92, 156.45], [333.3, 156.05], [322.33, 155.97]]}, {\"category\": \"car\", \"corners_3d\": [[961.23, 222.21], [1002.58, 222.03], [1069.07, 228.2], [1021.87, 228.43], [961.23, 181.29], [1002.58, 181.28], [1069.07, 181.4], [1021.87, 181.4]]}, {\"category\": \"car\", \"corners_3d\": [[1034.48, 222.37], [1076.15, 222.17], [1155.39, 228.42], [1107.89, 228.69], [1034.48, 183.51], [1076.15, 183.5], [1155.39, 183.94], [1107.89, 183.96]]}, {\"category\": \"car\", \"corners_3d\": [[1230.44, 229.55], [1184.73, 229.65], [1097.93, 222.57], [1137.21, 222.5], [1230.44, 184.52], [1184.73, 184.53], [1097.93, 183.95], [1137.21, 183.94]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[351.2, 190.68], [364.57, 190.68], [359.44, 190.9], [345.78, 190.9], [351.2, 155.31], [364.57, 155.31], [359.44, 154.77], [345.78, 154.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[388.79, 250.73], [414.2, 248.81], [449.14, 253.5], [422.99, 255.7], [388.79, 167.07], [414.2, 167.44], [449.14, 166.54], [422.99, 166.12]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[498.15, 252.38], [524.59, 249.43], [563.0, 252.56], [537.02, 255.78], [498.15, 158.63], [524.59, 159.52], [563.0, 158.57], [537.02, 157.59]]}, {\"category\": \"cyclist\", \"corners_3d\": [[677.81, 256.85], [709.77, 255.05], [784.02, 267.21], [748.79, 269.65], [677.81, 156.42], [709.77, 156.98], [784.02, 153.15], [748.79, 152.38]]}, {\"category\": \"cyclist\", \"corners_3d\": [[136.2, 334.65], [82.92, 335.0], [182.96, 304.05], [225.41, 303.83], [136.2, 151.71], [82.92, 151.65], [182.96, 157.43], [225.41, 157.47]]}, {\"category\": \"car\", \"corners_3d\": [[601.83, 197.3], [630.8, 197.27], [637.72, 199.4], [605.07, 199.44], [601.83, 172.68], [630.8, 172.7], [637.72, 171.7], [605.07, 171.69]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[315.18, 195.19], [297.54, 195.25], [299.96, 195.04], [317.35, 194.98], [315.18, 156.01], [297.54, 155.9], [299.96, 156.26], [317.35, 156.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[222.3, 197.07], [228.52, 196.79], [236.54, 196.78], [230.46, 197.06], [222.3, 153.26], [228.52, 153.72], [236.54, 153.73], [230.46, 153.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1004.89, 210.12], [995.77, 209.46], [1010.65, 209.45], [1020.11, 210.11], [1004.89, 173.82], [995.77, 173.97], [1010.65, 173.97], [1020.11, 173.82]]}]\n```" - } - ] - }, - { - "id": 281, - "data_source": "Mono3DRefer", - "scene_name": "000148", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[782.43, 201.45], [803.04, 200.88], [860.61, 203.46], [839.34, 204.14], [782.43, 174.51], [803.04, 174.48], [860.61, 174.63], [839.34, 174.67]]}]\n```" - } - ] - }, - { - "id": 282, - "data_source": "Mono3DRefer", - "scene_name": "000150", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[286.93, 226.44], [273.48, 226.47], [294.48, 222.88], [307.02, 222.85], [286.93, 180.8], [273.48, 180.81], [294.48, 180.28], [307.02, 180.27]]}]\n```" - } - ] - }, - { - "id": 283, - "data_source": "Mono3DRefer", - "scene_name": "000150", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[570.67, 219.14], [615.35, 219.12], [617.03, 225.38], [566.3, 225.42], [570.67, 175.81], [615.35, 175.8], [617.03, 176.2], [566.3, 176.21]]}, {\"category\": \"car\", \"corners_3d\": [[483.4, 211.66], [453.64, 211.66], [467.9, 208.1], [494.92, 208.1], [483.4, 184.08], [453.64, 184.08], [467.9, 183.05], [494.92, 183.05]]}, {\"category\": \"car\", \"corners_3d\": [[557.23, 199.92], [532.35, 199.94], [537.17, 197.84], [560.12, 197.83], [557.23, 178.08], [532.35, 178.09], [537.17, 177.68], [560.12, 177.68]]}]\n```" - } - ] - }, - { - "id": 284, - "data_source": "Mono3DRefer", - "scene_name": "000150", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[570.67, 219.14], [615.35, 219.12], [617.03, 225.38], [566.3, 225.42], [570.67, 175.81], [615.35, 175.8], [617.03, 176.2], [566.3, 176.21]]}, {\"category\": \"car\", \"corners_3d\": [[483.4, 211.66], [453.64, 211.66], [467.9, 208.1], [494.92, 208.1], [483.4, 184.08], [453.64, 184.08], [467.9, 183.05], [494.92, 183.05]]}, {\"category\": \"car\", \"corners_3d\": [[557.23, 199.92], [532.35, 199.94], [537.17, 197.84], [560.12, 197.83], [557.23, 178.08], [532.35, 178.09], [537.17, 177.68], [560.12, 177.68]]}, {\"category\": \"cyclist\", \"corners_3d\": [[286.93, 226.44], [273.48, 226.47], [294.48, 222.88], [307.02, 222.85], [286.93, 180.8], [273.48, 180.81], [294.48, 180.28], [307.02, 180.27]]}]\n```" - } - ] - }, - { - "id": 285, - "data_source": "Mono3DRefer", - "scene_name": "000151", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[60.53, 407.31], [302.45, 407.37], [-436.39, 962.03], [-1268.51, 961.29], [60.53, 203.38], [302.45, 203.39], [-436.39, 259.32], [-1268.51, 259.25]]}, {\"category\": \"car\", \"corners_3d\": [[976.25, 207.53], [1004.82, 209.74], [864.15, 209.56], [846.27, 207.39], [976.25, 163.41], [1004.82, 162.02], [864.15, 162.13], [846.27, 163.51]]}]\n```" - } - ] - }, - { - "id": 286, - "data_source": "Mono3DRefer", - "scene_name": "000151", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[992.59, 212.94], [1025.72, 213.57], [1029.49, 216.42], [993.49, 215.68], [992.59, 150.13], [1025.72, 149.55], [1029.49, 146.88], [993.49, 147.57]]}]\n```" - } - ] - }, - { - "id": 287, - "data_source": "Mono3DRefer", - "scene_name": "000151", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[60.53, 407.31], [302.45, 407.37], [-436.39, 962.03], [-1268.51, 961.29], [60.53, 203.38], [302.45, 203.39], [-436.39, 259.32], [-1268.51, 259.25]]}, {\"category\": \"cyclist\", \"corners_3d\": [[992.59, 212.94], [1025.72, 213.57], [1029.49, 216.42], [993.49, 215.68], [992.59, 150.13], [1025.72, 149.55], [1029.49, 146.88], [993.49, 147.57]]}, {\"category\": \"car\", \"corners_3d\": [[976.25, 207.53], [1004.82, 209.74], [864.15, 209.56], [846.27, 207.39], [976.25, 163.41], [1004.82, 162.02], [864.15, 162.13], [846.27, 163.51]]}]\n```" - } - ] - }, - { - "id": 288, - "data_source": "Mono3DRefer", - "scene_name": "000153", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[652.52, 211.77], [638.5, 211.42], [648.68, 210.59], [662.43, 210.92], [652.52, 162.49], [638.5, 162.69], [648.68, 163.17], [662.43, 162.98]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[639.87, 213.13], [623.63, 212.8], [632.48, 211.84], [648.33, 212.14], [639.87, 164.53], [623.63, 164.69], [632.48, 165.17], [648.33, 165.02]]}]\n```" - } - ] - }, - { - "id": 289, - "data_source": "Mono3DRefer", - "scene_name": "000153", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[669.95, 192.78], [697.14, 192.8], [704.44, 194.2], [674.15, 194.17], [669.95, 161.45], [697.14, 161.42], [704.44, 159.25], [674.15, 159.29]]}]\n```" - } - ] - }, - { - "id": 290, - "data_source": "Mono3DRefer", - "scene_name": "000153", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[669.95, 192.78], [697.14, 192.8], [704.44, 194.2], [674.15, 194.17], [669.95, 161.45], [697.14, 161.42], [704.44, 159.25], [674.15, 159.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[652.52, 211.77], [638.5, 211.42], [648.68, 210.59], [662.43, 210.92], [652.52, 162.49], [638.5, 162.69], [648.68, 163.17], [662.43, 162.98]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[639.87, 213.13], [623.63, 212.8], [632.48, 211.84], [648.33, 212.14], [639.87, 164.53], [623.63, 164.69], [632.48, 165.17], [648.33, 165.02]]}]\n```" - } - ] - }, - { - "id": 291, - "data_source": "Mono3DRefer", - "scene_name": "000156", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[149.59, 271.52], [47.62, 289.42], [-327.93, 283.8], [-171.33, 267.56], [149.59, 124.23], [47.62, 113.15], [-327.93, 116.63], [-171.33, 126.67]]}]\n```" - } - ] - }, - { - "id": 292, - "data_source": "Mono3DRefer", - "scene_name": "000157", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[22.01, 355.08], [-51.07, 361.73], [6.56, 324.63], [64.05, 320.31], [22.01, 178.61], [-51.07, 178.82], [6.56, 177.64], [64.05, 177.51]]}]\n```" - } - ] - }, - { - "id": 293, - "data_source": "Mono3DRefer", - "scene_name": "000157", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1371.89, 300.8], [1622.82, 327.45], [1535.79, 409.7], [1204.71, 352.41], [1371.89, 174.53], [1622.82, 174.88], [1535.79, 175.96], [1204.71, 175.21]]}, {\"category\": \"car\", \"corners_3d\": [[1100.38, 271.67], [1231.55, 286.47], [1022.58, 318.96], [893.71, 295.36], [1100.38, 140.74], [1231.55, 135.93], [1022.58, 125.37], [893.71, 133.04]]}, {\"category\": \"car\", \"corners_3d\": [[832.13, 242.09], [892.44, 248.56], [710.58, 263.08], [658.09, 254.04], [832.13, 174.17], [892.44, 174.29], [710.58, 174.57], [658.09, 174.4]]}, {\"category\": \"car\", \"corners_3d\": [[692.0, 233.91], [728.99, 238.3], [613.78, 245.25], [581.63, 239.92], [692.0, 183.32], [728.99, 184.07], [613.78, 185.26], [581.63, 184.35]]}, {\"category\": \"car\", \"corners_3d\": [[319.61, 230.2], [369.87, 228.14], [406.08, 236.76], [349.22, 239.54], [319.61, 192.46], [369.87, 191.75], [406.08, 194.7], [349.22, 195.65]]}, {\"category\": \"car\", \"corners_3d\": [[219.98, 233.32], [268.2, 231.55], [279.53, 240.92], [223.74, 243.3], [219.98, 194.57], [268.2, 193.94], [279.53, 197.3], [223.74, 198.16]]}]\n```" - } - ] - }, - { - "id": 294, - "data_source": "Mono3DRefer", - "scene_name": "000157", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[199.08, 223.87], [257.82, 222.08], [271.69, 233.34], [199.53, 236.07], [199.08, 155.63], [257.82, 156.23], [271.69, 152.43], [199.53, 151.51]]}]\n```" - } - ] - }, - { - "id": 295, - "data_source": "Mono3DRefer", - "scene_name": "000157", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1371.89, 300.8], [1622.82, 327.45], [1535.79, 409.7], [1204.71, 352.41], [1371.89, 174.53], [1622.82, 174.88], [1535.79, 175.96], [1204.71, 175.21]]}, {\"category\": \"car\", \"corners_3d\": [[1100.38, 271.67], [1231.55, 286.47], [1022.58, 318.96], [893.71, 295.36], [1100.38, 140.74], [1231.55, 135.93], [1022.58, 125.37], [893.71, 133.04]]}, {\"category\": \"car\", \"corners_3d\": [[832.13, 242.09], [892.44, 248.56], [710.58, 263.08], [658.09, 254.04], [832.13, 174.17], [892.44, 174.29], [710.58, 174.57], [658.09, 174.4]]}, {\"category\": \"cyclist\", \"corners_3d\": [[22.01, 355.08], [-51.07, 361.73], [6.56, 324.63], [64.05, 320.31], [22.01, 178.61], [-51.07, 178.82], [6.56, 177.64], [64.05, 177.51]]}, {\"category\": \"car\", \"corners_3d\": [[692.0, 233.91], [728.99, 238.3], [613.78, 245.25], [581.63, 239.92], [692.0, 183.32], [728.99, 184.07], [613.78, 185.26], [581.63, 184.35]]}, {\"category\": \"car\", \"corners_3d\": [[319.61, 230.2], [369.87, 228.14], [406.08, 236.76], [349.22, 239.54], [319.61, 192.46], [369.87, 191.75], [406.08, 194.7], [349.22, 195.65]]}, {\"category\": \"car\", \"corners_3d\": [[219.98, 233.32], [268.2, 231.55], [279.53, 240.92], [223.74, 243.3], [219.98, 194.57], [268.2, 193.94], [279.53, 197.3], [223.74, 198.16]]}]\n```" - } - ] - }, - { - "id": 296, - "data_source": "Mono3DRefer", - "scene_name": "000157", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[22.01, 355.08], [-51.07, 361.73], [6.56, 324.63], [64.05, 320.31], [22.01, 178.61], [-51.07, 178.82], [6.56, 177.64], [64.05, 177.51]]}, {\"category\": \"truck\", \"corners_3d\": [[199.08, 223.87], [257.82, 222.08], [271.69, 233.34], [199.53, 236.07], [199.08, 155.63], [257.82, 156.23], [271.69, 152.43], [199.53, 151.51]]}]\n```" - } - ] - }, - { - "id": 297, - "data_source": "Mono3DRefer", - "scene_name": "000157", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1371.89, 300.8], [1622.82, 327.45], [1535.79, 409.7], [1204.71, 352.41], [1371.89, 174.53], [1622.82, 174.88], [1535.79, 175.96], [1204.71, 175.21]]}, {\"category\": \"car\", \"corners_3d\": [[1100.38, 271.67], [1231.55, 286.47], [1022.58, 318.96], [893.71, 295.36], [1100.38, 140.74], [1231.55, 135.93], [1022.58, 125.37], [893.71, 133.04]]}, {\"category\": \"car\", \"corners_3d\": [[832.13, 242.09], [892.44, 248.56], [710.58, 263.08], [658.09, 254.04], [832.13, 174.17], [892.44, 174.29], [710.58, 174.57], [658.09, 174.4]]}, {\"category\": \"car\", \"corners_3d\": [[692.0, 233.91], [728.99, 238.3], [613.78, 245.25], [581.63, 239.92], [692.0, 183.32], [728.99, 184.07], [613.78, 185.26], [581.63, 184.35]]}, {\"category\": \"car\", \"corners_3d\": [[319.61, 230.2], [369.87, 228.14], [406.08, 236.76], [349.22, 239.54], [319.61, 192.46], [369.87, 191.75], [406.08, 194.7], [349.22, 195.65]]}, {\"category\": \"car\", \"corners_3d\": [[219.98, 233.32], [268.2, 231.55], [279.53, 240.92], [223.74, 243.3], [219.98, 194.57], [268.2, 193.94], [279.53, 197.3], [223.74, 198.16]]}, {\"category\": \"truck\", \"corners_3d\": [[199.08, 223.87], [257.82, 222.08], [271.69, 233.34], [199.53, 236.07], [199.08, 155.63], [257.82, 156.23], [271.69, 152.43], [199.53, 151.51]]}]\n```" - } - ] - }, - { - "id": 298, - "data_source": "Mono3DRefer", - "scene_name": "000157", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1371.89, 300.8], [1622.82, 327.45], [1535.79, 409.7], [1204.71, 352.41], [1371.89, 174.53], [1622.82, 174.88], [1535.79, 175.96], [1204.71, 175.21]]}, {\"category\": \"car\", \"corners_3d\": [[1100.38, 271.67], [1231.55, 286.47], [1022.58, 318.96], [893.71, 295.36], [1100.38, 140.74], [1231.55, 135.93], [1022.58, 125.37], [893.71, 133.04]]}, {\"category\": \"car\", \"corners_3d\": [[832.13, 242.09], [892.44, 248.56], [710.58, 263.08], [658.09, 254.04], [832.13, 174.17], [892.44, 174.29], [710.58, 174.57], [658.09, 174.4]]}, {\"category\": \"cyclist\", \"corners_3d\": [[22.01, 355.08], [-51.07, 361.73], [6.56, 324.63], [64.05, 320.31], [22.01, 178.61], [-51.07, 178.82], [6.56, 177.64], [64.05, 177.51]]}, {\"category\": \"car\", \"corners_3d\": [[692.0, 233.91], [728.99, 238.3], [613.78, 245.25], [581.63, 239.92], [692.0, 183.32], [728.99, 184.07], [613.78, 185.26], [581.63, 184.35]]}, {\"category\": \"car\", \"corners_3d\": [[319.61, 230.2], [369.87, 228.14], [406.08, 236.76], [349.22, 239.54], [319.61, 192.46], [369.87, 191.75], [406.08, 194.7], [349.22, 195.65]]}, {\"category\": \"car\", \"corners_3d\": [[219.98, 233.32], [268.2, 231.55], [279.53, 240.92], [223.74, 243.3], [219.98, 194.57], [268.2, 193.94], [279.53, 197.3], [223.74, 198.16]]}, {\"category\": \"truck\", \"corners_3d\": [[199.08, 223.87], [257.82, 222.08], [271.69, 233.34], [199.53, 236.07], [199.08, 155.63], [257.82, 156.23], [271.69, 152.43], [199.53, 151.51]]}]\n```" - } - ] - }, - { - "id": 299, - "data_source": "Mono3DRefer", - "scene_name": "000160", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[540.47, 226.34], [494.19, 226.31], [507.96, 220.28], [549.02, 220.3], [540.47, 184.0], [494.19, 183.99], [507.96, 182.74], [549.02, 182.74]]}, {\"category\": \"car\", \"corners_3d\": [[577.91, 200.67], [558.76, 200.65], [562.76, 198.93], [580.73, 198.95], [577.91, 180.98], [558.76, 180.97], [562.76, 180.47], [580.73, 180.48]]}, {\"category\": \"car\", \"corners_3d\": [[363.62, 203.64], [343.84, 203.63], [361.28, 201.66], [379.81, 201.67], [363.62, 179.51], [343.84, 179.51], [361.28, 179.08], [379.81, 179.08]]}, {\"category\": \"car\", \"corners_3d\": [[594.25, 214.85], [629.73, 214.85], [632.8, 221.45], [591.76, 221.45], [594.25, 179.59], [629.73, 179.59], [632.8, 180.65], [591.76, 180.65]]}]\n```" - } - ] - }, - { - "id": 300, - "data_source": "Mono3DRefer", - "scene_name": "000162", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[79.71, 214.77], [121.61, 214.78], [51.45, 220.8], [3.53, 220.79], [79.71, 174.05], [121.61, 174.05], [51.45, 174.22], [3.53, 174.22]]}, {\"category\": \"car\", \"corners_3d\": [[180.7, 206.22], [207.65, 206.22], [165.18, 209.74], [135.38, 209.74], [180.7, 176.15], [207.65, 176.15], [165.18, 176.49], [135.38, 176.49]]}, {\"category\": \"car\", \"corners_3d\": [[201.69, 193.82], [223.22, 194.05], [182.0, 195.5], [159.49, 195.24], [201.69, 167.68], [223.22, 167.62], [182.0, 167.27], [159.49, 167.33]]}, {\"category\": \"car\", \"corners_3d\": [[578.39, 220.56], [626.37, 220.5], [633.26, 231.2], [574.5, 231.29], [578.39, 172.84], [626.37, 172.84], [633.26, 172.84], [574.5, 172.84]]}]\n```" - } - ] - }, - { - "id": 301, - "data_source": "Mono3DRefer", - "scene_name": "000162", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[760.66, 199.29], [766.02, 199.29], [771.58, 200.23], [766.04, 200.23], [760.66, 173.48], [766.02, 173.48], [771.58, 173.5], [766.04, 173.5]]}]\n```" - } - ] - }, - { - "id": 302, - "data_source": "Mono3DRefer", - "scene_name": "000162", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[79.71, 214.77], [121.61, 214.78], [51.45, 220.8], [3.53, 220.79], [79.71, 174.05], [121.61, 174.05], [51.45, 174.22], [3.53, 174.22]]}, {\"category\": \"car\", \"corners_3d\": [[180.7, 206.22], [207.65, 206.22], [165.18, 209.74], [135.38, 209.74], [180.7, 176.15], [207.65, 176.15], [165.18, 176.49], [135.38, 176.49]]}, {\"category\": \"car\", \"corners_3d\": [[201.69, 193.82], [223.22, 194.05], [182.0, 195.5], [159.49, 195.24], [201.69, 167.68], [223.22, 167.62], [182.0, 167.27], [159.49, 167.33]]}, {\"category\": \"car\", \"corners_3d\": [[578.39, 220.56], [626.37, 220.5], [633.26, 231.2], [574.5, 231.29], [578.39, 172.84], [626.37, 172.84], [633.26, 172.84], [574.5, 172.84]]}, {\"category\": \"cyclist\", \"corners_3d\": [[760.66, 199.29], [766.02, 199.29], [771.58, 200.23], [766.04, 200.23], [760.66, 173.48], [766.02, 173.48], [771.58, 173.5], [766.04, 173.5]]}]\n```" - } - ] - }, - { - "id": 303, - "data_source": "Mono3DRefer", - "scene_name": "000163", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[258.87, 247.49], [292.84, 240.57], [438.21, 240.87], [419.01, 247.85], [258.87, 166.29], [292.84, 166.9], [438.21, 166.87], [419.01, 166.26]]}, {\"category\": \"car\", \"corners_3d\": [[322.44, 229.89], [344.33, 225.67], [472.77, 225.78], [461.13, 230.02], [322.44, 172.5], [344.33, 172.53], [472.77, 172.53], [461.13, 172.5]]}, {\"category\": \"car\", \"corners_3d\": [[805.29, 194.48], [843.73, 194.55], [871.39, 197.71], [827.28, 197.62], [805.29, 155.77], [843.73, 155.71], [871.39, 153.22], [827.28, 153.29]]}, {\"category\": \"car\", \"corners_3d\": [[857.69, 191.32], [887.33, 191.34], [917.0, 193.43], [883.99, 193.41], [857.69, 163.23], [887.33, 163.22], [917.0, 162.14], [883.99, 162.15]]}, {\"category\": \"car\", \"corners_3d\": [[375.23, 216.68], [388.35, 214.22], [495.58, 214.21], [488.86, 216.67], [375.23, 175.88], [388.35, 175.71], [495.58, 175.71], [488.86, 175.88]]}, {\"category\": \"car\", \"corners_3d\": [[399.06, 213.97], [408.88, 211.82], [492.44, 211.68], [487.25, 213.81], [399.06, 176.11], [408.88, 175.93], [492.44, 175.92], [487.25, 176.09]]}, {\"category\": \"car\", \"corners_3d\": [[404.77, 208.42], [416.08, 206.53], [515.89, 206.59], [510.16, 208.48], [404.77, 168.23], [416.08, 168.47], [515.89, 168.46], [510.16, 168.22]]}]\n```" - } - ] - }, - { - "id": 304, - "data_source": "Mono3DRefer", - "scene_name": "000165", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[357.39, 234.26], [413.4, 230.92], [485.95, 239.79], [425.68, 244.27], [357.39, 166.14], [413.4, 166.51], [485.95, 165.54], [425.68, 165.05]]}, {\"category\": \"car\", \"corners_3d\": [[412.39, 218.24], [446.44, 216.2], [511.86, 219.91], [478.09, 222.32], [412.39, 177.46], [446.44, 177.26], [511.86, 177.64], [478.09, 177.88]]}, {\"category\": \"car\", \"corners_3d\": [[443.62, 212.76], [475.12, 211.17], [536.32, 214.51], [504.73, 216.4], [443.62, 168.95], [475.12, 169.11], [536.32, 168.78], [504.73, 168.6]]}, {\"category\": \"car\", \"corners_3d\": [[586.66, 201.94], [617.61, 201.95], [617.64, 204.93], [583.53, 204.91], [586.66, 173.77], [617.61, 173.77], [617.64, 173.87], [583.53, 173.87]]}, {\"category\": \"car\", \"corners_3d\": [[585.97, 217.01], [633.21, 217.01], [636.69, 223.66], [582.33, 223.66], [585.97, 178.99], [633.21, 178.99], [636.69, 179.92], [582.33, 179.92]]}, {\"category\": \"car\", \"corners_3d\": [[-357.14, 385.86], [-66.45, 341.97], [164.45, 420.59], [-219.98, 527.72], [-357.14, 191.93], [-66.45, 188.0], [164.45, 195.04], [-219.98, 204.63]]}, {\"category\": \"car\", \"corners_3d\": [[287.58, 246.58], [358.85, 240.86], [471.37, 253.19], [397.22, 261.29], [287.58, 166.96], [358.85, 167.41], [471.37, 166.43], [397.22, 165.78]]}]\n```" - } - ] - }, - { - "id": 305, - "data_source": "Mono3DRefer", - "scene_name": "000165", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[70.9, 237.03], [-28.12, 237.13], [156.63, 218.28], [226.44, 218.23], [70.9, 118.21], [-28.12, 118.13], [156.63, 134.18], [226.44, 134.22]]}]\n```" - } - ] - }, - { - "id": 306, - "data_source": "Mono3DRefer", - "scene_name": "000165", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[653.21, 201.88], [695.64, 201.85], [715.06, 207.49], [664.4, 207.53], [653.21, 153.22], [695.64, 153.24], [715.06, 149.43], [664.4, 149.4]]}, {\"category\": \"van\", \"corners_3d\": [[-5.01, 309.88], [151.89, 289.76], [374.93, 327.59], [208.8, 364.92], [-5.01, 147.34], [151.89, 151.09], [374.93, 144.05], [208.8, 137.1]]}]\n```" - } - ] - }, - { - "id": 307, - "data_source": "Mono3DRefer", - "scene_name": "000165", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[357.39, 234.26], [413.4, 230.92], [485.95, 239.79], [425.68, 244.27], [357.39, 166.14], [413.4, 166.51], [485.95, 165.54], [425.68, 165.05]]}, {\"category\": \"car\", \"corners_3d\": [[412.39, 218.24], [446.44, 216.2], [511.86, 219.91], [478.09, 222.32], [412.39, 177.46], [446.44, 177.26], [511.86, 177.64], [478.09, 177.88]]}, {\"category\": \"car\", \"corners_3d\": [[443.62, 212.76], [475.12, 211.17], [536.32, 214.51], [504.73, 216.4], [443.62, 168.95], [475.12, 169.11], [536.32, 168.78], [504.73, 168.6]]}, {\"category\": \"car\", \"corners_3d\": [[586.66, 201.94], [617.61, 201.95], [617.64, 204.93], [583.53, 204.91], [586.66, 173.77], [617.61, 173.77], [617.64, 173.87], [583.53, 173.87]]}, {\"category\": \"truck\", \"corners_3d\": [[70.9, 237.03], [-28.12, 237.13], [156.63, 218.28], [226.44, 218.23], [70.9, 118.21], [-28.12, 118.13], [156.63, 134.18], [226.44, 134.22]]}, {\"category\": \"car\", \"corners_3d\": [[585.97, 217.01], [633.21, 217.01], [636.69, 223.66], [582.33, 223.66], [585.97, 178.99], [633.21, 178.99], [636.69, 179.92], [582.33, 179.92]]}, {\"category\": \"car\", \"corners_3d\": [[-357.14, 385.86], [-66.45, 341.97], [164.45, 420.59], [-219.98, 527.72], [-357.14, 191.93], [-66.45, 188.0], [164.45, 195.04], [-219.98, 204.63]]}, {\"category\": \"car\", \"corners_3d\": [[287.58, 246.58], [358.85, 240.86], [471.37, 253.19], [397.22, 261.29], [287.58, 166.96], [358.85, 167.41], [471.37, 166.43], [397.22, 165.78]]}]\n```" - } - ] - }, - { - "id": 308, - "data_source": "Mono3DRefer", - "scene_name": "000165", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[357.39, 234.26], [413.4, 230.92], [485.95, 239.79], [425.68, 244.27], [357.39, 166.14], [413.4, 166.51], [485.95, 165.54], [425.68, 165.05]]}, {\"category\": \"car\", \"corners_3d\": [[412.39, 218.24], [446.44, 216.2], [511.86, 219.91], [478.09, 222.32], [412.39, 177.46], [446.44, 177.26], [511.86, 177.64], [478.09, 177.88]]}, {\"category\": \"car\", \"corners_3d\": [[443.62, 212.76], [475.12, 211.17], [536.32, 214.51], [504.73, 216.4], [443.62, 168.95], [475.12, 169.11], [536.32, 168.78], [504.73, 168.6]]}, {\"category\": \"van\", \"corners_3d\": [[653.21, 201.88], [695.64, 201.85], [715.06, 207.49], [664.4, 207.53], [653.21, 153.22], [695.64, 153.24], [715.06, 149.43], [664.4, 149.4]]}, {\"category\": \"car\", \"corners_3d\": [[586.66, 201.94], [617.61, 201.95], [617.64, 204.93], [583.53, 204.91], [586.66, 173.77], [617.61, 173.77], [617.64, 173.87], [583.53, 173.87]]}, {\"category\": \"van\", \"corners_3d\": [[-5.01, 309.88], [151.89, 289.76], [374.93, 327.59], [208.8, 364.92], [-5.01, 147.34], [151.89, 151.09], [374.93, 144.05], [208.8, 137.1]]}, {\"category\": \"car\", \"corners_3d\": [[585.97, 217.01], [633.21, 217.01], [636.69, 223.66], [582.33, 223.66], [585.97, 178.99], [633.21, 178.99], [636.69, 179.92], [582.33, 179.92]]}, {\"category\": \"car\", \"corners_3d\": [[-357.14, 385.86], [-66.45, 341.97], [164.45, 420.59], [-219.98, 527.72], [-357.14, 191.93], [-66.45, 188.0], [164.45, 195.04], [-219.98, 204.63]]}, {\"category\": \"car\", \"corners_3d\": [[287.58, 246.58], [358.85, 240.86], [471.37, 253.19], [397.22, 261.29], [287.58, 166.96], [358.85, 167.41], [471.37, 166.43], [397.22, 165.78]]}]\n```" - } - ] - }, - { - "id": 309, - "data_source": "Mono3DRefer", - "scene_name": "000165", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[653.21, 201.88], [695.64, 201.85], [715.06, 207.49], [664.4, 207.53], [653.21, 153.22], [695.64, 153.24], [715.06, 149.43], [664.4, 149.4]]}, {\"category\": \"van\", \"corners_3d\": [[-5.01, 309.88], [151.89, 289.76], [374.93, 327.59], [208.8, 364.92], [-5.01, 147.34], [151.89, 151.09], [374.93, 144.05], [208.8, 137.1]]}, {\"category\": \"truck\", \"corners_3d\": [[70.9, 237.03], [-28.12, 237.13], [156.63, 218.28], [226.44, 218.23], [70.9, 118.21], [-28.12, 118.13], [156.63, 134.18], [226.44, 134.22]]}]\n```" - } - ] - }, - { - "id": 310, - "data_source": "Mono3DRefer", - "scene_name": "000165", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[357.39, 234.26], [413.4, 230.92], [485.95, 239.79], [425.68, 244.27], [357.39, 166.14], [413.4, 166.51], [485.95, 165.54], [425.68, 165.05]]}, {\"category\": \"car\", \"corners_3d\": [[412.39, 218.24], [446.44, 216.2], [511.86, 219.91], [478.09, 222.32], [412.39, 177.46], [446.44, 177.26], [511.86, 177.64], [478.09, 177.88]]}, {\"category\": \"car\", \"corners_3d\": [[443.62, 212.76], [475.12, 211.17], [536.32, 214.51], [504.73, 216.4], [443.62, 168.95], [475.12, 169.11], [536.32, 168.78], [504.73, 168.6]]}, {\"category\": \"van\", \"corners_3d\": [[653.21, 201.88], [695.64, 201.85], [715.06, 207.49], [664.4, 207.53], [653.21, 153.22], [695.64, 153.24], [715.06, 149.43], [664.4, 149.4]]}, {\"category\": \"car\", \"corners_3d\": [[586.66, 201.94], [617.61, 201.95], [617.64, 204.93], [583.53, 204.91], [586.66, 173.77], [617.61, 173.77], [617.64, 173.87], [583.53, 173.87]]}, {\"category\": \"van\", \"corners_3d\": [[-5.01, 309.88], [151.89, 289.76], [374.93, 327.59], [208.8, 364.92], [-5.01, 147.34], [151.89, 151.09], [374.93, 144.05], [208.8, 137.1]]}, {\"category\": \"truck\", \"corners_3d\": [[70.9, 237.03], [-28.12, 237.13], [156.63, 218.28], [226.44, 218.23], [70.9, 118.21], [-28.12, 118.13], [156.63, 134.18], [226.44, 134.22]]}, {\"category\": \"car\", \"corners_3d\": [[585.97, 217.01], [633.21, 217.01], [636.69, 223.66], [582.33, 223.66], [585.97, 178.99], [633.21, 178.99], [636.69, 179.92], [582.33, 179.92]]}, {\"category\": \"car\", \"corners_3d\": [[-357.14, 385.86], [-66.45, 341.97], [164.45, 420.59], [-219.98, 527.72], [-357.14, 191.93], [-66.45, 188.0], [164.45, 195.04], [-219.98, 204.63]]}, {\"category\": \"car\", \"corners_3d\": [[287.58, 246.58], [358.85, 240.86], [471.37, 253.19], [397.22, 261.29], [287.58, 166.96], [358.85, 167.41], [471.37, 166.43], [397.22, 165.78]]}]\n```" - } - ] - }, - { - "id": 311, - "data_source": "Mono3DRefer", - "scene_name": "000166", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-32.73, 222.47], [-89.54, 222.39], [-5.64, 216.57], [44.59, 216.62], [-32.73, 175.71], [-89.54, 175.7], [-5.64, 175.37], [44.59, 175.37]]}, {\"category\": \"car\", \"corners_3d\": [[318.45, 190.94], [339.32, 190.94], [319.04, 192.26], [296.66, 192.25], [318.45, 173.22], [339.32, 173.22], [319.04, 173.25], [296.66, 173.25]]}, {\"category\": \"car\", \"corners_3d\": [[253.68, 188.81], [275.45, 188.82], [246.01, 190.19], [222.37, 190.19], [253.68, 158.82], [275.45, 158.82], [246.01, 157.61], [222.37, 157.61]]}, {\"category\": \"car\", \"corners_3d\": [[576.15, 209.95], [615.16, 209.93], [617.31, 216.43], [571.45, 216.45], [576.15, 171.17], [615.16, 171.17], [617.31, 170.88], [571.45, 170.87]]}]\n```" - } - ] - }, - { - "id": 312, - "data_source": "Mono3DRefer", - "scene_name": "000167", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[938.13, 261.58], [983.69, 261.42], [1014.29, 268.27], [965.26, 268.47], [938.13, 153.18], [983.69, 153.22], [1014.29, 151.7], [965.26, 151.65]]}]\n```" - } - ] - }, - { - "id": 313, - "data_source": "Mono3DRefer", - "scene_name": "000167", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[707.75, 213.19], [728.44, 213.24], [735.33, 216.35], [713.03, 216.3], [707.75, 162.42], [728.44, 162.41], [735.33, 161.61], [713.03, 161.62]]}, {\"category\": \"cyclist\", \"corners_3d\": [[683.75, 200.61], [692.77, 200.63], [694.16, 201.84], [684.75, 201.82], [683.75, 169.57], [692.77, 169.57], [694.16, 169.43], [684.75, 169.43]]}]\n```" - } - ] - }, - { - "id": 314, - "data_source": "Mono3DRefer", - "scene_name": "000167", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[938.13, 261.58], [983.69, 261.42], [1014.29, 268.27], [965.26, 268.47], [938.13, 153.18], [983.69, 153.22], [1014.29, 151.7], [965.26, 151.65]]}, {\"category\": \"cyclist\", \"corners_3d\": [[707.75, 213.19], [728.44, 213.24], [735.33, 216.35], [713.03, 216.3], [707.75, 162.42], [728.44, 162.41], [735.33, 161.61], [713.03, 161.62]]}, {\"category\": \"cyclist\", \"corners_3d\": [[683.75, 200.61], [692.77, 200.63], [694.16, 201.84], [684.75, 201.82], [683.75, 169.57], [692.77, 169.57], [694.16, 169.43], [684.75, 169.43]]}]\n```" - } - ] - }, - { - "id": 315, - "data_source": "Mono3DRefer", - "scene_name": "000168", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[560.35, 196.42], [581.7, 196.36], [585.47, 198.78], [561.93, 198.85], [560.35, 166.99], [581.7, 167.0], [585.47, 166.4], [561.93, 166.38]]}]\n```" - } - ] - }, - { - "id": 316, - "data_source": "Mono3DRefer", - "scene_name": "000168", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[554.56, 215.11], [593.61, 214.95], [598.25, 220.68], [553.87, 220.89], [554.56, 184.33], [593.61, 184.29], [598.25, 185.84], [553.87, 185.9]]}, {\"category\": \"car\", \"corners_3d\": [[509.1, 207.64], [537.55, 207.59], [533.44, 210.85], [502.31, 210.92], [509.1, 183.81], [537.55, 183.79], [533.44, 184.82], [502.31, 184.84]]}, {\"category\": \"car\", \"corners_3d\": [[856.15, 223.55], [838.15, 221.51], [901.13, 220.22], [921.21, 222.15], [856.15, 191.45], [838.15, 190.71], [901.13, 190.23], [921.21, 190.94]]}, {\"category\": \"car\", \"corners_3d\": [[987.78, 226.03], [956.1, 223.6], [1013.98, 221.35], [1046.84, 223.56], [987.78, 191.37], [956.1, 190.53], [1013.98, 189.74], [1046.84, 190.51]]}]\n```" - } - ] - }, - { - "id": 317, - "data_source": "Mono3DRefer", - "scene_name": "000168", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[554.56, 215.11], [593.61, 214.95], [598.25, 220.68], [553.87, 220.89], [554.56, 184.33], [593.61, 184.29], [598.25, 185.84], [553.87, 185.9]]}, {\"category\": \"car\", \"corners_3d\": [[509.1, 207.64], [537.55, 207.59], [533.44, 210.85], [502.31, 210.92], [509.1, 183.81], [537.55, 183.79], [533.44, 184.82], [502.31, 184.84]]}, {\"category\": \"car\", \"corners_3d\": [[856.15, 223.55], [838.15, 221.51], [901.13, 220.22], [921.21, 222.15], [856.15, 191.45], [838.15, 190.71], [901.13, 190.23], [921.21, 190.94]]}, {\"category\": \"car\", \"corners_3d\": [[987.78, 226.03], [956.1, 223.6], [1013.98, 221.35], [1046.84, 223.56], [987.78, 191.37], [956.1, 190.53], [1013.98, 189.74], [1046.84, 190.51]]}, {\"category\": \"van\", \"corners_3d\": [[560.35, 196.42], [581.7, 196.36], [585.47, 198.78], [561.93, 198.85], [560.35, 166.99], [581.7, 167.0], [585.47, 166.4], [561.93, 166.38]]}]\n```" - } - ] - }, - { - "id": 318, - "data_source": "Mono3DRefer", - "scene_name": "000172", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[584.56, 201.73], [616.59, 201.72], [618.11, 204.92], [582.53, 204.94], [584.56, 171.47], [616.59, 171.47], [618.11, 171.32], [582.53, 171.32]]}, {\"category\": \"car\", \"corners_3d\": [[599.14, 194.11], [625.76, 194.11], [626.63, 196.33], [597.24, 196.32], [599.14, 166.23], [625.76, 166.23], [626.63, 165.54], [597.24, 165.55]]}, {\"category\": \"car\", \"corners_3d\": [[823.41, 338.97], [985.92, 337.87], [1525.93, 562.08], [1147.69, 568.27], [823.41, 180.27], [985.92, 180.22], [1525.93, 190.23], [1147.69, 190.51]]}, {\"category\": \"car\", \"corners_3d\": [[728.31, 268.94], [821.1, 269.44], [889.68, 305.82], [761.7, 304.88], [728.31, 173.45], [821.1, 173.45], [889.68, 173.68], [761.7, 173.68]]}, {\"category\": \"car\", \"corners_3d\": [[709.3, 237.07], [776.11, 237.25], [814.19, 254.24], [729.68, 253.95], [709.3, 171.13], [776.11, 171.12], [814.19, 170.66], [729.68, 170.67]]}, {\"category\": \"car\", \"corners_3d\": [[658.61, 207.43], [696.94, 207.41], [711.68, 212.83], [667.35, 212.85], [658.61, 166.21], [696.94, 166.22], [711.68, 165.18], [667.35, 165.17]]}, {\"category\": \"car\", \"corners_3d\": [[106.16, 304.13], [-8.88, 304.35], [170.71, 265.68], [251.74, 265.57], [106.16, 195.19], [-8.88, 195.23], [170.71, 188.65], [251.74, 188.63]]}, {\"category\": \"car\", \"corners_3d\": [[324.59, 244.25], [257.61, 244.25], [321.35, 231.3], [376.17, 231.3], [324.59, 182.98], [257.61, 182.98], [321.35, 181.15], [376.17, 181.15]]}, {\"category\": \"car\", \"corners_3d\": [[637.74, 189.69], [658.27, 189.69], [661.5, 190.83], [639.59, 190.82], [637.74, 168.64], [658.27, 168.64], [661.5, 168.35], [639.59, 168.36]]}]\n```" - } - ] - }, - { - "id": 319, - "data_source": "Mono3DRefer", - "scene_name": "000176", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[569.11, 243.77], [501.77, 243.38], [530.72, 229.82], [585.19, 230.08], [569.11, 165.46], [501.77, 165.5], [530.72, 166.91], [585.19, 166.89]]}, {\"category\": \"car\", \"corners_3d\": [[605.59, 226.56], [554.28, 226.22], [572.57, 218.07], [616.11, 218.32], [605.59, 170.39], [554.28, 170.4], [572.57, 170.78], [616.11, 170.77]]}, {\"category\": \"car\", \"corners_3d\": [[654.13, 207.48], [625.01, 207.07], [642.79, 204.02], [669.47, 204.36], [654.13, 179.51], [625.01, 179.44], [642.79, 178.85], [669.47, 178.91]]}, {\"category\": \"car\", \"corners_3d\": [[845.85, 406.24], [1060.61, 406.29], [2136.34, 963.74], [1408.3, 963.1], [845.85, 207.69], [1060.61, 207.69], [2136.34, 290.89], [1408.3, 290.8]]}, {\"category\": \"car\", \"corners_3d\": [[516.39, 274.82], [426.77, 274.45], [472.94, 251.44], [542.34, 251.66], [516.39, 183.91], [426.77, 183.87], [472.94, 181.37], [542.34, 181.4]]}]\n```" - } - ] - }, - { - "id": 320, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[486.16, 221.05], [541.48, 221.74], [488.78, 234.63], [420.09, 233.52], [486.16, 151.7], [541.48, 151.4], [488.78, 145.74], [420.09, 146.23]]}]\n```" - } - ] - }, - { - "id": 321, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[348.21, 211.07], [365.49, 209.4], [438.56, 210.12], [424.34, 211.87], [348.21, 181.16], [365.49, 180.79], [438.56, 180.95], [424.34, 181.33]]}]\n```" - } - ] - }, - { - "id": 322, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[92.7, 336.67], [36.58, 336.56], [154.27, 303.32], [199.06, 303.39], [92.7, 167.07], [36.58, 167.07], [154.27, 168.25], [199.06, 168.25]]}, {\"category\": \"cyclist\", \"corners_3d\": [[668.69, 223.37], [682.9, 224.07], [653.53, 227.31], [638.87, 226.52], [668.69, 172.56], [682.9, 172.56], [653.53, 172.54], [638.87, 172.55]]}]\n```" - } - ] - }, - { - "id": 323, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1029.21, 247.64], [1013.84, 245.17], [1041.92, 244.99], [1058.22, 247.45], [1029.21, 164.91], [1013.84, 165.17], [1041.92, 165.19], [1058.22, 164.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[923.44, 234.1], [947.18, 234.12], [963.11, 237.08], [938.22, 237.06], [923.44, 166.43], [947.18, 166.43], [963.11, 166.12], [938.22, 166.12]]}]\n```" - } - ] - }, - { - "id": 324, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[842.78, 230.25], [836.71, 228.75], [880.79, 228.75], [888.05, 230.25], [842.78, 179.3], [836.71, 179.13], [880.79, 179.14], [888.05, 179.3]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[828.91, 227.79], [824.23, 226.48], [870.46, 226.59], [876.29, 227.9], [828.91, 183.41], [824.23, 183.16], [870.46, 183.18], [876.29, 183.43]]}]\n```" - } - ] - }, - { - "id": 325, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[486.16, 221.05], [541.48, 221.74], [488.78, 234.63], [420.09, 233.52], [486.16, 151.7], [541.48, 151.4], [488.78, 145.74], [420.09, 146.23]]}, {\"category\": \"car\", \"corners_3d\": [[348.21, 211.07], [365.49, 209.4], [438.56, 210.12], [424.34, 211.87], [348.21, 181.16], [365.49, 180.79], [438.56, 180.95], [424.34, 181.33]]}]\n```" - } - ] - }, - { - "id": 326, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[92.7, 336.67], [36.58, 336.56], [154.27, 303.32], [199.06, 303.39], [92.7, 167.07], [36.58, 167.07], [154.27, 168.25], [199.06, 168.25]]}, {\"category\": \"van\", \"corners_3d\": [[486.16, 221.05], [541.48, 221.74], [488.78, 234.63], [420.09, 233.52], [486.16, 151.7], [541.48, 151.4], [488.78, 145.74], [420.09, 146.23]]}, {\"category\": \"cyclist\", \"corners_3d\": [[668.69, 223.37], [682.9, 224.07], [653.53, 227.31], [638.87, 226.52], [668.69, 172.56], [682.9, 172.56], [653.53, 172.54], [638.87, 172.55]]}]\n```" - } - ] - }, - { - "id": 327, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[486.16, 221.05], [541.48, 221.74], [488.78, 234.63], [420.09, 233.52], [486.16, 151.7], [541.48, 151.4], [488.78, 145.74], [420.09, 146.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1029.21, 247.64], [1013.84, 245.17], [1041.92, 244.99], [1058.22, 247.45], [1029.21, 164.91], [1013.84, 165.17], [1041.92, 165.19], [1058.22, 164.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[923.44, 234.1], [947.18, 234.12], [963.11, 237.08], [938.22, 237.06], [923.44, 166.43], [947.18, 166.43], [963.11, 166.12], [938.22, 166.12]]}]\n```" - } - ] - }, - { - "id": 328, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[486.16, 221.05], [541.48, 221.74], [488.78, 234.63], [420.09, 233.52], [486.16, 151.7], [541.48, 151.4], [488.78, 145.74], [420.09, 146.23]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[842.78, 230.25], [836.71, 228.75], [880.79, 228.75], [888.05, 230.25], [842.78, 179.3], [836.71, 179.13], [880.79, 179.14], [888.05, 179.3]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[828.91, 227.79], [824.23, 226.48], [870.46, 226.59], [876.29, 227.9], [828.91, 183.41], [824.23, 183.16], [870.46, 183.18], [876.29, 183.43]]}]\n```" - } - ] - }, - { - "id": 329, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[92.7, 336.67], [36.58, 336.56], [154.27, 303.32], [199.06, 303.39], [92.7, 167.07], [36.58, 167.07], [154.27, 168.25], [199.06, 168.25]]}, {\"category\": \"cyclist\", \"corners_3d\": [[668.69, 223.37], [682.9, 224.07], [653.53, 227.31], [638.87, 226.52], [668.69, 172.56], [682.9, 172.56], [653.53, 172.54], [638.87, 172.55]]}, {\"category\": \"car\", \"corners_3d\": [[348.21, 211.07], [365.49, 209.4], [438.56, 210.12], [424.34, 211.87], [348.21, 181.16], [365.49, 180.79], [438.56, 180.95], [424.34, 181.33]]}]\n```" - } - ] - }, - { - "id": 330, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1029.21, 247.64], [1013.84, 245.17], [1041.92, 244.99], [1058.22, 247.45], [1029.21, 164.91], [1013.84, 165.17], [1041.92, 165.19], [1058.22, 164.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[923.44, 234.1], [947.18, 234.12], [963.11, 237.08], [938.22, 237.06], [923.44, 166.43], [947.18, 166.43], [963.11, 166.12], [938.22, 166.12]]}, {\"category\": \"car\", \"corners_3d\": [[348.21, 211.07], [365.49, 209.4], [438.56, 210.12], [424.34, 211.87], [348.21, 181.16], [365.49, 180.79], [438.56, 180.95], [424.34, 181.33]]}]\n```" - } - ] - }, - { - "id": 331, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[842.78, 230.25], [836.71, 228.75], [880.79, 228.75], [888.05, 230.25], [842.78, 179.3], [836.71, 179.13], [880.79, 179.14], [888.05, 179.3]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[828.91, 227.79], [824.23, 226.48], [870.46, 226.59], [876.29, 227.9], [828.91, 183.41], [824.23, 183.16], [870.46, 183.18], [876.29, 183.43]]}, {\"category\": \"car\", \"corners_3d\": [[348.21, 211.07], [365.49, 209.4], [438.56, 210.12], [424.34, 211.87], [348.21, 181.16], [365.49, 180.79], [438.56, 180.95], [424.34, 181.33]]}]\n```" - } - ] - }, - { - "id": 332, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[92.7, 336.67], [36.58, 336.56], [154.27, 303.32], [199.06, 303.39], [92.7, 167.07], [36.58, 167.07], [154.27, 168.25], [199.06, 168.25]]}, {\"category\": \"cyclist\", \"corners_3d\": [[668.69, 223.37], [682.9, 224.07], [653.53, 227.31], [638.87, 226.52], [668.69, 172.56], [682.9, 172.56], [653.53, 172.54], [638.87, 172.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1029.21, 247.64], [1013.84, 245.17], [1041.92, 244.99], [1058.22, 247.45], [1029.21, 164.91], [1013.84, 165.17], [1041.92, 165.19], [1058.22, 164.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[923.44, 234.1], [947.18, 234.12], [963.11, 237.08], [938.22, 237.06], [923.44, 166.43], [947.18, 166.43], [963.11, 166.12], [938.22, 166.12]]}]\n```" - } - ] - }, - { - "id": 333, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[92.7, 336.67], [36.58, 336.56], [154.27, 303.32], [199.06, 303.39], [92.7, 167.07], [36.58, 167.07], [154.27, 168.25], [199.06, 168.25]]}, {\"category\": \"cyclist\", \"corners_3d\": [[668.69, 223.37], [682.9, 224.07], [653.53, 227.31], [638.87, 226.52], [668.69, 172.56], [682.9, 172.56], [653.53, 172.54], [638.87, 172.55]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[842.78, 230.25], [836.71, 228.75], [880.79, 228.75], [888.05, 230.25], [842.78, 179.3], [836.71, 179.13], [880.79, 179.14], [888.05, 179.3]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[828.91, 227.79], [824.23, 226.48], [870.46, 226.59], [876.29, 227.9], [828.91, 183.41], [824.23, 183.16], [870.46, 183.18], [876.29, 183.43]]}]\n```" - } - ] - }, - { - "id": 334, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1029.21, 247.64], [1013.84, 245.17], [1041.92, 244.99], [1058.22, 247.45], [1029.21, 164.91], [1013.84, 165.17], [1041.92, 165.19], [1058.22, 164.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[923.44, 234.1], [947.18, 234.12], [963.11, 237.08], [938.22, 237.06], [923.44, 166.43], [947.18, 166.43], [963.11, 166.12], [938.22, 166.12]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[842.78, 230.25], [836.71, 228.75], [880.79, 228.75], [888.05, 230.25], [842.78, 179.3], [836.71, 179.13], [880.79, 179.14], [888.05, 179.3]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[828.91, 227.79], [824.23, 226.48], [870.46, 226.59], [876.29, 227.9], [828.91, 183.41], [824.23, 183.16], [870.46, 183.18], [876.29, 183.43]]}]\n```" - } - ] - }, - { - "id": 335, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[92.7, 336.67], [36.58, 336.56], [154.27, 303.32], [199.06, 303.39], [92.7, 167.07], [36.58, 167.07], [154.27, 168.25], [199.06, 168.25]]}, {\"category\": \"van\", \"corners_3d\": [[486.16, 221.05], [541.48, 221.74], [488.78, 234.63], [420.09, 233.52], [486.16, 151.7], [541.48, 151.4], [488.78, 145.74], [420.09, 146.23]]}, {\"category\": \"cyclist\", \"corners_3d\": [[668.69, 223.37], [682.9, 224.07], [653.53, 227.31], [638.87, 226.52], [668.69, 172.56], [682.9, 172.56], [653.53, 172.54], [638.87, 172.55]]}, {\"category\": \"car\", \"corners_3d\": [[348.21, 211.07], [365.49, 209.4], [438.56, 210.12], [424.34, 211.87], [348.21, 181.16], [365.49, 180.79], [438.56, 180.95], [424.34, 181.33]]}]\n```" - } - ] - }, - { - "id": 336, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[486.16, 221.05], [541.48, 221.74], [488.78, 234.63], [420.09, 233.52], [486.16, 151.7], [541.48, 151.4], [488.78, 145.74], [420.09, 146.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1029.21, 247.64], [1013.84, 245.17], [1041.92, 244.99], [1058.22, 247.45], [1029.21, 164.91], [1013.84, 165.17], [1041.92, 165.19], [1058.22, 164.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[923.44, 234.1], [947.18, 234.12], [963.11, 237.08], [938.22, 237.06], [923.44, 166.43], [947.18, 166.43], [963.11, 166.12], [938.22, 166.12]]}, {\"category\": \"car\", \"corners_3d\": [[348.21, 211.07], [365.49, 209.4], [438.56, 210.12], [424.34, 211.87], [348.21, 181.16], [365.49, 180.79], [438.56, 180.95], [424.34, 181.33]]}]\n```" - } - ] - }, - { - "id": 337, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[486.16, 221.05], [541.48, 221.74], [488.78, 234.63], [420.09, 233.52], [486.16, 151.7], [541.48, 151.4], [488.78, 145.74], [420.09, 146.23]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[842.78, 230.25], [836.71, 228.75], [880.79, 228.75], [888.05, 230.25], [842.78, 179.3], [836.71, 179.13], [880.79, 179.14], [888.05, 179.3]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[828.91, 227.79], [824.23, 226.48], [870.46, 226.59], [876.29, 227.9], [828.91, 183.41], [824.23, 183.16], [870.46, 183.18], [876.29, 183.43]]}, {\"category\": \"car\", \"corners_3d\": [[348.21, 211.07], [365.49, 209.4], [438.56, 210.12], [424.34, 211.87], [348.21, 181.16], [365.49, 180.79], [438.56, 180.95], [424.34, 181.33]]}]\n```" - } - ] - }, - { - "id": 338, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[92.7, 336.67], [36.58, 336.56], [154.27, 303.32], [199.06, 303.39], [92.7, 167.07], [36.58, 167.07], [154.27, 168.25], [199.06, 168.25]]}, {\"category\": \"van\", \"corners_3d\": [[486.16, 221.05], [541.48, 221.74], [488.78, 234.63], [420.09, 233.52], [486.16, 151.7], [541.48, 151.4], [488.78, 145.74], [420.09, 146.23]]}, {\"category\": \"cyclist\", \"corners_3d\": [[668.69, 223.37], [682.9, 224.07], [653.53, 227.31], [638.87, 226.52], [668.69, 172.56], [682.9, 172.56], [653.53, 172.54], [638.87, 172.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1029.21, 247.64], [1013.84, 245.17], [1041.92, 244.99], [1058.22, 247.45], [1029.21, 164.91], [1013.84, 165.17], [1041.92, 165.19], [1058.22, 164.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[923.44, 234.1], [947.18, 234.12], [963.11, 237.08], [938.22, 237.06], [923.44, 166.43], [947.18, 166.43], [963.11, 166.12], [938.22, 166.12]]}]\n```" - } - ] - }, - { - "id": 339, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[92.7, 336.67], [36.58, 336.56], [154.27, 303.32], [199.06, 303.39], [92.7, 167.07], [36.58, 167.07], [154.27, 168.25], [199.06, 168.25]]}, {\"category\": \"van\", \"corners_3d\": [[486.16, 221.05], [541.48, 221.74], [488.78, 234.63], [420.09, 233.52], [486.16, 151.7], [541.48, 151.4], [488.78, 145.74], [420.09, 146.23]]}, {\"category\": \"cyclist\", \"corners_3d\": [[668.69, 223.37], [682.9, 224.07], [653.53, 227.31], [638.87, 226.52], [668.69, 172.56], [682.9, 172.56], [653.53, 172.54], [638.87, 172.55]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[842.78, 230.25], [836.71, 228.75], [880.79, 228.75], [888.05, 230.25], [842.78, 179.3], [836.71, 179.13], [880.79, 179.14], [888.05, 179.3]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[828.91, 227.79], [824.23, 226.48], [870.46, 226.59], [876.29, 227.9], [828.91, 183.41], [824.23, 183.16], [870.46, 183.18], [876.29, 183.43]]}]\n```" - } - ] - }, - { - "id": 340, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[486.16, 221.05], [541.48, 221.74], [488.78, 234.63], [420.09, 233.52], [486.16, 151.7], [541.48, 151.4], [488.78, 145.74], [420.09, 146.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1029.21, 247.64], [1013.84, 245.17], [1041.92, 244.99], [1058.22, 247.45], [1029.21, 164.91], [1013.84, 165.17], [1041.92, 165.19], [1058.22, 164.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[923.44, 234.1], [947.18, 234.12], [963.11, 237.08], [938.22, 237.06], [923.44, 166.43], [947.18, 166.43], [963.11, 166.12], [938.22, 166.12]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[842.78, 230.25], [836.71, 228.75], [880.79, 228.75], [888.05, 230.25], [842.78, 179.3], [836.71, 179.13], [880.79, 179.14], [888.05, 179.3]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[828.91, 227.79], [824.23, 226.48], [870.46, 226.59], [876.29, 227.9], [828.91, 183.41], [824.23, 183.16], [870.46, 183.18], [876.29, 183.43]]}]\n```" - } - ] - }, - { - "id": 341, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[92.7, 336.67], [36.58, 336.56], [154.27, 303.32], [199.06, 303.39], [92.7, 167.07], [36.58, 167.07], [154.27, 168.25], [199.06, 168.25]]}, {\"category\": \"cyclist\", \"corners_3d\": [[668.69, 223.37], [682.9, 224.07], [653.53, 227.31], [638.87, 226.52], [668.69, 172.56], [682.9, 172.56], [653.53, 172.54], [638.87, 172.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1029.21, 247.64], [1013.84, 245.17], [1041.92, 244.99], [1058.22, 247.45], [1029.21, 164.91], [1013.84, 165.17], [1041.92, 165.19], [1058.22, 164.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[923.44, 234.1], [947.18, 234.12], [963.11, 237.08], [938.22, 237.06], [923.44, 166.43], [947.18, 166.43], [963.11, 166.12], [938.22, 166.12]]}, {\"category\": \"car\", \"corners_3d\": [[348.21, 211.07], [365.49, 209.4], [438.56, 210.12], [424.34, 211.87], [348.21, 181.16], [365.49, 180.79], [438.56, 180.95], [424.34, 181.33]]}]\n```" - } - ] - }, - { - "id": 342, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[92.7, 336.67], [36.58, 336.56], [154.27, 303.32], [199.06, 303.39], [92.7, 167.07], [36.58, 167.07], [154.27, 168.25], [199.06, 168.25]]}, {\"category\": \"cyclist\", \"corners_3d\": [[668.69, 223.37], [682.9, 224.07], [653.53, 227.31], [638.87, 226.52], [668.69, 172.56], [682.9, 172.56], [653.53, 172.54], [638.87, 172.55]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[842.78, 230.25], [836.71, 228.75], [880.79, 228.75], [888.05, 230.25], [842.78, 179.3], [836.71, 179.13], [880.79, 179.14], [888.05, 179.3]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[828.91, 227.79], [824.23, 226.48], [870.46, 226.59], [876.29, 227.9], [828.91, 183.41], [824.23, 183.16], [870.46, 183.18], [876.29, 183.43]]}, {\"category\": \"car\", \"corners_3d\": [[348.21, 211.07], [365.49, 209.4], [438.56, 210.12], [424.34, 211.87], [348.21, 181.16], [365.49, 180.79], [438.56, 180.95], [424.34, 181.33]]}]\n```" - } - ] - }, - { - "id": 343, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1029.21, 247.64], [1013.84, 245.17], [1041.92, 244.99], [1058.22, 247.45], [1029.21, 164.91], [1013.84, 165.17], [1041.92, 165.19], [1058.22, 164.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[923.44, 234.1], [947.18, 234.12], [963.11, 237.08], [938.22, 237.06], [923.44, 166.43], [947.18, 166.43], [963.11, 166.12], [938.22, 166.12]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[842.78, 230.25], [836.71, 228.75], [880.79, 228.75], [888.05, 230.25], [842.78, 179.3], [836.71, 179.13], [880.79, 179.14], [888.05, 179.3]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[828.91, 227.79], [824.23, 226.48], [870.46, 226.59], [876.29, 227.9], [828.91, 183.41], [824.23, 183.16], [870.46, 183.18], [876.29, 183.43]]}, {\"category\": \"car\", \"corners_3d\": [[348.21, 211.07], [365.49, 209.4], [438.56, 210.12], [424.34, 211.87], [348.21, 181.16], [365.49, 180.79], [438.56, 180.95], [424.34, 181.33]]}]\n```" - } - ] - }, - { - "id": 344, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[92.7, 336.67], [36.58, 336.56], [154.27, 303.32], [199.06, 303.39], [92.7, 167.07], [36.58, 167.07], [154.27, 168.25], [199.06, 168.25]]}, {\"category\": \"cyclist\", \"corners_3d\": [[668.69, 223.37], [682.9, 224.07], [653.53, 227.31], [638.87, 226.52], [668.69, 172.56], [682.9, 172.56], [653.53, 172.54], [638.87, 172.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1029.21, 247.64], [1013.84, 245.17], [1041.92, 244.99], [1058.22, 247.45], [1029.21, 164.91], [1013.84, 165.17], [1041.92, 165.19], [1058.22, 164.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[923.44, 234.1], [947.18, 234.12], [963.11, 237.08], [938.22, 237.06], [923.44, 166.43], [947.18, 166.43], [963.11, 166.12], [938.22, 166.12]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[842.78, 230.25], [836.71, 228.75], [880.79, 228.75], [888.05, 230.25], [842.78, 179.3], [836.71, 179.13], [880.79, 179.14], [888.05, 179.3]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[828.91, 227.79], [824.23, 226.48], [870.46, 226.59], [876.29, 227.9], [828.91, 183.41], [824.23, 183.16], [870.46, 183.18], [876.29, 183.43]]}]\n```" - } - ] - }, - { - "id": 345, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[92.7, 336.67], [36.58, 336.56], [154.27, 303.32], [199.06, 303.39], [92.7, 167.07], [36.58, 167.07], [154.27, 168.25], [199.06, 168.25]]}, {\"category\": \"van\", \"corners_3d\": [[486.16, 221.05], [541.48, 221.74], [488.78, 234.63], [420.09, 233.52], [486.16, 151.7], [541.48, 151.4], [488.78, 145.74], [420.09, 146.23]]}, {\"category\": \"cyclist\", \"corners_3d\": [[668.69, 223.37], [682.9, 224.07], [653.53, 227.31], [638.87, 226.52], [668.69, 172.56], [682.9, 172.56], [653.53, 172.54], [638.87, 172.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1029.21, 247.64], [1013.84, 245.17], [1041.92, 244.99], [1058.22, 247.45], [1029.21, 164.91], [1013.84, 165.17], [1041.92, 165.19], [1058.22, 164.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[923.44, 234.1], [947.18, 234.12], [963.11, 237.08], [938.22, 237.06], [923.44, 166.43], [947.18, 166.43], [963.11, 166.12], [938.22, 166.12]]}, {\"category\": \"car\", \"corners_3d\": [[348.21, 211.07], [365.49, 209.4], [438.56, 210.12], [424.34, 211.87], [348.21, 181.16], [365.49, 180.79], [438.56, 180.95], [424.34, 181.33]]}]\n```" - } - ] - }, - { - "id": 346, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, cyclist, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[92.7, 336.67], [36.58, 336.56], [154.27, 303.32], [199.06, 303.39], [92.7, 167.07], [36.58, 167.07], [154.27, 168.25], [199.06, 168.25]]}, {\"category\": \"van\", \"corners_3d\": [[486.16, 221.05], [541.48, 221.74], [488.78, 234.63], [420.09, 233.52], [486.16, 151.7], [541.48, 151.4], [488.78, 145.74], [420.09, 146.23]]}, {\"category\": \"cyclist\", \"corners_3d\": [[668.69, 223.37], [682.9, 224.07], [653.53, 227.31], [638.87, 226.52], [668.69, 172.56], [682.9, 172.56], [653.53, 172.54], [638.87, 172.55]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[842.78, 230.25], [836.71, 228.75], [880.79, 228.75], [888.05, 230.25], [842.78, 179.3], [836.71, 179.13], [880.79, 179.14], [888.05, 179.3]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[828.91, 227.79], [824.23, 226.48], [870.46, 226.59], [876.29, 227.9], [828.91, 183.41], [824.23, 183.16], [870.46, 183.18], [876.29, 183.43]]}, {\"category\": \"car\", \"corners_3d\": [[348.21, 211.07], [365.49, 209.4], [438.56, 210.12], [424.34, 211.87], [348.21, 181.16], [365.49, 180.79], [438.56, 180.95], [424.34, 181.33]]}]\n```" - } - ] - }, - { - "id": 347, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, pedestrian, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[486.16, 221.05], [541.48, 221.74], [488.78, 234.63], [420.09, 233.52], [486.16, 151.7], [541.48, 151.4], [488.78, 145.74], [420.09, 146.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1029.21, 247.64], [1013.84, 245.17], [1041.92, 244.99], [1058.22, 247.45], [1029.21, 164.91], [1013.84, 165.17], [1041.92, 165.19], [1058.22, 164.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[923.44, 234.1], [947.18, 234.12], [963.11, 237.08], [938.22, 237.06], [923.44, 166.43], [947.18, 166.43], [963.11, 166.12], [938.22, 166.12]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[842.78, 230.25], [836.71, 228.75], [880.79, 228.75], [888.05, 230.25], [842.78, 179.3], [836.71, 179.13], [880.79, 179.14], [888.05, 179.3]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[828.91, 227.79], [824.23, 226.48], [870.46, 226.59], [876.29, 227.9], [828.91, 183.41], [824.23, 183.16], [870.46, 183.18], [876.29, 183.43]]}, {\"category\": \"car\", \"corners_3d\": [[348.21, 211.07], [365.49, 209.4], [438.56, 210.12], [424.34, 211.87], [348.21, 181.16], [365.49, 180.79], [438.56, 180.95], [424.34, 181.33]]}]\n```" - } - ] - }, - { - "id": 348, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, pedestrian, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[92.7, 336.67], [36.58, 336.56], [154.27, 303.32], [199.06, 303.39], [92.7, 167.07], [36.58, 167.07], [154.27, 168.25], [199.06, 168.25]]}, {\"category\": \"van\", \"corners_3d\": [[486.16, 221.05], [541.48, 221.74], [488.78, 234.63], [420.09, 233.52], [486.16, 151.7], [541.48, 151.4], [488.78, 145.74], [420.09, 146.23]]}, {\"category\": \"cyclist\", \"corners_3d\": [[668.69, 223.37], [682.9, 224.07], [653.53, 227.31], [638.87, 226.52], [668.69, 172.56], [682.9, 172.56], [653.53, 172.54], [638.87, 172.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1029.21, 247.64], [1013.84, 245.17], [1041.92, 244.99], [1058.22, 247.45], [1029.21, 164.91], [1013.84, 165.17], [1041.92, 165.19], [1058.22, 164.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[923.44, 234.1], [947.18, 234.12], [963.11, 237.08], [938.22, 237.06], [923.44, 166.43], [947.18, 166.43], [963.11, 166.12], [938.22, 166.12]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[842.78, 230.25], [836.71, 228.75], [880.79, 228.75], [888.05, 230.25], [842.78, 179.3], [836.71, 179.13], [880.79, 179.14], [888.05, 179.3]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[828.91, 227.79], [824.23, 226.48], [870.46, 226.59], [876.29, 227.9], [828.91, 183.41], [824.23, 183.16], [870.46, 183.18], [876.29, 183.43]]}]\n```" - } - ] - }, - { - "id": 349, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[92.7, 336.67], [36.58, 336.56], [154.27, 303.32], [199.06, 303.39], [92.7, 167.07], [36.58, 167.07], [154.27, 168.25], [199.06, 168.25]]}, {\"category\": \"cyclist\", \"corners_3d\": [[668.69, 223.37], [682.9, 224.07], [653.53, 227.31], [638.87, 226.52], [668.69, 172.56], [682.9, 172.56], [653.53, 172.54], [638.87, 172.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1029.21, 247.64], [1013.84, 245.17], [1041.92, 244.99], [1058.22, 247.45], [1029.21, 164.91], [1013.84, 165.17], [1041.92, 165.19], [1058.22, 164.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[923.44, 234.1], [947.18, 234.12], [963.11, 237.08], [938.22, 237.06], [923.44, 166.43], [947.18, 166.43], [963.11, 166.12], [938.22, 166.12]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[842.78, 230.25], [836.71, 228.75], [880.79, 228.75], [888.05, 230.25], [842.78, 179.3], [836.71, 179.13], [880.79, 179.14], [888.05, 179.3]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[828.91, 227.79], [824.23, 226.48], [870.46, 226.59], [876.29, 227.9], [828.91, 183.41], [824.23, 183.16], [870.46, 183.18], [876.29, 183.43]]}, {\"category\": \"car\", \"corners_3d\": [[348.21, 211.07], [365.49, 209.4], [438.56, 210.12], [424.34, 211.87], [348.21, 181.16], [365.49, 180.79], [438.56, 180.95], [424.34, 181.33]]}]\n```" - } - ] - }, - { - "id": 350, - "data_source": "Mono3DRefer", - "scene_name": "000177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, cyclist, pedestrian, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[92.7, 336.67], [36.58, 336.56], [154.27, 303.32], [199.06, 303.39], [92.7, 167.07], [36.58, 167.07], [154.27, 168.25], [199.06, 168.25]]}, {\"category\": \"van\", \"corners_3d\": [[486.16, 221.05], [541.48, 221.74], [488.78, 234.63], [420.09, 233.52], [486.16, 151.7], [541.48, 151.4], [488.78, 145.74], [420.09, 146.23]]}, {\"category\": \"cyclist\", \"corners_3d\": [[668.69, 223.37], [682.9, 224.07], [653.53, 227.31], [638.87, 226.52], [668.69, 172.56], [682.9, 172.56], [653.53, 172.54], [638.87, 172.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1029.21, 247.64], [1013.84, 245.17], [1041.92, 244.99], [1058.22, 247.45], [1029.21, 164.91], [1013.84, 165.17], [1041.92, 165.19], [1058.22, 164.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[923.44, 234.1], [947.18, 234.12], [963.11, 237.08], [938.22, 237.06], [923.44, 166.43], [947.18, 166.43], [963.11, 166.12], [938.22, 166.12]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[842.78, 230.25], [836.71, 228.75], [880.79, 228.75], [888.05, 230.25], [842.78, 179.3], [836.71, 179.13], [880.79, 179.14], [888.05, 179.3]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[828.91, 227.79], [824.23, 226.48], [870.46, 226.59], [876.29, 227.9], [828.91, 183.41], [824.23, 183.16], [870.46, 183.18], [876.29, 183.43]]}, {\"category\": \"car\", \"corners_3d\": [[348.21, 211.07], [365.49, 209.4], [438.56, 210.12], [424.34, 211.87], [348.21, 181.16], [365.49, 180.79], [438.56, 180.95], [424.34, 181.33]]}]\n```" - } - ] - }, - { - "id": 351, - "data_source": "Mono3DRefer", - "scene_name": "000179", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[881.31, 220.65], [809.05, 220.46], [740.01, 201.26], [782.9, 201.33], [881.31, 100.97], [809.05, 101.25], [740.01, 130.13], [782.9, 130.03]]}, {\"category\": \"tram\", \"corners_3d\": [[786.5, 203.68], [741.38, 203.59], [720.63, 196.97], [755.97, 197.03], [786.5, 132.0], [741.38, 132.12], [720.63, 140.89], [755.97, 140.82]]}, {\"category\": \"tram\", \"corners_3d\": [[757.18, 197.05], [721.68, 196.95], [711.54, 191.81], [739.41, 191.88], [757.18, 140.63], [721.68, 140.77], [711.54, 147.61], [739.41, 147.52]]}]\n```" - } - ] - }, - { - "id": 352, - "data_source": "Mono3DRefer", - "scene_name": "000179", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[281.65, 231.03], [266.17, 230.96], [280.17, 228.85], [295.1, 228.92], [281.65, 180.98], [266.17, 180.97], [280.17, 180.67], [295.1, 180.68]]}]\n```" - } - ] - }, - { - "id": 353, - "data_source": "Mono3DRefer", - "scene_name": "000179", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[598.22, 229.48], [653.6, 229.61], [657.39, 239.47], [592.4, 239.29], [598.22, 175.84], [653.6, 175.85], [657.39, 176.37], [592.4, 176.36]]}]\n```" - } - ] - }, - { - "id": 354, - "data_source": "Mono3DRefer", - "scene_name": "000179", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[881.31, 220.65], [809.05, 220.46], [740.01, 201.26], [782.9, 201.33], [881.31, 100.97], [809.05, 101.25], [740.01, 130.13], [782.9, 130.03]]}, {\"category\": \"tram\", \"corners_3d\": [[786.5, 203.68], [741.38, 203.59], [720.63, 196.97], [755.97, 197.03], [786.5, 132.0], [741.38, 132.12], [720.63, 140.89], [755.97, 140.82]]}, {\"category\": \"tram\", \"corners_3d\": [[757.18, 197.05], [721.68, 196.95], [711.54, 191.81], [739.41, 191.88], [757.18, 140.63], [721.68, 140.77], [711.54, 147.61], [739.41, 147.52]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[281.65, 231.03], [266.17, 230.96], [280.17, 228.85], [295.1, 228.92], [281.65, 180.98], [266.17, 180.97], [280.17, 180.67], [295.1, 180.68]]}]\n```" - } - ] - }, - { - "id": 355, - "data_source": "Mono3DRefer", - "scene_name": "000179", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[598.22, 229.48], [653.6, 229.61], [657.39, 239.47], [592.4, 239.29], [598.22, 175.84], [653.6, 175.85], [657.39, 176.37], [592.4, 176.36]]}, {\"category\": \"tram\", \"corners_3d\": [[881.31, 220.65], [809.05, 220.46], [740.01, 201.26], [782.9, 201.33], [881.31, 100.97], [809.05, 101.25], [740.01, 130.13], [782.9, 130.03]]}, {\"category\": \"tram\", \"corners_3d\": [[786.5, 203.68], [741.38, 203.59], [720.63, 196.97], [755.97, 197.03], [786.5, 132.0], [741.38, 132.12], [720.63, 140.89], [755.97, 140.82]]}, {\"category\": \"tram\", \"corners_3d\": [[757.18, 197.05], [721.68, 196.95], [711.54, 191.81], [739.41, 191.88], [757.18, 140.63], [721.68, 140.77], [711.54, 147.61], [739.41, 147.52]]}]\n```" - } - ] - }, - { - "id": 356, - "data_source": "Mono3DRefer", - "scene_name": "000179", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[598.22, 229.48], [653.6, 229.61], [657.39, 239.47], [592.4, 239.29], [598.22, 175.84], [653.6, 175.85], [657.39, 176.37], [592.4, 176.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[281.65, 231.03], [266.17, 230.96], [280.17, 228.85], [295.1, 228.92], [281.65, 180.98], [266.17, 180.97], [280.17, 180.67], [295.1, 180.68]]}]\n```" - } - ] - }, - { - "id": 357, - "data_source": "Mono3DRefer", - "scene_name": "000179", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[598.22, 229.48], [653.6, 229.61], [657.39, 239.47], [592.4, 239.29], [598.22, 175.84], [653.6, 175.85], [657.39, 176.37], [592.4, 176.36]]}, {\"category\": \"tram\", \"corners_3d\": [[881.31, 220.65], [809.05, 220.46], [740.01, 201.26], [782.9, 201.33], [881.31, 100.97], [809.05, 101.25], [740.01, 130.13], [782.9, 130.03]]}, {\"category\": \"tram\", \"corners_3d\": [[786.5, 203.68], [741.38, 203.59], [720.63, 196.97], [755.97, 197.03], [786.5, 132.0], [741.38, 132.12], [720.63, 140.89], [755.97, 140.82]]}, {\"category\": \"tram\", \"corners_3d\": [[757.18, 197.05], [721.68, 196.95], [711.54, 191.81], [739.41, 191.88], [757.18, 140.63], [721.68, 140.77], [711.54, 147.61], [739.41, 147.52]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[281.65, 231.03], [266.17, 230.96], [280.17, 228.85], [295.1, 228.92], [281.65, 180.98], [266.17, 180.97], [280.17, 180.67], [295.1, 180.68]]}]\n```" - } - ] - }, - { - "id": 358, - "data_source": "Mono3DRefer", - "scene_name": "000180", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[562.09, 197.96], [541.15, 197.94], [546.81, 196.35], [566.43, 196.37], [562.09, 181.88], [541.15, 181.88], [546.81, 181.31], [566.43, 181.31]]}, {\"category\": \"car\", \"corners_3d\": [[507.61, 205.3], [503.92, 206.4], [439.82, 206.37], [445.61, 205.28], [507.61, 183.76], [503.92, 184.13], [439.82, 184.12], [445.61, 183.75]]}]\n```" - } - ] - }, - { - "id": 359, - "data_source": "Mono3DRefer", - "scene_name": "000183", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[697.82, 227.02], [769.13, 227.29], [803.36, 242.46], [712.09, 242.02], [697.82, 153.62], [769.13, 153.52], [803.36, 148.13], [712.09, 148.29]]}]\n```" - } - ] - }, - { - "id": 360, - "data_source": "Mono3DRefer", - "scene_name": "000183", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[887.35, 362.54], [1111.76, 365.48], [3017.9, 1154.22], [1802.31, 1082.36], [887.35, 172.81], [1111.76, 172.81], [3017.9, 172.61], [1802.31, 172.63]]}, {\"category\": \"car\", \"corners_3d\": [[51.37, 304.73], [133.82, 286.02], [365.16, 287.19], [320.44, 306.33], [51.37, 172.09], [133.82, 172.2], [365.16, 172.19], [320.44, 172.08]]}, {\"category\": \"car\", \"corners_3d\": [[740.92, 220.71], [692.65, 220.65], [682.71, 213.79], [724.03, 213.83], [740.92, 168.41], [692.65, 168.42], [682.71, 169.05], [724.03, 169.05]]}, {\"category\": \"car\", \"corners_3d\": [[662.84, 197.74], [689.75, 197.76], [696.64, 200.39], [666.88, 200.37], [662.84, 170.06], [689.75, 170.06], [696.64, 169.77], [666.88, 169.77]]}, {\"category\": \"car\", \"corners_3d\": [[577.27, 201.1], [551.56, 201.09], [556.37, 198.99], [580.17, 198.99], [577.27, 176.77], [551.56, 176.77], [556.37, 176.48], [580.17, 176.48]]}, {\"category\": \"car\", \"corners_3d\": [[596.93, 192.47], [618.22, 192.53], [612.48, 194.11], [589.51, 194.03], [596.93, 173.2], [618.22, 173.21], [612.48, 173.23], [589.51, 173.23]]}]\n```" - } - ] - }, - { - "id": 361, - "data_source": "Mono3DRefer", - "scene_name": "000183", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[887.35, 362.54], [1111.76, 365.48], [3017.9, 1154.22], [1802.31, 1082.36], [887.35, 172.81], [1111.76, 172.81], [3017.9, 172.61], [1802.31, 172.63]]}, {\"category\": \"car\", \"corners_3d\": [[51.37, 304.73], [133.82, 286.02], [365.16, 287.19], [320.44, 306.33], [51.37, 172.09], [133.82, 172.2], [365.16, 172.19], [320.44, 172.08]]}, {\"category\": \"van\", \"corners_3d\": [[697.82, 227.02], [769.13, 227.29], [803.36, 242.46], [712.09, 242.02], [697.82, 153.62], [769.13, 153.52], [803.36, 148.13], [712.09, 148.29]]}, {\"category\": \"car\", \"corners_3d\": [[740.92, 220.71], [692.65, 220.65], [682.71, 213.79], [724.03, 213.83], [740.92, 168.41], [692.65, 168.42], [682.71, 169.05], [724.03, 169.05]]}, {\"category\": \"car\", \"corners_3d\": [[662.84, 197.74], [689.75, 197.76], [696.64, 200.39], [666.88, 200.37], [662.84, 170.06], [689.75, 170.06], [696.64, 169.77], [666.88, 169.77]]}, {\"category\": \"car\", \"corners_3d\": [[577.27, 201.1], [551.56, 201.09], [556.37, 198.99], [580.17, 198.99], [577.27, 176.77], [551.56, 176.77], [556.37, 176.48], [580.17, 176.48]]}, {\"category\": \"car\", \"corners_3d\": [[596.93, 192.47], [618.22, 192.53], [612.48, 194.11], [589.51, 194.03], [596.93, 173.2], [618.22, 173.21], [612.48, 173.23], [589.51, 173.23]]}]\n```" - } - ] - }, - { - "id": 362, - "data_source": "Mono3DRefer", - "scene_name": "000184", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[620.73, 199.27], [643.06, 199.42], [635.16, 201.41], [611.22, 201.23], [620.73, 177.97], [643.06, 178.0], [635.16, 178.38], [611.22, 178.35]]}, {\"category\": \"car\", \"corners_3d\": [[475.23, 218.21], [443.32, 217.85], [471.07, 213.54], [500.1, 213.83], [475.23, 185.14], [443.32, 185.04], [471.07, 183.88], [500.1, 183.95]]}, {\"category\": \"car\", \"corners_3d\": [[503.17, 213.22], [474.6, 212.98], [497.15, 209.26], [523.18, 209.46], [503.17, 177.15], [474.6, 177.12], [497.15, 176.73], [523.18, 176.75]]}, {\"category\": \"car\", \"corners_3d\": [[659.16, 200.34], [685.21, 200.39], [688.72, 202.87], [660.32, 202.81], [659.16, 175.8], [685.21, 175.81], [688.72, 176.07], [660.32, 176.06]]}, {\"category\": \"car\", \"corners_3d\": [[975.15, 192.5], [962.84, 191.67], [1037.38, 191.86], [1053.13, 192.71], [975.15, 163.12], [962.84, 163.53], [1037.38, 163.44], [1053.13, 163.02]]}]\n```" - } - ] - }, - { - "id": 363, - "data_source": "Mono3DRefer", - "scene_name": "000185", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-653.88, 1197.73], [-1666.29, 1252.52], [28.88, 437.78], [265.25, 434.35], [-653.88, 294.27], [-1666.29, 300.76], [28.88, 204.24], [265.25, 203.83]]}, {\"category\": \"car\", \"corners_3d\": [[375.67, 376.78], [182.15, 379.51], [339.02, 296.55], [454.21, 295.56], [375.67, 160.35], [182.15, 160.18], [339.02, 165.27], [454.21, 165.33]]}, {\"category\": \"car\", \"corners_3d\": [[500.61, 281.66], [401.43, 281.82], [453.1, 253.72], [526.67, 253.63], [500.61, 174.59], [401.43, 174.59], [453.1, 174.14], [526.67, 174.14]]}, {\"category\": \"car\", \"corners_3d\": [[585.56, 226.01], [545.06, 225.41], [570.82, 218.96], [606.56, 219.43], [585.56, 187.04], [545.06, 186.88], [570.82, 185.16], [606.56, 185.28]]}, {\"category\": \"car\", \"corners_3d\": [[615.18, 218.23], [586.55, 217.68], [608.66, 213.64], [634.93, 214.1], [615.18, 185.24], [586.55, 185.09], [608.66, 183.99], [634.93, 184.12]]}, {\"category\": \"car\", \"corners_3d\": [[656.83, 212.84], [630.76, 212.33], [649.03, 209.34], [673.32, 209.78], [656.83, 186.12], [630.76, 185.95], [649.03, 184.96], [673.32, 185.1]]}]\n```" - } - ] - }, - { - "id": 364, - "data_source": "Mono3DRefer", - "scene_name": "000189", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[426.15, 193.47], [414.67, 193.33], [442.35, 192.74], [453.6, 192.87], [426.15, 145.3], [414.67, 145.69], [442.35, 147.29], [453.6, 146.93]]}]\n```" - } - ] - }, - { - "id": 365, - "data_source": "Mono3DRefer", - "scene_name": "000189", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[709.43, 260.39], [681.69, 261.6], [659.94, 256.36], [686.21, 255.3], [709.43, 145.51], [681.69, 144.98], [659.94, 147.28], [686.21, 147.74]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[585.08, 224.76], [554.53, 225.13], [550.35, 222.98], [579.47, 222.65], [585.08, 129.75], [554.53, 129.33], [550.35, 131.79], [579.47, 132.17]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[626.69, 224.3], [599.3, 224.83], [592.26, 223.38], [618.85, 222.88], [626.69, 142.75], [599.3, 142.29], [592.26, 143.54], [618.85, 143.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[680.92, 222.34], [660.51, 222.68], [650.01, 220.93], [669.66, 220.61], [680.92, 147.46], [660.51, 147.19], [650.01, 148.57], [669.66, 148.82]]}]\n```" - } - ] - }, - { - "id": 366, - "data_source": "Mono3DRefer", - "scene_name": "000189", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[709.43, 260.39], [681.69, 261.6], [659.94, 256.36], [686.21, 255.3], [709.43, 145.51], [681.69, 144.98], [659.94, 147.28], [686.21, 147.74]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[585.08, 224.76], [554.53, 225.13], [550.35, 222.98], [579.47, 222.65], [585.08, 129.75], [554.53, 129.33], [550.35, 131.79], [579.47, 132.17]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[626.69, 224.3], [599.3, 224.83], [592.26, 223.38], [618.85, 222.88], [626.69, 142.75], [599.3, 142.29], [592.26, 143.54], [618.85, 143.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[680.92, 222.34], [660.51, 222.68], [650.01, 220.93], [669.66, 220.61], [680.92, 147.46], [660.51, 147.19], [650.01, 148.57], [669.66, 148.82]]}, {\"category\": \"cyclist\", \"corners_3d\": [[426.15, 193.47], [414.67, 193.33], [442.35, 192.74], [453.6, 192.87], [426.15, 145.3], [414.67, 145.69], [442.35, 147.29], [453.6, 146.93]]}]\n```" - } - ] - }, - { - "id": 367, - "data_source": "Mono3DRefer", - "scene_name": "000191", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[579.33, 208.46], [612.08, 208.44], [614.14, 211.01], [577.78, 211.03], [579.33, 178.68], [612.08, 178.69], [614.14, 177.96], [577.78, 177.96]]}, {\"category\": \"car\", \"corners_3d\": [[-58.45, 439.06], [196.89, 438.23], [-1620.16, 1577.11], [-3078.68, 1602.38], [-58.45, 209.82], [196.89, 209.74], [-1620.16, 320.14], [-3078.68, 322.59]]}, {\"category\": \"car\", \"corners_3d\": [[303.45, 240.19], [362.65, 240.15], [307.24, 252.94], [234.19, 253.0], [303.45, 189.55], [362.65, 189.54], [307.24, 190.55], [234.19, 190.56]]}, {\"category\": \"car\", \"corners_3d\": [[434.02, 249.49], [502.04, 249.56], [464.79, 270.78], [374.41, 270.67], [434.02, 191.52], [502.04, 191.52], [464.79, 193.6], [374.41, 193.59]]}, {\"category\": \"car\", \"corners_3d\": [[391.95, 221.4], [432.62, 221.4], [405.71, 226.96], [358.8, 226.96], [391.95, 185.44], [432.62, 185.44], [405.71, 185.47], [358.8, 185.47]]}, {\"category\": \"car\", \"corners_3d\": [[481.16, 215.68], [518.71, 215.67], [505.72, 220.5], [462.2, 220.52], [481.16, 185.41], [518.71, 185.41], [505.72, 185.44], [462.2, 185.44]]}, {\"category\": \"car\", \"corners_3d\": [[57.16, 252.0], [124.38, 252.01], [-40.4, 274.77], [-130.51, 274.76], [57.16, 188.97], [124.38, 188.98], [-40.4, 190.26], [-130.51, 190.26]]}, {\"category\": \"car\", \"corners_3d\": [[262.93, 223.28], [303.36, 223.28], [254.21, 229.43], [207.25, 229.43], [262.93, 188.58], [303.36, 188.58], [254.21, 189.12], [207.25, 189.12]]}, {\"category\": \"car\", \"corners_3d\": [[326.55, 213.49], [361.73, 213.48], [329.84, 217.26], [289.95, 217.27], [326.55, 180.07], [361.73, 180.07], [329.84, 179.38], [289.95, 179.38]]}]\n```" - } - ] - }, - { - "id": 368, - "data_source": "Mono3DRefer", - "scene_name": "000191", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[519.72, 202.22], [550.44, 202.2], [545.5, 204.55], [510.52, 204.57], [519.72, 162.02], [550.44, 162.04], [545.5, 158.84], [510.52, 158.8]]}]\n```" - } - ] - }, - { - "id": 369, - "data_source": "Mono3DRefer", - "scene_name": "000191", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[579.33, 208.46], [612.08, 208.44], [614.14, 211.01], [577.78, 211.03], [579.33, 178.68], [612.08, 178.69], [614.14, 177.96], [577.78, 177.96]]}, {\"category\": \"car\", \"corners_3d\": [[-58.45, 439.06], [196.89, 438.23], [-1620.16, 1577.11], [-3078.68, 1602.38], [-58.45, 209.82], [196.89, 209.74], [-1620.16, 320.14], [-3078.68, 322.59]]}, {\"category\": \"car\", \"corners_3d\": [[303.45, 240.19], [362.65, 240.15], [307.24, 252.94], [234.19, 253.0], [303.45, 189.55], [362.65, 189.54], [307.24, 190.55], [234.19, 190.56]]}, {\"category\": \"car\", \"corners_3d\": [[434.02, 249.49], [502.04, 249.56], [464.79, 270.78], [374.41, 270.67], [434.02, 191.52], [502.04, 191.52], [464.79, 193.6], [374.41, 193.59]]}, {\"category\": \"car\", \"corners_3d\": [[391.95, 221.4], [432.62, 221.4], [405.71, 226.96], [358.8, 226.96], [391.95, 185.44], [432.62, 185.44], [405.71, 185.47], [358.8, 185.47]]}, {\"category\": \"car\", \"corners_3d\": [[481.16, 215.68], [518.71, 215.67], [505.72, 220.5], [462.2, 220.52], [481.16, 185.41], [518.71, 185.41], [505.72, 185.44], [462.2, 185.44]]}, {\"category\": \"van\", \"corners_3d\": [[519.72, 202.22], [550.44, 202.2], [545.5, 204.55], [510.52, 204.57], [519.72, 162.02], [550.44, 162.04], [545.5, 158.84], [510.52, 158.8]]}, {\"category\": \"car\", \"corners_3d\": [[57.16, 252.0], [124.38, 252.01], [-40.4, 274.77], [-130.51, 274.76], [57.16, 188.97], [124.38, 188.98], [-40.4, 190.26], [-130.51, 190.26]]}, {\"category\": \"car\", \"corners_3d\": [[262.93, 223.28], [303.36, 223.28], [254.21, 229.43], [207.25, 229.43], [262.93, 188.58], [303.36, 188.58], [254.21, 189.12], [207.25, 189.12]]}, {\"category\": \"car\", \"corners_3d\": [[326.55, 213.49], [361.73, 213.48], [329.84, 217.26], [289.95, 217.27], [326.55, 180.07], [361.73, 180.07], [329.84, 179.38], [289.95, 179.38]]}]\n```" - } - ] - }, - { - "id": 370, - "data_source": "Mono3DRefer", - "scene_name": "000192", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[575.05, 206.71], [604.41, 206.7], [605.48, 208.8], [573.26, 208.82], [575.05, 180.03], [604.41, 180.03], [605.48, 179.52], [573.26, 179.52]]}, {\"category\": \"car\", \"corners_3d\": [[72.09, 284.1], [174.6, 283.69], [-21.88, 330.71], [-174.81, 331.59], [72.09, 190.03], [174.6, 190.01], [-21.88, 192.3], [-174.81, 192.35]]}, {\"category\": \"car\", \"corners_3d\": [[247.82, 248.8], [315.05, 248.81], [229.13, 267.47], [142.17, 267.46], [247.82, 188.79], [315.05, 188.79], [229.13, 189.85], [142.17, 189.85]]}, {\"category\": \"car\", \"corners_3d\": [[471.71, 222.73], [512.44, 222.69], [498.98, 228.92], [451.46, 228.98], [471.71, 186.47], [512.44, 186.47], [498.98, 186.68], [451.46, 186.68]]}, {\"category\": \"car\", \"corners_3d\": [[499.15, 210.75], [528.97, 210.75], [521.01, 213.33], [488.18, 213.33], [499.15, 184.45], [528.97, 184.45], [521.01, 184.38], [488.18, 184.38]]}, {\"category\": \"car\", \"corners_3d\": [[30.62, 261.18], [111.86, 261.1], [-47.05, 285.77], [-154.96, 285.91], [30.62, 186.11], [111.86, 186.11], [-47.05, 186.4], [-154.96, 186.4]]}]\n```" - } - ] - }, - { - "id": 371, - "data_source": "Mono3DRefer", - "scene_name": "000192", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[370.58, 253.11], [461.77, 253.04], [383.46, 291.3], [240.58, 291.49], [370.58, 133.77], [461.77, 133.83], [383.46, 104.84], [240.58, 104.69]]}]\n```" - } - ] - }, - { - "id": 372, - "data_source": "Mono3DRefer", - "scene_name": "000192", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[575.05, 206.71], [604.41, 206.7], [605.48, 208.8], [573.26, 208.82], [575.05, 180.03], [604.41, 180.03], [605.48, 179.52], [573.26, 179.52]]}, {\"category\": \"car\", \"corners_3d\": [[72.09, 284.1], [174.6, 283.69], [-21.88, 330.71], [-174.81, 331.59], [72.09, 190.03], [174.6, 190.01], [-21.88, 192.3], [-174.81, 192.35]]}, {\"category\": \"car\", \"corners_3d\": [[247.82, 248.8], [315.05, 248.81], [229.13, 267.47], [142.17, 267.46], [247.82, 188.79], [315.05, 188.79], [229.13, 189.85], [142.17, 189.85]]}, {\"category\": \"van\", \"corners_3d\": [[370.58, 253.11], [461.77, 253.04], [383.46, 291.3], [240.58, 291.49], [370.58, 133.77], [461.77, 133.83], [383.46, 104.84], [240.58, 104.69]]}, {\"category\": \"car\", \"corners_3d\": [[471.71, 222.73], [512.44, 222.69], [498.98, 228.92], [451.46, 228.98], [471.71, 186.47], [512.44, 186.47], [498.98, 186.68], [451.46, 186.68]]}, {\"category\": \"car\", \"corners_3d\": [[499.15, 210.75], [528.97, 210.75], [521.01, 213.33], [488.18, 213.33], [499.15, 184.45], [528.97, 184.45], [521.01, 184.38], [488.18, 184.38]]}, {\"category\": \"car\", \"corners_3d\": [[30.62, 261.18], [111.86, 261.1], [-47.05, 285.77], [-154.96, 285.91], [30.62, 186.11], [111.86, 186.11], [-47.05, 186.4], [-154.96, 186.4]]}]\n```" - } - ] - }, - { - "id": 373, - "data_source": "Mono3DRefer", - "scene_name": "000195", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[681.72, 203.98], [694.38, 204.0], [696.92, 205.42], [683.67, 205.4], [681.72, 172.85], [694.38, 172.85], [696.92, 172.85], [683.67, 172.85]]}]\n```" - } - ] - }, - { - "id": 374, - "data_source": "Mono3DRefer", - "scene_name": "000195", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[988.0, 273.91], [942.96, 274.17], [924.23, 268.94], [966.99, 268.71], [988.0, 167.32], [942.96, 167.3], [924.23, 167.59], [966.99, 167.6]]}]\n```" - } - ] - }, - { - "id": 375, - "data_source": "Mono3DRefer", - "scene_name": "000195", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[988.0, 273.91], [942.96, 274.17], [924.23, 268.94], [966.99, 268.71], [988.0, 167.32], [942.96, 167.3], [924.23, 167.59], [966.99, 167.6]]}, {\"category\": \"cyclist\", \"corners_3d\": [[681.72, 203.98], [694.38, 204.0], [696.92, 205.42], [683.67, 205.4], [681.72, 172.85], [694.38, 172.85], [696.92, 172.85], [683.67, 172.85]]}]\n```" - } - ] - }, - { - "id": 376, - "data_source": "Mono3DRefer", - "scene_name": "000200", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[607.04, 229.24], [659.14, 229.37], [663.56, 238.51], [603.04, 238.34], [607.04, 178.8], [659.14, 178.81], [663.56, 179.77], [603.04, 179.75]]}, {\"category\": \"car\", \"corners_3d\": [[798.64, 279.29], [897.4, 279.59], [1018.11, 326.81], [875.35, 326.18], [798.64, 185.89], [897.4, 185.93], [1018.11, 191.71], [875.35, 191.64]]}]\n```" - } - ] - }, - { - "id": 377, - "data_source": "Mono3DRefer", - "scene_name": "000200", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[895.73, 335.22], [877.78, 326.59], [907.72, 326.12], [927.3, 334.69], [895.73, 173.81], [877.78, 173.76], [907.72, 173.76], [927.3, 173.81]]}]\n```" - } - ] - }, - { - "id": 378, - "data_source": "Mono3DRefer", - "scene_name": "000200", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[607.04, 229.24], [659.14, 229.37], [663.56, 238.51], [603.04, 238.34], [607.04, 178.8], [659.14, 178.81], [663.56, 179.77], [603.04, 179.75]]}, {\"category\": \"car\", \"corners_3d\": [[798.64, 279.29], [897.4, 279.59], [1018.11, 326.81], [875.35, 326.18], [798.64, 185.89], [897.4, 185.93], [1018.11, 191.71], [875.35, 191.64]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[895.73, 335.22], [877.78, 326.59], [907.72, 326.12], [927.3, 334.69], [895.73, 173.81], [877.78, 173.76], [907.72, 173.76], [927.3, 173.81]]}]\n```" - } - ] - }, - { - "id": 379, - "data_source": "Mono3DRefer", - "scene_name": "000202", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[563.83, 205.01], [535.31, 205.01], [541.71, 202.22], [567.75, 202.22], [563.83, 181.14], [535.31, 181.14], [541.71, 180.42], [567.75, 180.42]]}]\n```" - } - ] - }, - { - "id": 380, - "data_source": "Mono3DRefer", - "scene_name": "000210", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-27.09, 324.61], [-156.47, 325.16], [65.24, 280.2], [155.96, 279.92], [-27.09, 183.05], [-156.47, 183.09], [65.24, 180.07], [155.96, 180.05]]}, {\"category\": \"car\", \"corners_3d\": [[192.48, 270.59], [124.64, 269.62], [230.19, 251.28], [285.92, 251.91], [192.48, 197.53], [124.64, 197.28], [230.19, 192.65], [285.92, 192.81]]}, {\"category\": \"car\", \"corners_3d\": [[304.23, 247.95], [239.02, 247.82], [310.81, 233.82], [363.92, 233.91], [304.23, 180.54], [239.02, 180.53], [310.81, 179.1], [363.92, 179.11]]}, {\"category\": \"car\", \"corners_3d\": [[417.81, 221.8], [376.33, 221.72], [407.55, 215.72], [443.97, 215.78], [417.81, 183.19], [376.33, 183.17], [407.55, 181.91], [443.97, 181.92]]}, {\"category\": \"car\", \"corners_3d\": [[452.89, 214.0], [419.25, 213.91], [442.52, 209.68], [472.73, 209.75], [452.89, 182.44], [419.25, 182.42], [442.52, 181.43], [472.73, 181.45]]}, {\"category\": \"car\", \"corners_3d\": [[521.34, 199.38], [498.94, 199.33], [510.43, 197.25], [531.08, 197.29], [521.34, 177.59], [498.94, 177.58], [510.43, 177.21], [531.08, 177.21]]}]\n```" - } - ] - }, - { - "id": 381, - "data_source": "Mono3DRefer", - "scene_name": "000210", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[681.5, 183.22], [700.76, 183.23], [727.76, 186.91], [701.67, 186.9], [681.5, 156.24], [700.76, 156.23], [727.76, 150.33], [701.67, 150.35]]}]\n```" - } - ] - }, - { - "id": 382, - "data_source": "Mono3DRefer", - "scene_name": "000210", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[681.5, 183.22], [700.76, 183.23], [727.76, 186.91], [701.67, 186.9], [681.5, 156.24], [700.76, 156.23], [727.76, 150.33], [701.67, 150.35]]}, {\"category\": \"car\", \"corners_3d\": [[-27.09, 324.61], [-156.47, 325.16], [65.24, 280.2], [155.96, 279.92], [-27.09, 183.05], [-156.47, 183.09], [65.24, 180.07], [155.96, 180.05]]}, {\"category\": \"car\", \"corners_3d\": [[192.48, 270.59], [124.64, 269.62], [230.19, 251.28], [285.92, 251.91], [192.48, 197.53], [124.64, 197.28], [230.19, 192.65], [285.92, 192.81]]}, {\"category\": \"car\", \"corners_3d\": [[304.23, 247.95], [239.02, 247.82], [310.81, 233.82], [363.92, 233.91], [304.23, 180.54], [239.02, 180.53], [310.81, 179.1], [363.92, 179.11]]}, {\"category\": \"car\", \"corners_3d\": [[417.81, 221.8], [376.33, 221.72], [407.55, 215.72], [443.97, 215.78], [417.81, 183.19], [376.33, 183.17], [407.55, 181.91], [443.97, 181.92]]}, {\"category\": \"car\", \"corners_3d\": [[452.89, 214.0], [419.25, 213.91], [442.52, 209.68], [472.73, 209.75], [452.89, 182.44], [419.25, 182.42], [442.52, 181.43], [472.73, 181.45]]}, {\"category\": \"car\", \"corners_3d\": [[521.34, 199.38], [498.94, 199.33], [510.43, 197.25], [531.08, 197.29], [521.34, 177.59], [498.94, 177.58], [510.43, 177.21], [531.08, 177.21]]}]\n```" - } - ] - }, - { - "id": 383, - "data_source": "Mono3DRefer", - "scene_name": "000214", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1116.3, 243.04], [1145.79, 243.15], [1231.12, 254.99], [1196.51, 254.83], [1116.3, 140.82], [1145.79, 140.77], [1231.12, 135.37], [1196.51, 135.44]]}, {\"category\": \"cyclist\", \"corners_3d\": [[983.16, 225.53], [1002.8, 225.62], [1043.48, 231.76], [1021.48, 231.66], [983.16, 153.8], [1002.8, 153.77], [1043.48, 151.54], [1021.48, 151.58]]}, {\"category\": \"cyclist\", \"corners_3d\": [[941.12, 217.48], [957.51, 217.89], [955.95, 221.54], [938.26, 221.05], [941.12, 163.11], [957.51, 163.02], [955.95, 162.22], [938.26, 162.33]]}]\n```" - } - ] - }, - { - "id": 384, - "data_source": "Mono3DRefer", - "scene_name": "000214", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[412.75, 227.46], [399.79, 231.51], [277.59, 231.8], [299.08, 227.72], [412.75, 181.34], [399.79, 181.97], [277.59, 182.01], [299.08, 181.38]]}, {\"category\": \"car\", \"corners_3d\": [[334.43, 222.26], [353.48, 219.5], [443.22, 220.05], [429.31, 222.88], [334.43, 182.62], [353.48, 182.08], [443.22, 182.19], [429.31, 182.74]]}, {\"category\": \"car\", \"corners_3d\": [[902.71, 184.41], [877.22, 184.3], [872.57, 183.43], [896.06, 183.52], [902.71, 163.12], [877.22, 163.21], [872.57, 163.95], [896.06, 163.87]]}]\n```" - } - ] - }, - { - "id": 385, - "data_source": "Mono3DRefer", - "scene_name": "000214", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[412.75, 227.46], [399.79, 231.51], [277.59, 231.8], [299.08, 227.72], [412.75, 181.34], [399.79, 181.97], [277.59, 182.01], [299.08, 181.38]]}, {\"category\": \"car\", \"corners_3d\": [[334.43, 222.26], [353.48, 219.5], [443.22, 220.05], [429.31, 222.88], [334.43, 182.62], [353.48, 182.08], [443.22, 182.19], [429.31, 182.74]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1116.3, 243.04], [1145.79, 243.15], [1231.12, 254.99], [1196.51, 254.83], [1116.3, 140.82], [1145.79, 140.77], [1231.12, 135.37], [1196.51, 135.44]]}, {\"category\": \"cyclist\", \"corners_3d\": [[983.16, 225.53], [1002.8, 225.62], [1043.48, 231.76], [1021.48, 231.66], [983.16, 153.8], [1002.8, 153.77], [1043.48, 151.54], [1021.48, 151.58]]}, {\"category\": \"cyclist\", \"corners_3d\": [[941.12, 217.48], [957.51, 217.89], [955.95, 221.54], [938.26, 221.05], [941.12, 163.11], [957.51, 163.02], [955.95, 162.22], [938.26, 162.33]]}, {\"category\": \"car\", \"corners_3d\": [[902.71, 184.41], [877.22, 184.3], [872.57, 183.43], [896.06, 183.52], [902.71, 163.12], [877.22, 163.21], [872.57, 163.95], [896.06, 163.87]]}]\n```" - } - ] - }, - { - "id": 386, - "data_source": "Mono3DRefer", - "scene_name": "000215", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[570.82, 284.54], [585.63, 303.97], [215.44, 315.46], [258.43, 292.77], [570.82, 190.91], [585.63, 194.05], [215.44, 195.91], [258.43, 192.24]]}, {\"category\": \"car\", \"corners_3d\": [[1163.41, 286.17], [1269.33, 303.5], [1023.55, 313.88], [945.23, 293.9], [1163.41, 188.71], [1269.33, 191.14], [1023.55, 192.59], [945.23, 189.79]]}]\n```" - } - ] - }, - { - "id": 387, - "data_source": "Mono3DRefer", - "scene_name": "000217", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[543.06, 212.51], [500.54, 212.62], [510.54, 207.13], [547.17, 207.04], [543.06, 160.99], [500.54, 160.96], [510.54, 162.6], [547.17, 162.63]]}]\n```" - } - ] - }, - { - "id": 388, - "data_source": "Mono3DRefer", - "scene_name": "000217", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1173.67, 373.69], [1098.41, 374.58], [983.49, 329.85], [1042.51, 329.31], [1173.67, 142.61], [1098.41, 142.48], [983.49, 149.21], [1042.51, 149.29]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1054.02, 298.2], [1107.72, 297.82], [1256.48, 333.14], [1188.13, 333.75], [1054.02, 143.01], [1107.72, 143.1], [1256.48, 134.7], [1188.13, 134.55]]}]\n```" - } - ] - }, - { - "id": 389, - "data_source": "Mono3DRefer", - "scene_name": "000217", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[81.59, 491.84], [-250.78, 493.42], [204.33, 322.29], [358.64, 321.95], [81.59, 222.01], [-250.78, 222.26], [204.33, 195.88], [358.64, 195.83]]}, {\"category\": \"car\", \"corners_3d\": [[420.09, 282.35], [317.24, 283.13], [387.28, 252.92], [461.73, 252.5], [420.09, 198.06], [317.24, 198.24], [387.28, 191.28], [461.73, 191.19]]}, {\"category\": \"car\", \"corners_3d\": [[515.03, 223.04], [475.69, 223.04], [490.4, 217.5], [525.41, 217.5], [515.03, 185.05], [475.69, 185.05], [490.4, 183.71], [525.41, 183.71]]}, {\"category\": \"car\", \"corners_3d\": [[594.39, 206.44], [562.5, 206.5], [564.18, 203.29], [593.03, 203.24], [594.39, 179.1], [562.5, 179.11], [564.18, 178.51], [593.03, 178.5]]}, {\"category\": \"car\", \"corners_3d\": [[557.58, 200.83], [537.76, 200.84], [543.0, 198.55], [561.19, 198.55], [557.58, 180.29], [537.76, 180.29], [543.0, 179.69], [561.19, 179.68]]}]\n```" - } - ] - }, - { - "id": 390, - "data_source": "Mono3DRefer", - "scene_name": "000217", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[543.06, 212.51], [500.54, 212.62], [510.54, 207.13], [547.17, 207.04], [543.06, 160.99], [500.54, 160.96], [510.54, 162.6], [547.17, 162.63]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1173.67, 373.69], [1098.41, 374.58], [983.49, 329.85], [1042.51, 329.31], [1173.67, 142.61], [1098.41, 142.48], [983.49, 149.21], [1042.51, 149.29]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1054.02, 298.2], [1107.72, 297.82], [1256.48, 333.14], [1188.13, 333.75], [1054.02, 143.01], [1107.72, 143.1], [1256.48, 134.7], [1188.13, 134.55]]}]\n```" - } - ] - }, - { - "id": 391, - "data_source": "Mono3DRefer", - "scene_name": "000217", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[81.59, 491.84], [-250.78, 493.42], [204.33, 322.29], [358.64, 321.95], [81.59, 222.01], [-250.78, 222.26], [204.33, 195.88], [358.64, 195.83]]}, {\"category\": \"car\", \"corners_3d\": [[420.09, 282.35], [317.24, 283.13], [387.28, 252.92], [461.73, 252.5], [420.09, 198.06], [317.24, 198.24], [387.28, 191.28], [461.73, 191.19]]}, {\"category\": \"car\", \"corners_3d\": [[515.03, 223.04], [475.69, 223.04], [490.4, 217.5], [525.41, 217.5], [515.03, 185.05], [475.69, 185.05], [490.4, 183.71], [525.41, 183.71]]}, {\"category\": \"van\", \"corners_3d\": [[543.06, 212.51], [500.54, 212.62], [510.54, 207.13], [547.17, 207.04], [543.06, 160.99], [500.54, 160.96], [510.54, 162.6], [547.17, 162.63]]}, {\"category\": \"car\", \"corners_3d\": [[594.39, 206.44], [562.5, 206.5], [564.18, 203.29], [593.03, 203.24], [594.39, 179.1], [562.5, 179.11], [564.18, 178.51], [593.03, 178.5]]}, {\"category\": \"car\", \"corners_3d\": [[557.58, 200.83], [537.76, 200.84], [543.0, 198.55], [561.19, 198.55], [557.58, 180.29], [537.76, 180.29], [543.0, 179.69], [561.19, 179.68]]}]\n```" - } - ] - }, - { - "id": 392, - "data_source": "Mono3DRefer", - "scene_name": "000217", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[81.59, 491.84], [-250.78, 493.42], [204.33, 322.29], [358.64, 321.95], [81.59, 222.01], [-250.78, 222.26], [204.33, 195.88], [358.64, 195.83]]}, {\"category\": \"car\", \"corners_3d\": [[420.09, 282.35], [317.24, 283.13], [387.28, 252.92], [461.73, 252.5], [420.09, 198.06], [317.24, 198.24], [387.28, 191.28], [461.73, 191.19]]}, {\"category\": \"car\", \"corners_3d\": [[515.03, 223.04], [475.69, 223.04], [490.4, 217.5], [525.41, 217.5], [515.03, 185.05], [475.69, 185.05], [490.4, 183.71], [525.41, 183.71]]}, {\"category\": \"car\", \"corners_3d\": [[594.39, 206.44], [562.5, 206.5], [564.18, 203.29], [593.03, 203.24], [594.39, 179.1], [562.5, 179.11], [564.18, 178.51], [593.03, 178.5]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1173.67, 373.69], [1098.41, 374.58], [983.49, 329.85], [1042.51, 329.31], [1173.67, 142.61], [1098.41, 142.48], [983.49, 149.21], [1042.51, 149.29]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1054.02, 298.2], [1107.72, 297.82], [1256.48, 333.14], [1188.13, 333.75], [1054.02, 143.01], [1107.72, 143.1], [1256.48, 134.7], [1188.13, 134.55]]}, {\"category\": \"car\", \"corners_3d\": [[557.58, 200.83], [537.76, 200.84], [543.0, 198.55], [561.19, 198.55], [557.58, 180.29], [537.76, 180.29], [543.0, 179.69], [561.19, 179.68]]}]\n```" - } - ] - }, - { - "id": 393, - "data_source": "Mono3DRefer", - "scene_name": "000217", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[81.59, 491.84], [-250.78, 493.42], [204.33, 322.29], [358.64, 321.95], [81.59, 222.01], [-250.78, 222.26], [204.33, 195.88], [358.64, 195.83]]}, {\"category\": \"car\", \"corners_3d\": [[420.09, 282.35], [317.24, 283.13], [387.28, 252.92], [461.73, 252.5], [420.09, 198.06], [317.24, 198.24], [387.28, 191.28], [461.73, 191.19]]}, {\"category\": \"car\", \"corners_3d\": [[515.03, 223.04], [475.69, 223.04], [490.4, 217.5], [525.41, 217.5], [515.03, 185.05], [475.69, 185.05], [490.4, 183.71], [525.41, 183.71]]}, {\"category\": \"van\", \"corners_3d\": [[543.06, 212.51], [500.54, 212.62], [510.54, 207.13], [547.17, 207.04], [543.06, 160.99], [500.54, 160.96], [510.54, 162.6], [547.17, 162.63]]}, {\"category\": \"car\", \"corners_3d\": [[594.39, 206.44], [562.5, 206.5], [564.18, 203.29], [593.03, 203.24], [594.39, 179.1], [562.5, 179.11], [564.18, 178.51], [593.03, 178.5]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1173.67, 373.69], [1098.41, 374.58], [983.49, 329.85], [1042.51, 329.31], [1173.67, 142.61], [1098.41, 142.48], [983.49, 149.21], [1042.51, 149.29]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1054.02, 298.2], [1107.72, 297.82], [1256.48, 333.14], [1188.13, 333.75], [1054.02, 143.01], [1107.72, 143.1], [1256.48, 134.7], [1188.13, 134.55]]}, {\"category\": \"car\", \"corners_3d\": [[557.58, 200.83], [537.76, 200.84], [543.0, 198.55], [561.19, 198.55], [557.58, 180.29], [537.76, 180.29], [543.0, 179.69], [561.19, 179.68]]}]\n```" - } - ] - }, - { - "id": 394, - "data_source": "Mono3DRefer", - "scene_name": "000219", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[570.33, 219.47], [614.3, 219.42], [616.77, 225.6], [566.96, 225.67], [570.33, 176.82], [614.3, 176.81], [616.77, 177.34], [566.96, 177.35]]}]\n```" - } - ] - }, - { - "id": 395, - "data_source": "Mono3DRefer", - "scene_name": "000219", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[940.07, 188.24], [983.25, 188.24], [1050.49, 191.01], [999.52, 191.01], [940.07, 144.7], [983.25, 144.7], [1050.49, 139.62], [999.52, 139.62]]}]\n```" - } - ] - }, - { - "id": 396, - "data_source": "Mono3DRefer", - "scene_name": "000219", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[822.38, 185.19], [844.94, 185.2], [864.82, 186.32], [840.22, 186.31], [822.38, 163.86], [844.94, 163.85], [864.82, 163.04], [840.22, 163.05]]}, {\"category\": \"van\", \"corners_3d\": [[870.23, 188.65], [902.39, 188.66], [936.28, 190.59], [900.17, 190.57], [870.23, 159.56], [902.39, 159.55], [936.28, 157.93], [900.17, 157.94]]}]\n```" - } - ] - }, - { - "id": 397, - "data_source": "Mono3DRefer", - "scene_name": "000219", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[570.33, 219.47], [614.3, 219.42], [616.77, 225.6], [566.96, 225.67], [570.33, 176.82], [614.3, 176.81], [616.77, 177.34], [566.96, 177.35]]}, {\"category\": \"truck\", \"corners_3d\": [[940.07, 188.24], [983.25, 188.24], [1050.49, 191.01], [999.52, 191.01], [940.07, 144.7], [983.25, 144.7], [1050.49, 139.62], [999.52, 139.62]]}]\n```" - } - ] - }, - { - "id": 398, - "data_source": "Mono3DRefer", - "scene_name": "000219", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[570.33, 219.47], [614.3, 219.42], [616.77, 225.6], [566.96, 225.67], [570.33, 176.82], [614.3, 176.81], [616.77, 177.34], [566.96, 177.35]]}, {\"category\": \"van\", \"corners_3d\": [[822.38, 185.19], [844.94, 185.2], [864.82, 186.32], [840.22, 186.31], [822.38, 163.86], [844.94, 163.85], [864.82, 163.04], [840.22, 163.05]]}, {\"category\": \"van\", \"corners_3d\": [[870.23, 188.65], [902.39, 188.66], [936.28, 190.59], [900.17, 190.57], [870.23, 159.56], [902.39, 159.55], [936.28, 157.93], [900.17, 157.94]]}]\n```" - } - ] - }, - { - "id": 399, - "data_source": "Mono3DRefer", - "scene_name": "000219", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[822.38, 185.19], [844.94, 185.2], [864.82, 186.32], [840.22, 186.31], [822.38, 163.86], [844.94, 163.85], [864.82, 163.04], [840.22, 163.05]]}, {\"category\": \"van\", \"corners_3d\": [[870.23, 188.65], [902.39, 188.66], [936.28, 190.59], [900.17, 190.57], [870.23, 159.56], [902.39, 159.55], [936.28, 157.93], [900.17, 157.94]]}, {\"category\": \"truck\", \"corners_3d\": [[940.07, 188.24], [983.25, 188.24], [1050.49, 191.01], [999.52, 191.01], [940.07, 144.7], [983.25, 144.7], [1050.49, 139.62], [999.52, 139.62]]}]\n```" - } - ] - }, - { - "id": 400, - "data_source": "Mono3DRefer", - "scene_name": "000219", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[570.33, 219.47], [614.3, 219.42], [616.77, 225.6], [566.96, 225.67], [570.33, 176.82], [614.3, 176.81], [616.77, 177.34], [566.96, 177.35]]}, {\"category\": \"van\", \"corners_3d\": [[822.38, 185.19], [844.94, 185.2], [864.82, 186.32], [840.22, 186.31], [822.38, 163.86], [844.94, 163.85], [864.82, 163.04], [840.22, 163.05]]}, {\"category\": \"van\", \"corners_3d\": [[870.23, 188.65], [902.39, 188.66], [936.28, 190.59], [900.17, 190.57], [870.23, 159.56], [902.39, 159.55], [936.28, 157.93], [900.17, 157.94]]}, {\"category\": \"truck\", \"corners_3d\": [[940.07, 188.24], [983.25, 188.24], [1050.49, 191.01], [999.52, 191.01], [940.07, 144.7], [983.25, 144.7], [1050.49, 139.62], [999.52, 139.62]]}]\n```" - } - ] - }, - { - "id": 401, - "data_source": "Mono3DRefer", - "scene_name": "000221", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[849.89, 308.74], [836.78, 302.64], [862.18, 302.31], [876.44, 308.37], [849.89, 173.65], [836.78, 173.62], [862.18, 173.62], [876.44, 173.65]]}]\n```" - } - ] - }, - { - "id": 402, - "data_source": "Mono3DRefer", - "scene_name": "000221", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[778.26, 267.16], [865.69, 267.4], [955.65, 302.67], [835.41, 302.23], [778.26, 184.41], [865.69, 184.44], [955.65, 188.76], [835.41, 188.7]]}, {\"category\": \"car\", \"corners_3d\": [[607.67, 229.07], [659.61, 229.2], [664.1, 238.28], [603.79, 238.11], [607.67, 178.78], [659.61, 178.79], [664.1, 179.75], [603.79, 179.73]]}]\n```" - } - ] - }, - { - "id": 403, - "data_source": "Mono3DRefer", - "scene_name": "000221", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[778.26, 267.16], [865.69, 267.4], [955.65, 302.67], [835.41, 302.23], [778.26, 184.41], [865.69, 184.44], [955.65, 188.76], [835.41, 188.7]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[849.89, 308.74], [836.78, 302.64], [862.18, 302.31], [876.44, 308.37], [849.89, 173.65], [836.78, 173.62], [862.18, 173.62], [876.44, 173.65]]}, {\"category\": \"car\", \"corners_3d\": [[607.67, 229.07], [659.61, 229.2], [664.1, 238.28], [603.79, 238.11], [607.67, 178.78], [659.61, 178.79], [664.1, 179.75], [603.79, 179.73]]}]\n```" - } - ] - }, - { - "id": 404, - "data_source": "Mono3DRefer", - "scene_name": "000222", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[450.18, 201.68], [425.74, 201.67], [441.8, 199.24], [464.18, 199.25], [450.18, 182.09], [425.74, 182.08], [441.8, 181.31], [464.18, 181.31]]}, {\"category\": \"car\", \"corners_3d\": [[555.17, 299.27], [690.78, 299.29], [740.8, 377.72], [521.08, 377.67], [555.17, 170.33], [690.78, 170.33], [740.8, 168.76], [521.08, 168.76]]}, {\"category\": \"car\", \"corners_3d\": [[345.03, 218.02], [365.6, 215.59], [457.2, 216.53], [441.49, 219.07], [345.03, 185.21], [365.6, 184.54], [457.2, 184.8], [441.49, 185.49]]}, {\"category\": \"car\", \"corners_3d\": [[402.22, 205.41], [376.73, 205.39], [394.43, 203.06], [418.1, 203.07], [402.22, 186.1], [376.73, 186.09], [394.43, 185.14], [418.1, 185.15]]}, {\"category\": \"car\", \"corners_3d\": [[176.09, 226.42], [205.63, 223.9], [259.89, 225.11], [232.37, 227.76], [176.09, 195.22], [205.63, 194.17], [259.89, 194.67], [232.37, 195.78]]}, {\"category\": \"car\", \"corners_3d\": [[426.95, 210.26], [439.79, 208.73], [498.56, 209.29], [488.04, 210.87], [426.95, 179.08], [439.79, 178.83], [498.56, 178.92], [488.04, 179.18]]}, {\"category\": \"car\", \"corners_3d\": [[360.03, 208.31], [334.93, 208.28], [350.62, 206.36], [374.36, 206.38], [360.03, 188.81], [334.93, 188.8], [350.62, 187.93], [374.36, 187.94]]}]\n```" - } - ] - }, - { - "id": 405, - "data_source": "Mono3DRefer", - "scene_name": "000225", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[548.77, 217.27], [494.23, 217.31], [517.4, 207.73], [560.18, 207.71], [548.77, 141.18], [494.23, 141.15], [517.4, 147.98], [560.18, 148.0]]}]\n```" - } - ] - }, - { - "id": 406, - "data_source": "Mono3DRefer", - "scene_name": "000225", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[45.62, 230.75], [1.12, 230.71], [83.44, 222.97], [122.02, 223.0], [45.62, 176.25], [1.12, 176.25], [83.44, 175.79], [122.02, 175.8]]}, {\"category\": \"car\", \"corners_3d\": [[574.46, 197.07], [555.54, 197.08], [558.59, 195.48], [576.26, 195.48], [574.46, 178.28], [555.54, 178.28], [558.59, 177.93], [576.26, 177.93]]}, {\"category\": \"car\", \"corners_3d\": [[390.94, 197.09], [408.75, 197.09], [399.4, 198.21], [380.77, 198.21], [390.94, 180.93], [408.75, 180.93], [399.4, 181.3], [380.77, 181.3]]}, {\"category\": \"car\", \"corners_3d\": [[587.93, 209.51], [620.77, 209.52], [622.3, 214.78], [584.74, 214.78], [587.93, 176.88], [620.77, 176.88], [622.3, 177.45], [584.74, 177.45]]}]\n```" - } - ] - }, - { - "id": 407, - "data_source": "Mono3DRefer", - "scene_name": "000225", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[45.62, 230.75], [1.12, 230.71], [83.44, 222.97], [122.02, 223.0], [45.62, 176.25], [1.12, 176.25], [83.44, 175.79], [122.02, 175.8]]}, {\"category\": \"truck\", \"corners_3d\": [[548.77, 217.27], [494.23, 217.31], [517.4, 207.73], [560.18, 207.71], [548.77, 141.18], [494.23, 141.15], [517.4, 147.98], [560.18, 148.0]]}, {\"category\": \"car\", \"corners_3d\": [[574.46, 197.07], [555.54, 197.08], [558.59, 195.48], [576.26, 195.48], [574.46, 178.28], [555.54, 178.28], [558.59, 177.93], [576.26, 177.93]]}, {\"category\": \"car\", \"corners_3d\": [[390.94, 197.09], [408.75, 197.09], [399.4, 198.21], [380.77, 198.21], [390.94, 180.93], [408.75, 180.93], [399.4, 181.3], [380.77, 181.3]]}, {\"category\": \"car\", \"corners_3d\": [[587.93, 209.51], [620.77, 209.52], [622.3, 214.78], [584.74, 214.78], [587.93, 176.88], [620.77, 176.88], [622.3, 177.45], [584.74, 177.45]]}]\n```" - } - ] - }, - { - "id": 408, - "data_source": "Mono3DRefer", - "scene_name": "000227", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1083.91, 199.89], [1042.36, 199.15], [1049.13, 196.63], [1086.77, 197.23], [1083.91, 157.03], [1042.36, 157.46], [1049.13, 158.94], [1086.77, 158.58]]}]\n```" - } - ] - }, - { - "id": 409, - "data_source": "Mono3DRefer", - "scene_name": "000227", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[848.65, 248.86], [784.33, 244.54], [853.03, 232.46], [909.38, 235.42], [848.65, 181.08], [784.33, 180.61], [853.03, 179.31], [909.38, 179.63]]}, {\"category\": \"car\", \"corners_3d\": [[1123.83, 204.37], [1082.56, 203.44], [1091.84, 201.1], [1130.13, 201.9], [1123.83, 175.24], [1082.56, 175.17], [1091.84, 174.99], [1130.13, 175.05]]}, {\"category\": \"car\", \"corners_3d\": [[925.85, 233.97], [865.39, 230.92], [909.97, 222.99], [963.82, 225.25], [925.85, 182.91], [865.39, 182.41], [909.97, 181.1], [963.82, 181.48]]}, {\"category\": \"car\", \"corners_3d\": [[987.49, 218.06], [946.49, 216.58], [971.54, 212.33], [1009.2, 213.53], [987.49, 180.79], [946.49, 180.53], [971.54, 179.78], [1009.2, 180.0]]}, {\"category\": \"car\", \"corners_3d\": [[1057.22, 205.92], [1020.43, 205.1], [1025.99, 203.17], [1060.64, 203.89], [1057.22, 177.91], [1020.43, 177.78], [1025.99, 177.48], [1060.64, 177.6]]}]\n```" - } - ] - }, - { - "id": 410, - "data_source": "Mono3DRefer", - "scene_name": "000227", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[848.65, 248.86], [784.33, 244.54], [853.03, 232.46], [909.38, 235.42], [848.65, 181.08], [784.33, 180.61], [853.03, 179.31], [909.38, 179.63]]}, {\"category\": \"car\", \"corners_3d\": [[1123.83, 204.37], [1082.56, 203.44], [1091.84, 201.1], [1130.13, 201.9], [1123.83, 175.24], [1082.56, 175.17], [1091.84, 174.99], [1130.13, 175.05]]}, {\"category\": \"car\", \"corners_3d\": [[925.85, 233.97], [865.39, 230.92], [909.97, 222.99], [963.82, 225.25], [925.85, 182.91], [865.39, 182.41], [909.97, 181.1], [963.82, 181.48]]}, {\"category\": \"car\", \"corners_3d\": [[987.49, 218.06], [946.49, 216.58], [971.54, 212.33], [1009.2, 213.53], [987.49, 180.79], [946.49, 180.53], [971.54, 179.78], [1009.2, 180.0]]}, {\"category\": \"van\", \"corners_3d\": [[1083.91, 199.89], [1042.36, 199.15], [1049.13, 196.63], [1086.77, 197.23], [1083.91, 157.03], [1042.36, 157.46], [1049.13, 158.94], [1086.77, 158.58]]}, {\"category\": \"car\", \"corners_3d\": [[1057.22, 205.92], [1020.43, 205.1], [1025.99, 203.17], [1060.64, 203.89], [1057.22, 177.91], [1020.43, 177.78], [1025.99, 177.48], [1060.64, 177.6]]}]\n```" - } - ] - }, - { - "id": 411, - "data_source": "Mono3DRefer", - "scene_name": "000228", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[316.63, 212.96], [342.53, 214.21], [245.65, 219.7], [219.74, 218.1], [316.63, 178.9], [342.53, 179.09], [245.65, 179.92], [219.74, 179.67]]}, {\"category\": \"car\", \"corners_3d\": [[499.88, 202.46], [522.28, 203.16], [467.49, 206.01], [444.42, 205.17], [499.88, 173.62], [522.28, 173.64], [467.49, 173.71], [444.42, 173.69]]}, {\"category\": \"car\", \"corners_3d\": [[613.1, 197.01], [633.86, 197.45], [603.0, 199.22], [581.36, 198.72], [613.1, 174.08], [633.86, 174.1], [603.0, 174.19], [581.36, 174.17]]}]\n```" - } - ] - }, - { - "id": 412, - "data_source": "Mono3DRefer", - "scene_name": "000228", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[444.27, 201.0], [421.75, 200.35], [479.89, 197.85], [501.55, 198.38], [444.27, 165.68], [421.75, 165.84], [479.89, 166.48], [501.55, 166.34]]}]\n```" - } - ] - }, - { - "id": 413, - "data_source": "Mono3DRefer", - "scene_name": "000228", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[866.29, 233.97], [888.1, 234.26], [892.85, 236.06], [870.38, 235.76], [866.29, 165.35], [888.1, 165.32], [892.85, 165.1], [870.38, 165.13]]}]\n```" - } - ] - }, - { - "id": 414, - "data_source": "Mono3DRefer", - "scene_name": "000228", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[316.63, 212.96], [342.53, 214.21], [245.65, 219.7], [219.74, 218.1], [316.63, 178.9], [342.53, 179.09], [245.65, 179.92], [219.74, 179.67]]}, {\"category\": \"van\", \"corners_3d\": [[444.27, 201.0], [421.75, 200.35], [479.89, 197.85], [501.55, 198.38], [444.27, 165.68], [421.75, 165.84], [479.89, 166.48], [501.55, 166.34]]}, {\"category\": \"car\", \"corners_3d\": [[499.88, 202.46], [522.28, 203.16], [467.49, 206.01], [444.42, 205.17], [499.88, 173.62], [522.28, 173.64], [467.49, 173.71], [444.42, 173.69]]}, {\"category\": \"car\", \"corners_3d\": [[613.1, 197.01], [633.86, 197.45], [603.0, 199.22], [581.36, 198.72], [613.1, 174.08], [633.86, 174.1], [603.0, 174.19], [581.36, 174.17]]}]\n```" - } - ] - }, - { - "id": 415, - "data_source": "Mono3DRefer", - "scene_name": "000228", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[316.63, 212.96], [342.53, 214.21], [245.65, 219.7], [219.74, 218.1], [316.63, 178.9], [342.53, 179.09], [245.65, 179.92], [219.74, 179.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[866.29, 233.97], [888.1, 234.26], [892.85, 236.06], [870.38, 235.76], [866.29, 165.35], [888.1, 165.32], [892.85, 165.1], [870.38, 165.13]]}, {\"category\": \"car\", \"corners_3d\": [[499.88, 202.46], [522.28, 203.16], [467.49, 206.01], [444.42, 205.17], [499.88, 173.62], [522.28, 173.64], [467.49, 173.71], [444.42, 173.69]]}, {\"category\": \"car\", \"corners_3d\": [[613.1, 197.01], [633.86, 197.45], [603.0, 199.22], [581.36, 198.72], [613.1, 174.08], [633.86, 174.1], [603.0, 174.19], [581.36, 174.17]]}]\n```" - } - ] - }, - { - "id": 416, - "data_source": "Mono3DRefer", - "scene_name": "000228", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[444.27, 201.0], [421.75, 200.35], [479.89, 197.85], [501.55, 198.38], [444.27, 165.68], [421.75, 165.84], [479.89, 166.48], [501.55, 166.34]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[866.29, 233.97], [888.1, 234.26], [892.85, 236.06], [870.38, 235.76], [866.29, 165.35], [888.1, 165.32], [892.85, 165.1], [870.38, 165.13]]}]\n```" - } - ] - }, - { - "id": 417, - "data_source": "Mono3DRefer", - "scene_name": "000228", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[316.63, 212.96], [342.53, 214.21], [245.65, 219.7], [219.74, 218.1], [316.63, 178.9], [342.53, 179.09], [245.65, 179.92], [219.74, 179.67]]}, {\"category\": \"van\", \"corners_3d\": [[444.27, 201.0], [421.75, 200.35], [479.89, 197.85], [501.55, 198.38], [444.27, 165.68], [421.75, 165.84], [479.89, 166.48], [501.55, 166.34]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[866.29, 233.97], [888.1, 234.26], [892.85, 236.06], [870.38, 235.76], [866.29, 165.35], [888.1, 165.32], [892.85, 165.1], [870.38, 165.13]]}, {\"category\": \"car\", \"corners_3d\": [[499.88, 202.46], [522.28, 203.16], [467.49, 206.01], [444.42, 205.17], [499.88, 173.62], [522.28, 173.64], [467.49, 173.71], [444.42, 173.69]]}, {\"category\": \"car\", \"corners_3d\": [[613.1, 197.01], [633.86, 197.45], [603.0, 199.22], [581.36, 198.72], [613.1, 174.08], [633.86, 174.1], [603.0, 174.19], [581.36, 174.17]]}]\n```" - } - ] - }, - { - "id": 418, - "data_source": "Mono3DRefer", - "scene_name": "000230", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[424.05, 206.73], [397.09, 206.75], [414.16, 203.92], [438.87, 203.91], [424.05, 184.57], [397.09, 184.58], [414.16, 183.6], [438.87, 183.6]]}]\n```" - } - ] - }, - { - "id": 419, - "data_source": "Mono3DRefer", - "scene_name": "000230", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[304.06, 214.82], [255.49, 214.85], [313.18, 207.86], [353.63, 207.84], [304.06, 147.47], [255.49, 147.45], [313.18, 151.68], [353.63, 151.69]]}, {\"category\": \"truck\", \"corners_3d\": [[359.83, 208.15], [320.94, 208.15], [360.43, 203.31], [393.99, 203.31], [359.83, 152.75], [320.94, 152.75], [360.43, 155.51], [393.99, 155.51]]}]\n```" - } - ] - }, - { - "id": 420, - "data_source": "Mono3DRefer", - "scene_name": "000230", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[304.06, 214.82], [255.49, 214.85], [313.18, 207.86], [353.63, 207.84], [304.06, 147.47], [255.49, 147.45], [313.18, 151.68], [353.63, 151.69]]}, {\"category\": \"truck\", \"corners_3d\": [[359.83, 208.15], [320.94, 208.15], [360.43, 203.31], [393.99, 203.31], [359.83, 152.75], [320.94, 152.75], [360.43, 155.51], [393.99, 155.51]]}, {\"category\": \"car\", \"corners_3d\": [[424.05, 206.73], [397.09, 206.75], [414.16, 203.92], [438.87, 203.91], [424.05, 184.57], [397.09, 184.58], [414.16, 183.6], [438.87, 183.6]]}]\n```" - } - ] - }, - { - "id": 421, - "data_source": "Mono3DRefer", - "scene_name": "000236", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[583.09, 210.39], [619.73, 210.37], [622.12, 213.52], [580.9, 213.54], [583.09, 177.07], [619.73, 177.08], [622.12, 176.06], [580.9, 176.05]]}, {\"category\": \"car\", \"corners_3d\": [[211.75, 253.85], [288.24, 253.85], [179.48, 277.22], [76.96, 277.21], [211.75, 177.79], [288.24, 177.79], [179.48, 175.26], [76.96, 175.26]]}, {\"category\": \"car\", \"corners_3d\": [[398.24, 257.06], [476.69, 257.14], [429.53, 281.68], [324.44, 281.52], [398.24, 186.12], [476.69, 186.12], [429.53, 186.43], [324.44, 186.43]]}, {\"category\": \"car\", \"corners_3d\": [[338.45, 230.86], [393.45, 230.87], [349.8, 240.16], [283.61, 240.16], [338.45, 184.54], [393.45, 184.54], [349.8, 184.4], [283.61, 184.4]]}, {\"category\": \"car\", \"corners_3d\": [[502.0, 219.29], [543.82, 219.33], [529.72, 225.47], [480.39, 225.41], [502.0, 183.62], [543.82, 183.62], [529.72, 183.34], [480.39, 183.34]]}, {\"category\": \"car\", \"corners_3d\": [[460.92, 205.92], [489.48, 205.92], [477.28, 208.06], [445.77, 208.06], [460.92, 180.67], [489.48, 180.67], [477.28, 180.2], [445.77, 180.2]]}, {\"category\": \"car\", \"corners_3d\": [[76.4, 247.08], [139.12, 247.09], [5.85, 264.68], [-74.69, 264.68], [76.4, 186.42], [139.12, 186.42], [5.85, 186.76], [-74.69, 186.76]]}, {\"category\": \"car\", \"corners_3d\": [[254.72, 225.25], [299.09, 225.25], [237.12, 233.29], [183.84, 233.29], [254.72, 181.72], [299.09, 181.72], [237.12, 181.02], [183.84, 181.02]]}]\n```" - } - ] - }, - { - "id": 422, - "data_source": "Mono3DRefer", - "scene_name": "000236", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[539.07, 197.05], [563.86, 197.03], [561.43, 198.32], [533.94, 198.33], [539.07, 164.59], [563.86, 164.61], [561.43, 162.37], [533.94, 162.35]]}]\n```" - } - ] - }, - { - "id": 423, - "data_source": "Mono3DRefer", - "scene_name": "000236", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[583.09, 210.39], [619.73, 210.37], [622.12, 213.52], [580.9, 213.54], [583.09, 177.07], [619.73, 177.08], [622.12, 176.06], [580.9, 176.05]]}, {\"category\": \"car\", \"corners_3d\": [[211.75, 253.85], [288.24, 253.85], [179.48, 277.22], [76.96, 277.21], [211.75, 177.79], [288.24, 177.79], [179.48, 175.26], [76.96, 175.26]]}, {\"category\": \"car\", \"corners_3d\": [[398.24, 257.06], [476.69, 257.14], [429.53, 281.68], [324.44, 281.52], [398.24, 186.12], [476.69, 186.12], [429.53, 186.43], [324.44, 186.43]]}, {\"category\": \"car\", \"corners_3d\": [[338.45, 230.86], [393.45, 230.87], [349.8, 240.16], [283.61, 240.16], [338.45, 184.54], [393.45, 184.54], [349.8, 184.4], [283.61, 184.4]]}, {\"category\": \"car\", \"corners_3d\": [[502.0, 219.29], [543.82, 219.33], [529.72, 225.47], [480.39, 225.41], [502.0, 183.62], [543.82, 183.62], [529.72, 183.34], [480.39, 183.34]]}, {\"category\": \"car\", \"corners_3d\": [[460.92, 205.92], [489.48, 205.92], [477.28, 208.06], [445.77, 208.06], [460.92, 180.67], [489.48, 180.67], [477.28, 180.2], [445.77, 180.2]]}, {\"category\": \"van\", \"corners_3d\": [[539.07, 197.05], [563.86, 197.03], [561.43, 198.32], [533.94, 198.33], [539.07, 164.59], [563.86, 164.61], [561.43, 162.37], [533.94, 162.35]]}, {\"category\": \"car\", \"corners_3d\": [[76.4, 247.08], [139.12, 247.09], [5.85, 264.68], [-74.69, 264.68], [76.4, 186.42], [139.12, 186.42], [5.85, 186.76], [-74.69, 186.76]]}, {\"category\": \"car\", \"corners_3d\": [[254.72, 225.25], [299.09, 225.25], [237.12, 233.29], [183.84, 233.29], [254.72, 181.72], [299.09, 181.72], [237.12, 181.02], [183.84, 181.02]]}]\n```" - } - ] - }, - { - "id": 424, - "data_source": "Mono3DRefer", - "scene_name": "000237", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[856.29, 212.07], [866.83, 210.48], [952.98, 213.0], [945.61, 214.82], [856.29, 178.52], [866.83, 178.29], [952.98, 178.65], [945.61, 178.91]]}, {\"category\": \"car\", \"corners_3d\": [[211.74, 222.46], [175.08, 222.43], [225.73, 216.74], [258.2, 216.76], [211.74, 194.36], [175.08, 194.35], [225.73, 191.89], [258.2, 191.89]]}]\n```" - } - ] - }, - { - "id": 425, - "data_source": "Mono3DRefer", - "scene_name": "000237", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[592.85, 189.55], [615.95, 189.54], [618.26, 192.49], [591.07, 192.5], [592.85, 164.5], [615.95, 164.51], [618.26, 163.03], [591.07, 163.03]]}]\n```" - } - ] - }, - { - "id": 426, - "data_source": "Mono3DRefer", - "scene_name": "000237", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[592.85, 189.55], [615.95, 189.54], [618.26, 192.49], [591.07, 192.5], [592.85, 164.5], [615.95, 164.51], [618.26, 163.03], [591.07, 163.03]]}, {\"category\": \"car\", \"corners_3d\": [[856.29, 212.07], [866.83, 210.48], [952.98, 213.0], [945.61, 214.82], [856.29, 178.52], [866.83, 178.29], [952.98, 178.65], [945.61, 178.91]]}, {\"category\": \"car\", \"corners_3d\": [[211.74, 222.46], [175.08, 222.43], [225.73, 216.74], [258.2, 216.76], [211.74, 194.36], [175.08, 194.35], [225.73, 191.89], [258.2, 191.89]]}]\n```" - } - ] - }, - { - "id": 427, - "data_source": "Mono3DRefer", - "scene_name": "000238", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[17.78, 212.95], [58.22, 213.04], [-32.96, 219.35], [-79.49, 219.23], [17.78, 170.62], [58.22, 170.61], [-32.96, 170.26], [-79.49, 170.27]]}, {\"category\": \"car\", \"corners_3d\": [[330.5, 194.26], [306.68, 194.22], [324.63, 193.09], [347.22, 193.12], [330.5, 174.59], [306.68, 174.59], [324.63, 174.49], [347.22, 174.5]]}, {\"category\": \"car\", \"corners_3d\": [[578.35, 209.08], [613.29, 209.06], [614.89, 214.65], [574.56, 214.67], [578.35, 174.35], [613.29, 174.35], [614.89, 174.58], [574.56, 174.58]]}]\n```" - } - ] - }, - { - "id": 428, - "data_source": "Mono3DRefer", - "scene_name": "000240", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[610.12, 215.91], [567.82, 215.96], [571.6, 209.86], [607.92, 209.82], [610.12, 178.43], [567.82, 178.43], [571.6, 177.64], [607.92, 177.64]]}]\n```" - } - ] - }, - { - "id": 429, - "data_source": "Mono3DRefer", - "scene_name": "000241", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[682.16, 214.65], [634.28, 213.77], [666.62, 207.58], [707.71, 208.21], [682.16, 155.18], [634.28, 155.56], [666.62, 158.18], [707.71, 157.91]]}]\n```" - } - ] - }, - { - "id": 430, - "data_source": "Mono3DRefer", - "scene_name": "000241", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[552.73, 191.02], [569.44, 190.95], [574.03, 191.95], [556.41, 192.02], [552.73, 174.86], [569.44, 174.85], [574.03, 174.96], [556.41, 174.97]]}]\n```" - } - ] - }, - { - "id": 431, - "data_source": "Mono3DRefer", - "scene_name": "000241", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[682.16, 214.65], [634.28, 213.77], [666.62, 207.58], [707.71, 208.21], [682.16, 155.18], [634.28, 155.56], [666.62, 158.18], [707.71, 157.91]]}, {\"category\": \"car\", \"corners_3d\": [[552.73, 191.02], [569.44, 190.95], [574.03, 191.95], [556.41, 192.02], [552.73, 174.86], [569.44, 174.85], [574.03, 174.96], [556.41, 174.97]]}]\n```" - } - ] - }, - { - "id": 432, - "data_source": "Mono3DRefer", - "scene_name": "000242", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[725.78, 229.56], [665.06, 230.31], [635.49, 220.02], [685.76, 219.51], [725.78, 164.89], [665.06, 164.79], [635.49, 166.23], [685.76, 166.3]]}, {\"category\": \"car\", \"corners_3d\": [[444.22, 238.04], [389.77, 238.67], [410.59, 227.7], [455.86, 227.26], [444.22, 186.66], [389.77, 186.79], [410.59, 184.47], [455.86, 184.38]]}, {\"category\": \"car\", \"corners_3d\": [[472.33, 220.37], [433.0, 220.65], [443.63, 215.37], [478.58, 215.15], [472.33, 188.6], [433.0, 188.7], [443.63, 186.95], [478.58, 186.87]]}]\n```" - } - ] - }, - { - "id": 433, - "data_source": "Mono3DRefer", - "scene_name": "000243", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.01, 233.44], [666.51, 233.5], [678.78, 248.63], [599.44, 248.54], [603.01, 177.72], [666.51, 177.72], [678.78, 178.94], [599.44, 178.93]]}, {\"category\": \"car\", \"corners_3d\": [[524.69, 220.85], [576.56, 220.99], [562.66, 231.71], [499.34, 231.5], [524.69, 171.3], [576.56, 171.3], [562.66, 170.95], [499.34, 170.96]]}, {\"category\": \"car\", \"corners_3d\": [[626.37, 211.35], [669.8, 211.49], [673.05, 219.18], [621.0, 218.98], [626.37, 170.72], [669.8, 170.71], [673.05, 170.29], [621.0, 170.3]]}, {\"category\": \"car\", \"corners_3d\": [[1314.13, 200.06], [1404.53, 202.78], [1275.64, 204.91], [1191.89, 201.8], [1314.13, 137.11], [1404.53, 133.54], [1275.64, 130.75], [1191.89, 134.82]]}, {\"category\": \"car\", \"corners_3d\": [[1175.24, 195.22], [1238.1, 197.08], [1128.65, 198.35], [1071.55, 196.3], [1175.24, 140.15], [1238.1, 137.44], [1128.65, 135.58], [1071.55, 138.58]]}]\n```" - } - ] - }, - { - "id": 434, - "data_source": "Mono3DRefer", - "scene_name": "000244", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[553.7, 223.64], [514.17, 223.53], [528.31, 217.74], [563.34, 217.83], [553.7, 183.62], [514.17, 183.6], [528.31, 182.37], [563.34, 182.39]]}, {\"category\": \"car\", \"corners_3d\": [[576.04, 214.34], [540.24, 214.3], [549.57, 209.65], [581.36, 209.69], [576.04, 173.06], [540.24, 173.06], [549.57, 173.04], [581.36, 173.04]]}, {\"category\": \"car\", \"corners_3d\": [[592.81, 209.9], [561.98, 209.81], [570.81, 206.19], [598.64, 206.27], [592.81, 176.33], [561.98, 176.32], [570.81, 175.98], [598.64, 175.99]]}, {\"category\": \"car\", \"corners_3d\": [[119.54, 669.72], [-333.7, 663.64], [238.06, 370.45], [422.03, 371.43], [119.54, 196.0], [-333.7, 195.72], [238.06, 182.06], [422.03, 182.1]]}, {\"category\": \"car\", \"corners_3d\": [[953.34, 479.15], [1216.69, 475.84], [4277.52, 1952.07], [2851.55, 2072.7], [953.34, 190.92], [1216.69, 190.72], [4277.52, 277.78], [2851.55, 284.89]]}, {\"category\": \"car\", \"corners_3d\": [[731.05, 307.75], [852.63, 307.54], [1000.1, 387.09], [806.9, 387.62], [731.05, 190.36], [852.63, 190.33], [1000.1, 200.65], [806.9, 200.72]]}, {\"category\": \"car\", \"corners_3d\": [[390.81, 286.21], [487.58, 285.48], [449.36, 335.27], [309.04, 336.81], [390.81, 187.56], [487.58, 187.47], [449.36, 193.93], [309.04, 194.13]]}, {\"category\": \"car\", \"corners_3d\": [[502.44, 272.88], [411.58, 273.02], [450.35, 252.6], [522.67, 252.51], [502.44, 192.84], [411.58, 192.87], [450.35, 188.79], [522.67, 188.77]]}, {\"category\": \"car\", \"corners_3d\": [[685.96, 251.59], [754.78, 250.99], [806.33, 270.89], [720.32, 271.83], [685.96, 181.44], [754.78, 181.37], [806.33, 183.54], [720.32, 183.64]]}, {\"category\": \"car\", \"corners_3d\": [[657.08, 233.11], [711.09, 233.02], [736.0, 246.01], [670.35, 246.13], [657.08, 183.0], [711.09, 182.98], [736.0, 185.17], [670.35, 185.19]]}]\n```" - } - ] - }, - { - "id": 435, - "data_source": "Mono3DRefer", - "scene_name": "000245", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[594.19, 215.97], [610.16, 215.94], [611.91, 219.41], [594.65, 219.44], [594.19, 172.59], [610.16, 172.59], [611.91, 172.57], [594.65, 172.57]]}]\n```" - } - ] - }, - { - "id": 436, - "data_source": "Mono3DRefer", - "scene_name": "000245", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[957.96, 368.25], [873.4, 324.59], [1273.95, 321.03], [1467.78, 362.38], [957.96, 139.41], [873.4, 146.89], [1273.95, 147.49], [1467.78, 140.42]]}, {\"category\": \"car\", \"corners_3d\": [[-201.14, 338.87], [-68.95, 311.27], [230.46, 310.06], [158.74, 337.13], [-201.14, 206.95], [-68.95, 201.28], [230.46, 201.04], [158.74, 206.6]]}, {\"category\": \"car\", \"corners_3d\": [[175.11, 263.15], [210.59, 254.28], [369.08, 252.55], [351.3, 261.04], [175.11, 193.83], [210.59, 191.77], [369.08, 191.37], [351.3, 193.34]]}, {\"category\": \"car\", \"corners_3d\": [[363.55, 221.41], [374.44, 218.78], [466.75, 218.43], [461.18, 221.01], [363.55, 183.58], [374.44, 183.0], [466.75, 182.92], [461.18, 183.49]]}]\n```" - } - ] - }, - { - "id": 437, - "data_source": "Mono3DRefer", - "scene_name": "000245", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[957.96, 368.25], [873.4, 324.59], [1273.95, 321.03], [1467.78, 362.38], [957.96, 139.41], [873.4, 146.89], [1273.95, 147.49], [1467.78, 140.42]]}, {\"category\": \"cyclist\", \"corners_3d\": [[594.19, 215.97], [610.16, 215.94], [611.91, 219.41], [594.65, 219.44], [594.19, 172.59], [610.16, 172.59], [611.91, 172.57], [594.65, 172.57]]}, {\"category\": \"car\", \"corners_3d\": [[-201.14, 338.87], [-68.95, 311.27], [230.46, 310.06], [158.74, 337.13], [-201.14, 206.95], [-68.95, 201.28], [230.46, 201.04], [158.74, 206.6]]}, {\"category\": \"car\", \"corners_3d\": [[175.11, 263.15], [210.59, 254.28], [369.08, 252.55], [351.3, 261.04], [175.11, 193.83], [210.59, 191.77], [369.08, 191.37], [351.3, 193.34]]}, {\"category\": \"car\", \"corners_3d\": [[363.55, 221.41], [374.44, 218.78], [466.75, 218.43], [461.18, 221.01], [363.55, 183.58], [374.44, 183.0], [466.75, 182.92], [461.18, 183.49]]}]\n```" - } - ] - }, - { - "id": 438, - "data_source": "Mono3DRefer", - "scene_name": "000247", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[309.19, 218.15], [275.82, 218.87], [279.67, 214.92], [310.16, 214.32], [309.19, 191.0], [275.82, 191.29], [279.67, 189.71], [310.16, 189.47]]}]\n```" - } - ] - }, - { - "id": 439, - "data_source": "Mono3DRefer", - "scene_name": "000248", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[579.18, 206.86], [609.67, 206.84], [611.33, 209.04], [577.73, 209.06], [579.18, 179.13], [609.67, 179.14], [611.33, 178.52], [577.73, 178.51]]}, {\"category\": \"car\", \"corners_3d\": [[162.65, 262.75], [245.44, 262.67], [110.12, 292.19], [-4.47, 292.34], [162.65, 183.72], [245.44, 183.72], [110.12, 183.15], [-4.47, 183.14]]}, {\"category\": \"car\", \"corners_3d\": [[361.4, 263.07], [441.32, 263.07], [376.33, 293.48], [265.21, 293.46], [361.4, 191.74], [441.32, 191.74], [376.33, 194.28], [265.21, 194.28]]}, {\"category\": \"car\", \"corners_3d\": [[317.12, 230.71], [369.41, 230.61], [328.42, 239.2], [266.12, 239.33], [317.12, 192.4], [369.41, 192.39], [328.42, 193.74], [266.12, 193.76]]}, {\"category\": \"car\", \"corners_3d\": [[456.2, 226.62], [502.2, 226.62], [484.8, 233.45], [431.21, 233.45], [456.2, 186.04], [502.2, 186.04], [484.8, 186.18], [431.21, 186.18]]}, {\"category\": \"car\", \"corners_3d\": [[405.15, 215.8], [439.76, 215.77], [420.88, 219.52], [381.99, 219.56], [405.15, 181.34], [439.76, 181.34], [420.88, 180.87], [381.99, 180.86]]}, {\"category\": \"car\", \"corners_3d\": [[488.43, 212.56], [521.71, 212.55], [511.72, 216.01], [474.22, 216.03], [488.43, 183.57], [521.71, 183.57], [511.72, 183.36], [474.22, 183.36]]}, {\"category\": \"car\", \"corners_3d\": [[514.99, 207.31], [540.36, 207.3], [534.87, 209.32], [507.19, 209.32], [514.99, 185.52], [540.36, 185.52], [534.87, 185.55], [507.19, 185.55]]}]\n```" - } - ] - }, - { - "id": 440, - "data_source": "Mono3DRefer", - "scene_name": "000248", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[298.07, 219.03], [340.63, 219.02], [298.02, 224.56], [248.45, 224.58], [298.07, 173.83], [340.63, 173.84], [298.02, 171.97], [248.45, 171.96]]}]\n```" - } - ] - }, - { - "id": 441, - "data_source": "Mono3DRefer", - "scene_name": "000248", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[579.18, 206.86], [609.67, 206.84], [611.33, 209.04], [577.73, 209.06], [579.18, 179.13], [609.67, 179.14], [611.33, 178.52], [577.73, 178.51]]}, {\"category\": \"car\", \"corners_3d\": [[162.65, 262.75], [245.44, 262.67], [110.12, 292.19], [-4.47, 292.34], [162.65, 183.72], [245.44, 183.72], [110.12, 183.15], [-4.47, 183.14]]}, {\"category\": \"car\", \"corners_3d\": [[361.4, 263.07], [441.32, 263.07], [376.33, 293.48], [265.21, 293.46], [361.4, 191.74], [441.32, 191.74], [376.33, 194.28], [265.21, 194.28]]}, {\"category\": \"car\", \"corners_3d\": [[317.12, 230.71], [369.41, 230.61], [328.42, 239.2], [266.12, 239.33], [317.12, 192.4], [369.41, 192.39], [328.42, 193.74], [266.12, 193.76]]}, {\"category\": \"car\", \"corners_3d\": [[456.2, 226.62], [502.2, 226.62], [484.8, 233.45], [431.21, 233.45], [456.2, 186.04], [502.2, 186.04], [484.8, 186.18], [431.21, 186.18]]}, {\"category\": \"car\", \"corners_3d\": [[405.15, 215.8], [439.76, 215.77], [420.88, 219.52], [381.99, 219.56], [405.15, 181.34], [439.76, 181.34], [420.88, 180.87], [381.99, 180.86]]}, {\"category\": \"car\", \"corners_3d\": [[488.43, 212.56], [521.71, 212.55], [511.72, 216.01], [474.22, 216.03], [488.43, 183.57], [521.71, 183.57], [511.72, 183.36], [474.22, 183.36]]}, {\"category\": \"car\", \"corners_3d\": [[514.99, 207.31], [540.36, 207.3], [534.87, 209.32], [507.19, 209.32], [514.99, 185.52], [540.36, 185.52], [534.87, 185.55], [507.19, 185.55]]}, {\"category\": \"van\", \"corners_3d\": [[298.07, 219.03], [340.63, 219.02], [298.02, 224.56], [248.45, 224.58], [298.07, 173.83], [340.63, 173.84], [298.02, 171.97], [248.45, 171.96]]}]\n```" - } - ] - }, - { - "id": 442, - "data_source": "Mono3DRefer", - "scene_name": "000249", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[524.29, 193.3], [546.61, 193.28], [547.55, 195.32], [519.56, 195.36], [524.29, 157.27], [546.61, 157.34], [547.55, 150.28], [519.56, 150.16]]}]\n```" - } - ] - }, - { - "id": 443, - "data_source": "Mono3DRefer", - "scene_name": "000249", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[561.67, 212.07], [600.1, 212.0], [603.82, 215.52], [560.33, 215.61], [561.67, 177.11], [600.1, 177.13], [603.82, 176.06], [560.33, 176.04]]}, {\"category\": \"car\", \"corners_3d\": [[298.95, 303.14], [425.58, 302.34], [320.15, 382.84], [104.27, 385.13], [298.95, 185.13], [425.58, 185.13], [320.15, 185.07], [104.27, 185.07]]}, {\"category\": \"car\", \"corners_3d\": [[-23.43, 305.06], [92.87, 304.51], [-283.9, 395.61], [-492.83, 397.33], [-23.43, 199.28], [92.87, 199.22], [-283.9, 209.91], [-492.83, 210.12]]}, {\"category\": \"car\", \"corners_3d\": [[404.74, 247.59], [467.8, 247.49], [436.84, 262.84], [358.13, 263.0], [404.74, 189.36], [467.8, 189.35], [436.84, 190.37], [358.13, 190.38]]}, {\"category\": \"car\", \"corners_3d\": [[271.42, 239.43], [331.45, 239.3], [264.76, 253.46], [188.77, 253.67], [271.42, 187.01], [331.45, 187.0], [264.76, 187.47], [188.77, 187.48]]}, {\"category\": \"car\", \"corners_3d\": [[181.99, 229.99], [229.04, 229.91], [154.18, 239.27], [97.08, 239.4], [181.99, 187.03], [229.04, 187.03], [154.18, 187.41], [97.08, 187.42]]}]\n```" - } - ] - }, - { - "id": 444, - "data_source": "Mono3DRefer", - "scene_name": "000249", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[412.05, 212.44], [450.81, 212.34], [427.27, 218.34], [379.81, 218.49], [412.05, 160.75], [450.81, 160.84], [427.27, 155.45], [379.81, 155.32]]}]\n```" - } - ] - }, - { - "id": 445, - "data_source": "Mono3DRefer", - "scene_name": "000249", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[561.67, 212.07], [600.1, 212.0], [603.82, 215.52], [560.33, 215.61], [561.67, 177.11], [600.1, 177.13], [603.82, 176.06], [560.33, 176.04]]}, {\"category\": \"car\", \"corners_3d\": [[298.95, 303.14], [425.58, 302.34], [320.15, 382.84], [104.27, 385.13], [298.95, 185.13], [425.58, 185.13], [320.15, 185.07], [104.27, 185.07]]}, {\"category\": \"car\", \"corners_3d\": [[-23.43, 305.06], [92.87, 304.51], [-283.9, 395.61], [-492.83, 397.33], [-23.43, 199.28], [92.87, 199.22], [-283.9, 209.91], [-492.83, 210.12]]}, {\"category\": \"car\", \"corners_3d\": [[404.74, 247.59], [467.8, 247.49], [436.84, 262.84], [358.13, 263.0], [404.74, 189.36], [467.8, 189.35], [436.84, 190.37], [358.13, 190.38]]}, {\"category\": \"car\", \"corners_3d\": [[271.42, 239.43], [331.45, 239.3], [264.76, 253.46], [188.77, 253.67], [271.42, 187.01], [331.45, 187.0], [264.76, 187.47], [188.77, 187.48]]}, {\"category\": \"car\", \"corners_3d\": [[181.99, 229.99], [229.04, 229.91], [154.18, 239.27], [97.08, 239.4], [181.99, 187.03], [229.04, 187.03], [154.18, 187.41], [97.08, 187.42]]}, {\"category\": \"truck\", \"corners_3d\": [[524.29, 193.3], [546.61, 193.28], [547.55, 195.32], [519.56, 195.36], [524.29, 157.27], [546.61, 157.34], [547.55, 150.28], [519.56, 150.16]]}]\n```" - } - ] - }, - { - "id": 446, - "data_source": "Mono3DRefer", - "scene_name": "000249", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[412.05, 212.44], [450.81, 212.34], [427.27, 218.34], [379.81, 218.49], [412.05, 160.75], [450.81, 160.84], [427.27, 155.45], [379.81, 155.32]]}, {\"category\": \"truck\", \"corners_3d\": [[524.29, 193.3], [546.61, 193.28], [547.55, 195.32], [519.56, 195.36], [524.29, 157.27], [546.61, 157.34], [547.55, 150.28], [519.56, 150.16]]}]\n```" - } - ] - }, - { - "id": 447, - "data_source": "Mono3DRefer", - "scene_name": "000249", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[561.67, 212.07], [600.1, 212.0], [603.82, 215.52], [560.33, 215.61], [561.67, 177.11], [600.1, 177.13], [603.82, 176.06], [560.33, 176.04]]}, {\"category\": \"car\", \"corners_3d\": [[298.95, 303.14], [425.58, 302.34], [320.15, 382.84], [104.27, 385.13], [298.95, 185.13], [425.58, 185.13], [320.15, 185.07], [104.27, 185.07]]}, {\"category\": \"car\", \"corners_3d\": [[-23.43, 305.06], [92.87, 304.51], [-283.9, 395.61], [-492.83, 397.33], [-23.43, 199.28], [92.87, 199.22], [-283.9, 209.91], [-492.83, 210.12]]}, {\"category\": \"car\", \"corners_3d\": [[404.74, 247.59], [467.8, 247.49], [436.84, 262.84], [358.13, 263.0], [404.74, 189.36], [467.8, 189.35], [436.84, 190.37], [358.13, 190.38]]}, {\"category\": \"car\", \"corners_3d\": [[271.42, 239.43], [331.45, 239.3], [264.76, 253.46], [188.77, 253.67], [271.42, 187.01], [331.45, 187.0], [264.76, 187.47], [188.77, 187.48]]}, {\"category\": \"car\", \"corners_3d\": [[181.99, 229.99], [229.04, 229.91], [154.18, 239.27], [97.08, 239.4], [181.99, 187.03], [229.04, 187.03], [154.18, 187.41], [97.08, 187.42]]}, {\"category\": \"van\", \"corners_3d\": [[412.05, 212.44], [450.81, 212.34], [427.27, 218.34], [379.81, 218.49], [412.05, 160.75], [450.81, 160.84], [427.27, 155.45], [379.81, 155.32]]}]\n```" - } - ] - }, - { - "id": 448, - "data_source": "Mono3DRefer", - "scene_name": "000249", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[561.67, 212.07], [600.1, 212.0], [603.82, 215.52], [560.33, 215.61], [561.67, 177.11], [600.1, 177.13], [603.82, 176.06], [560.33, 176.04]]}, {\"category\": \"car\", \"corners_3d\": [[298.95, 303.14], [425.58, 302.34], [320.15, 382.84], [104.27, 385.13], [298.95, 185.13], [425.58, 185.13], [320.15, 185.07], [104.27, 185.07]]}, {\"category\": \"car\", \"corners_3d\": [[-23.43, 305.06], [92.87, 304.51], [-283.9, 395.61], [-492.83, 397.33], [-23.43, 199.28], [92.87, 199.22], [-283.9, 209.91], [-492.83, 210.12]]}, {\"category\": \"car\", \"corners_3d\": [[404.74, 247.59], [467.8, 247.49], [436.84, 262.84], [358.13, 263.0], [404.74, 189.36], [467.8, 189.35], [436.84, 190.37], [358.13, 190.38]]}, {\"category\": \"car\", \"corners_3d\": [[271.42, 239.43], [331.45, 239.3], [264.76, 253.46], [188.77, 253.67], [271.42, 187.01], [331.45, 187.0], [264.76, 187.47], [188.77, 187.48]]}, {\"category\": \"car\", \"corners_3d\": [[181.99, 229.99], [229.04, 229.91], [154.18, 239.27], [97.08, 239.4], [181.99, 187.03], [229.04, 187.03], [154.18, 187.41], [97.08, 187.42]]}, {\"category\": \"van\", \"corners_3d\": [[412.05, 212.44], [450.81, 212.34], [427.27, 218.34], [379.81, 218.49], [412.05, 160.75], [450.81, 160.84], [427.27, 155.45], [379.81, 155.32]]}, {\"category\": \"truck\", \"corners_3d\": [[524.29, 193.3], [546.61, 193.28], [547.55, 195.32], [519.56, 195.36], [524.29, 157.27], [546.61, 157.34], [547.55, 150.28], [519.56, 150.16]]}]\n```" - } - ] - }, - { - "id": 449, - "data_source": "Mono3DRefer", - "scene_name": "000250", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[200.1, 336.46], [27.23, 347.42], [137.49, 291.24], [252.35, 286.09], [200.1, 205.55], [27.23, 207.74], [137.49, 196.51], [252.35, 195.48]]}, {\"category\": \"car\", \"corners_3d\": [[376.36, 224.6], [341.76, 225.1], [352.54, 220.13], [383.79, 219.72], [376.36, 194.63], [341.76, 194.83], [352.54, 192.75], [383.79, 192.57]]}, {\"category\": \"car\", \"corners_3d\": [[318.81, 271.0], [232.66, 273.72], [268.32, 254.99], [338.03, 253.18], [318.81, 198.98], [232.66, 199.7], [268.32, 194.72], [338.03, 194.24]]}]\n```" - } - ] - }, - { - "id": 450, - "data_source": "Mono3DRefer", - "scene_name": "000252", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[568.2, 189.89], [586.7, 189.66], [612.94, 190.85], [593.5, 191.11], [568.2, 162.14], [586.7, 162.28], [612.94, 161.54], [593.5, 161.37]]}]\n```" - } - ] - }, - { - "id": 451, - "data_source": "Mono3DRefer", - "scene_name": "000252", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[373.62, 328.27], [240.76, 328.0], [354.12, 281.16], [446.95, 281.29], [373.62, 206.2], [240.76, 206.14], [354.12, 196.09], [446.95, 196.12]]}, {\"category\": \"car\", \"corners_3d\": [[546.56, 209.51], [515.88, 209.59], [521.89, 205.71], [549.33, 205.65], [546.56, 184.57], [515.88, 184.6], [521.89, 183.36], [549.33, 183.34]]}, {\"category\": \"car\", \"corners_3d\": [[389.93, 218.65], [386.8, 216.82], [479.43, 214.76], [486.06, 216.41], [389.93, 183.66], [386.8, 183.23], [479.43, 182.75], [486.06, 183.14]]}]\n```" - } - ] - }, - { - "id": 452, - "data_source": "Mono3DRefer", - "scene_name": "000252", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[373.62, 328.27], [240.76, 328.0], [354.12, 281.16], [446.95, 281.29], [373.62, 206.2], [240.76, 206.14], [354.12, 196.09], [446.95, 196.12]]}, {\"category\": \"car\", \"corners_3d\": [[546.56, 209.51], [515.88, 209.59], [521.89, 205.71], [549.33, 205.65], [546.56, 184.57], [515.88, 184.6], [521.89, 183.36], [549.33, 183.34]]}, {\"category\": \"van\", \"corners_3d\": [[568.2, 189.89], [586.7, 189.66], [612.94, 190.85], [593.5, 191.11], [568.2, 162.14], [586.7, 162.28], [612.94, 161.54], [593.5, 161.37]]}, {\"category\": \"car\", \"corners_3d\": [[389.93, 218.65], [386.8, 216.82], [479.43, 214.76], [486.06, 216.41], [389.93, 183.66], [386.8, 183.23], [479.43, 182.75], [486.06, 183.14]]}]\n```" - } - ] - }, - { - "id": 453, - "data_source": "Mono3DRefer", - "scene_name": "000253", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[574.63, 223.12], [622.59, 223.12], [624.43, 230.51], [569.41, 230.5], [574.63, 176.6], [622.59, 176.6], [624.43, 177.15], [569.41, 177.15]]}, {\"category\": \"car\", \"corners_3d\": [[397.02, 251.61], [330.49, 251.54], [386.25, 236.19], [439.83, 236.23], [397.02, 192.03], [330.49, 192.02], [386.25, 188.28], [439.83, 188.29]]}, {\"category\": \"car\", \"corners_3d\": [[639.65, 193.13], [656.74, 193.13], [659.37, 194.27], [641.31, 194.27], [639.65, 176.84], [656.74, 176.84], [659.37, 177.06], [641.31, 177.06]]}]\n```" - } - ] - }, - { - "id": 454, - "data_source": "Mono3DRefer", - "scene_name": "000254", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[785.86, 275.56], [783.5, 270.84], [890.94, 273.27], [898.67, 278.23], [785.86, 179.64], [783.5, 179.33], [890.94, 179.49], [898.67, 179.82]]}, {\"category\": \"cyclist\", \"corners_3d\": [[816.34, 287.33], [820.27, 280.52], [939.13, 286.57], [942.93, 294.19], [816.34, 177.78], [820.27, 177.48], [939.13, 177.74], [942.93, 178.07]]}]\n```" - } - ] - }, - { - "id": 455, - "data_source": "Mono3DRefer", - "scene_name": "000254", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[168.98, 344.9], [27.97, 345.91], [203.37, 291.64], [299.64, 291.16], [168.98, 213.58], [27.97, 213.82], [203.37, 200.97], [299.64, 200.86]]}, {\"category\": \"car\", \"corners_3d\": [[473.56, 277.13], [382.13, 277.66], [429.39, 252.64], [498.9, 252.33], [473.56, 195.7], [382.13, 195.82], [429.39, 190.34], [498.9, 190.27]]}, {\"category\": \"car\", \"corners_3d\": [[321.66, 288.07], [227.28, 288.38], [313.21, 261.35], [385.38, 261.17], [321.66, 201.79], [227.28, 201.87], [313.21, 195.08], [385.38, 195.03]]}, {\"category\": \"car\", \"corners_3d\": [[1170.78, 253.51], [1111.55, 244.52], [1327.75, 245.39], [1415.29, 254.62], [1170.78, 152.37], [1111.55, 154.65], [1327.75, 154.43], [1415.29, 152.09]]}, {\"category\": \"car\", \"corners_3d\": [[540.89, 225.63], [498.27, 225.72], [509.67, 218.95], [546.82, 218.88], [540.89, 183.98], [498.27, 184.0], [509.67, 182.58], [546.82, 182.56]]}, {\"category\": \"car\", \"corners_3d\": [[567.36, 201.09], [549.13, 201.1], [551.71, 199.5], [568.9, 199.48], [567.36, 183.13], [549.13, 183.13], [551.71, 182.55], [568.9, 182.55]]}, {\"category\": \"car\", \"corners_3d\": [[718.22, 212.33], [724.64, 214.27], [635.45, 214.43], [633.21, 212.47], [718.22, 179.02], [724.64, 179.32], [635.45, 179.34], [633.21, 179.04]]}, {\"category\": \"car\", \"corners_3d\": [[675.19, 202.62], [676.29, 203.73], [613.05, 203.6], [614.23, 202.49], [675.19, 177.59], [676.29, 177.77], [613.05, 177.74], [614.23, 177.57]]}]\n```" - } - ] - }, - { - "id": 456, - "data_source": "Mono3DRefer", - "scene_name": "000254", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[517.54, 244.57], [453.55, 244.7], [480.29, 231.08], [532.12, 230.99], [517.54, 173.57], [453.55, 173.57], [480.29, 173.43], [532.12, 173.43]]}]\n```" - } - ] - }, - { - "id": 457, - "data_source": "Mono3DRefer", - "scene_name": "000254", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[168.98, 344.9], [27.97, 345.91], [203.37, 291.64], [299.64, 291.16], [168.98, 213.58], [27.97, 213.82], [203.37, 200.97], [299.64, 200.86]]}, {\"category\": \"car\", \"corners_3d\": [[473.56, 277.13], [382.13, 277.66], [429.39, 252.64], [498.9, 252.33], [473.56, 195.7], [382.13, 195.82], [429.39, 190.34], [498.9, 190.27]]}, {\"category\": \"car\", \"corners_3d\": [[321.66, 288.07], [227.28, 288.38], [313.21, 261.35], [385.38, 261.17], [321.66, 201.79], [227.28, 201.87], [313.21, 195.08], [385.38, 195.03]]}, {\"category\": \"cyclist\", \"corners_3d\": [[785.86, 275.56], [783.5, 270.84], [890.94, 273.27], [898.67, 278.23], [785.86, 179.64], [783.5, 179.33], [890.94, 179.49], [898.67, 179.82]]}, {\"category\": \"car\", \"corners_3d\": [[1170.78, 253.51], [1111.55, 244.52], [1327.75, 245.39], [1415.29, 254.62], [1170.78, 152.37], [1111.55, 154.65], [1327.75, 154.43], [1415.29, 152.09]]}, {\"category\": \"cyclist\", \"corners_3d\": [[816.34, 287.33], [820.27, 280.52], [939.13, 286.57], [942.93, 294.19], [816.34, 177.78], [820.27, 177.48], [939.13, 177.74], [942.93, 178.07]]}, {\"category\": \"car\", \"corners_3d\": [[540.89, 225.63], [498.27, 225.72], [509.67, 218.95], [546.82, 218.88], [540.89, 183.98], [498.27, 184.0], [509.67, 182.58], [546.82, 182.56]]}, {\"category\": \"car\", \"corners_3d\": [[567.36, 201.09], [549.13, 201.1], [551.71, 199.5], [568.9, 199.48], [567.36, 183.13], [549.13, 183.13], [551.71, 182.55], [568.9, 182.55]]}, {\"category\": \"car\", \"corners_3d\": [[718.22, 212.33], [724.64, 214.27], [635.45, 214.43], [633.21, 212.47], [718.22, 179.02], [724.64, 179.32], [635.45, 179.34], [633.21, 179.04]]}, {\"category\": \"car\", \"corners_3d\": [[675.19, 202.62], [676.29, 203.73], [613.05, 203.6], [614.23, 202.49], [675.19, 177.59], [676.29, 177.77], [613.05, 177.74], [614.23, 177.57]]}]\n```" - } - ] - }, - { - "id": 458, - "data_source": "Mono3DRefer", - "scene_name": "000254", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[785.86, 275.56], [783.5, 270.84], [890.94, 273.27], [898.67, 278.23], [785.86, 179.64], [783.5, 179.33], [890.94, 179.49], [898.67, 179.82]]}, {\"category\": \"cyclist\", \"corners_3d\": [[816.34, 287.33], [820.27, 280.52], [939.13, 286.57], [942.93, 294.19], [816.34, 177.78], [820.27, 177.48], [939.13, 177.74], [942.93, 178.07]]}, {\"category\": \"van\", \"corners_3d\": [[517.54, 244.57], [453.55, 244.7], [480.29, 231.08], [532.12, 230.99], [517.54, 173.57], [453.55, 173.57], [480.29, 173.43], [532.12, 173.43]]}]\n```" - } - ] - }, - { - "id": 459, - "data_source": "Mono3DRefer", - "scene_name": "000254", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[168.98, 344.9], [27.97, 345.91], [203.37, 291.64], [299.64, 291.16], [168.98, 213.58], [27.97, 213.82], [203.37, 200.97], [299.64, 200.86]]}, {\"category\": \"car\", \"corners_3d\": [[473.56, 277.13], [382.13, 277.66], [429.39, 252.64], [498.9, 252.33], [473.56, 195.7], [382.13, 195.82], [429.39, 190.34], [498.9, 190.27]]}, {\"category\": \"car\", \"corners_3d\": [[321.66, 288.07], [227.28, 288.38], [313.21, 261.35], [385.38, 261.17], [321.66, 201.79], [227.28, 201.87], [313.21, 195.08], [385.38, 195.03]]}, {\"category\": \"car\", \"corners_3d\": [[1170.78, 253.51], [1111.55, 244.52], [1327.75, 245.39], [1415.29, 254.62], [1170.78, 152.37], [1111.55, 154.65], [1327.75, 154.43], [1415.29, 152.09]]}, {\"category\": \"car\", \"corners_3d\": [[540.89, 225.63], [498.27, 225.72], [509.67, 218.95], [546.82, 218.88], [540.89, 183.98], [498.27, 184.0], [509.67, 182.58], [546.82, 182.56]]}, {\"category\": \"car\", \"corners_3d\": [[567.36, 201.09], [549.13, 201.1], [551.71, 199.5], [568.9, 199.48], [567.36, 183.13], [549.13, 183.13], [551.71, 182.55], [568.9, 182.55]]}, {\"category\": \"car\", \"corners_3d\": [[718.22, 212.33], [724.64, 214.27], [635.45, 214.43], [633.21, 212.47], [718.22, 179.02], [724.64, 179.32], [635.45, 179.34], [633.21, 179.04]]}, {\"category\": \"car\", \"corners_3d\": [[675.19, 202.62], [676.29, 203.73], [613.05, 203.6], [614.23, 202.49], [675.19, 177.59], [676.29, 177.77], [613.05, 177.74], [614.23, 177.57]]}, {\"category\": \"van\", \"corners_3d\": [[517.54, 244.57], [453.55, 244.7], [480.29, 231.08], [532.12, 230.99], [517.54, 173.57], [453.55, 173.57], [480.29, 173.43], [532.12, 173.43]]}]\n```" - } - ] - }, - { - "id": 460, - "data_source": "Mono3DRefer", - "scene_name": "000254", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[168.98, 344.9], [27.97, 345.91], [203.37, 291.64], [299.64, 291.16], [168.98, 213.58], [27.97, 213.82], [203.37, 200.97], [299.64, 200.86]]}, {\"category\": \"car\", \"corners_3d\": [[473.56, 277.13], [382.13, 277.66], [429.39, 252.64], [498.9, 252.33], [473.56, 195.7], [382.13, 195.82], [429.39, 190.34], [498.9, 190.27]]}, {\"category\": \"car\", \"corners_3d\": [[321.66, 288.07], [227.28, 288.38], [313.21, 261.35], [385.38, 261.17], [321.66, 201.79], [227.28, 201.87], [313.21, 195.08], [385.38, 195.03]]}, {\"category\": \"cyclist\", \"corners_3d\": [[785.86, 275.56], [783.5, 270.84], [890.94, 273.27], [898.67, 278.23], [785.86, 179.64], [783.5, 179.33], [890.94, 179.49], [898.67, 179.82]]}, {\"category\": \"car\", \"corners_3d\": [[1170.78, 253.51], [1111.55, 244.52], [1327.75, 245.39], [1415.29, 254.62], [1170.78, 152.37], [1111.55, 154.65], [1327.75, 154.43], [1415.29, 152.09]]}, {\"category\": \"cyclist\", \"corners_3d\": [[816.34, 287.33], [820.27, 280.52], [939.13, 286.57], [942.93, 294.19], [816.34, 177.78], [820.27, 177.48], [939.13, 177.74], [942.93, 178.07]]}, {\"category\": \"car\", \"corners_3d\": [[540.89, 225.63], [498.27, 225.72], [509.67, 218.95], [546.82, 218.88], [540.89, 183.98], [498.27, 184.0], [509.67, 182.58], [546.82, 182.56]]}, {\"category\": \"car\", \"corners_3d\": [[567.36, 201.09], [549.13, 201.1], [551.71, 199.5], [568.9, 199.48], [567.36, 183.13], [549.13, 183.13], [551.71, 182.55], [568.9, 182.55]]}, {\"category\": \"car\", \"corners_3d\": [[718.22, 212.33], [724.64, 214.27], [635.45, 214.43], [633.21, 212.47], [718.22, 179.02], [724.64, 179.32], [635.45, 179.34], [633.21, 179.04]]}, {\"category\": \"car\", \"corners_3d\": [[675.19, 202.62], [676.29, 203.73], [613.05, 203.6], [614.23, 202.49], [675.19, 177.59], [676.29, 177.77], [613.05, 177.74], [614.23, 177.57]]}, {\"category\": \"van\", \"corners_3d\": [[517.54, 244.57], [453.55, 244.7], [480.29, 231.08], [532.12, 230.99], [517.54, 173.57], [453.55, 173.57], [480.29, 173.43], [532.12, 173.43]]}]\n```" - } - ] - }, - { - "id": 461, - "data_source": "Mono3DRefer", - "scene_name": "000256", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-67.16, 248.69], [-125.88, 248.69], [-8.58, 236.59], [40.75, 236.58], [-67.16, 187.51], [-125.88, 187.51], [-8.58, 185.17], [40.75, 185.17]]}, {\"category\": \"car\", \"corners_3d\": [[79.53, 232.77], [27.28, 232.55], [130.87, 222.54], [174.65, 222.69], [79.53, 172.16], [27.28, 172.16], [130.87, 172.28], [174.65, 172.28]]}, {\"category\": \"car\", \"corners_3d\": [[198.86, 223.03], [161.27, 222.82], [221.15, 216.9], [254.5, 217.07], [198.86, 182.56], [161.27, 182.52], [221.15, 181.37], [254.5, 181.4]]}]\n```" - } - ] - }, - { - "id": 462, - "data_source": "Mono3DRefer", - "scene_name": "000257", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[792.91, 199.79], [808.72, 199.33], [857.53, 201.25], [841.47, 201.79], [792.91, 174.71], [808.72, 174.68], [857.53, 174.81], [841.47, 174.85]]}]\n```" - } - ] - }, - { - "id": 463, - "data_source": "Mono3DRefer", - "scene_name": "000257", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[696.36, 198.84], [725.25, 198.81], [740.2, 201.65], [708.17, 201.68], [696.36, 165.79], [725.25, 165.8], [740.2, 165.03], [708.17, 165.02]]}]\n```" - } - ] - }, - { - "id": 464, - "data_source": "Mono3DRefer", - "scene_name": "000257", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[792.91, 199.79], [808.72, 199.33], [857.53, 201.25], [841.47, 201.79], [792.91, 174.71], [808.72, 174.68], [857.53, 174.81], [841.47, 174.85]]}, {\"category\": \"van\", \"corners_3d\": [[696.36, 198.84], [725.25, 198.81], [740.2, 201.65], [708.17, 201.68], [696.36, 165.79], [725.25, 165.8], [740.2, 165.03], [708.17, 165.02]]}]\n```" - } - ] - }, - { - "id": 465, - "data_source": "Mono3DRefer", - "scene_name": "000258", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[682.65, 206.44], [707.05, 206.65], [703.67, 209.38], [677.33, 209.13], [682.65, 183.55], [707.05, 183.62], [703.67, 184.49], [677.33, 184.41]]}, {\"category\": \"car\", \"corners_3d\": [[345.17, 268.33], [279.95, 266.85], [361.14, 249.72], [415.35, 250.7], [345.17, 198.39], [279.95, 198.0], [361.14, 193.41], [415.35, 193.68]]}, {\"category\": \"car\", \"corners_3d\": [[428.16, 250.4], [375.9, 249.53], [430.49, 237.29], [474.85, 237.91], [428.16, 182.97], [375.9, 182.86], [430.49, 181.26], [474.85, 181.34]]}, {\"category\": \"car\", \"corners_3d\": [[705.98, 221.53], [749.02, 221.73], [762.91, 229.42], [713.06, 229.15], [705.98, 181.26], [749.02, 181.29], [762.91, 182.62], [713.06, 182.57]]}, {\"category\": \"car\", \"corners_3d\": [[716.96, 215.79], [750.26, 216.2], [748.5, 221.79], [710.96, 221.26], [716.96, 171.1], [750.26, 171.09], [748.5, 170.86], [710.96, 170.88]]}]\n```" - } - ] - }, - { - "id": 466, - "data_source": "Mono3DRefer", - "scene_name": "000259", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[437.76, 257.87], [522.31, 258.24], [476.04, 291.3], [355.64, 290.54], [437.76, 179.29], [522.31, 179.28], [476.04, 178.32], [355.64, 178.35]]}, {\"category\": \"car\", \"corners_3d\": [[507.41, 219.12], [555.21, 219.19], [543.73, 225.62], [487.81, 225.52], [507.41, 175.24], [555.21, 175.23], [543.73, 174.16], [487.81, 174.18]]}, {\"category\": \"car\", \"corners_3d\": [[601.08, 203.56], [639.14, 203.59], [642.98, 207.15], [598.79, 207.12], [601.08, 168.96], [639.14, 168.94], [642.98, 166.91], [598.79, 166.93]]}, {\"category\": \"car\", \"corners_3d\": [[564.11, 195.13], [593.42, 195.14], [591.16, 196.55], [558.82, 196.54], [564.11, 166.36], [593.42, 166.35], [591.16, 164.78], [558.82, 164.8]]}]\n```" - } - ] - }, - { - "id": 467, - "data_source": "Mono3DRefer", - "scene_name": "000262", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[757.42, 288.24], [866.87, 287.9], [1038.35, 360.65], [860.17, 361.55], [757.42, 180.97], [866.87, 180.94], [1038.35, 186.06], [860.17, 186.12]]}, {\"category\": \"car\", \"corners_3d\": [[692.56, 246.97], [761.69, 246.76], [822.38, 272.66], [729.17, 273.04], [692.56, 178.75], [761.69, 178.73], [822.38, 180.79], [729.17, 180.82]]}, {\"category\": \"car\", \"corners_3d\": [[482.16, 229.73], [534.01, 229.73], [518.39, 241.41], [455.89, 241.4], [482.16, 180.07], [534.01, 180.07], [518.39, 181.55], [455.89, 181.55]]}, {\"category\": \"car\", \"corners_3d\": [[674.42, 222.16], [721.7, 222.19], [743.16, 232.34], [686.14, 232.29], [674.42, 177.95], [721.7, 177.96], [743.16, 179.01], [686.14, 179.0]]}, {\"category\": \"car\", \"corners_3d\": [[651.41, 211.08], [687.4, 211.1], [696.19, 215.89], [655.68, 215.87], [651.41, 178.08], [687.4, 178.08], [696.19, 178.74], [655.68, 178.73]]}, {\"category\": \"car\", \"corners_3d\": [[636.25, 201.9], [664.15, 201.91], [669.45, 205.2], [638.39, 205.19], [636.25, 177.52], [664.15, 177.52], [669.45, 178.05], [638.39, 178.05]]}]\n```" - } - ] - }, - { - "id": 468, - "data_source": "Mono3DRefer", - "scene_name": "000264", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[740.07, 213.19], [731.19, 212.14], [746.06, 211.89], [755.27, 212.93], [740.07, 167.15], [731.19, 167.3], [746.06, 167.34], [755.27, 167.19]]}]\n```" - } - ] - }, - { - "id": 469, - "data_source": "Mono3DRefer", - "scene_name": "000264", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[819.28, 340.04], [966.79, 339.72], [1283.55, 484.99], [1008.29, 486.09], [819.28, 183.83], [966.79, 183.81], [1283.55, 193.35], [1008.29, 193.42]]}, {\"category\": \"car\", \"corners_3d\": [[654.41, 218.25], [698.28, 218.19], [715.17, 225.66], [664.09, 225.73], [654.41, 179.09], [698.28, 179.08], [715.17, 180.1], [664.09, 180.11]]}, {\"category\": \"car\", \"corners_3d\": [[127.0, 232.33], [71.6, 232.28], [158.69, 222.78], [205.28, 222.81], [127.0, 186.96], [71.6, 186.95], [158.69, 184.69], [205.28, 184.7]]}, {\"category\": \"car\", \"corners_3d\": [[586.8, 205.86], [619.23, 205.85], [620.99, 209.4], [585.07, 209.42], [586.8, 176.34], [619.23, 176.34], [620.99, 176.72], [585.07, 176.72]]}, {\"category\": \"car\", \"corners_3d\": [[320.39, 206.6], [290.82, 206.57], [320.91, 203.52], [347.82, 203.55], [320.39, 178.34], [290.82, 178.34], [320.91, 177.84], [347.82, 177.84]]}, {\"category\": \"car\", \"corners_3d\": [[574.8, 220.59], [624.01, 220.59], [626.2, 228.13], [569.22, 228.13], [574.8, 181.0], [624.01, 181.0], [626.2, 182.28], [569.22, 182.28]]}, {\"category\": \"car\", \"corners_3d\": [[215.53, 363.88], [399.82, 361.54], [167.97, 612.68], [-275.65, 625.61], [215.53, 196.84], [399.82, 196.55], [167.97, 228.08], [-275.65, 229.7]]}, {\"category\": \"car\", \"corners_3d\": [[373.89, 284.15], [478.65, 284.17], [421.2, 332.8], [270.69, 332.78], [373.89, 192.48], [478.65, 192.48], [421.2, 201.05], [270.69, 201.05]]}, {\"category\": \"car\", \"corners_3d\": [[327.85, 287.54], [221.25, 301.48], [134.36, 278.48], [231.51, 268.9], [327.85, 196.17], [221.25, 199.01], [134.36, 194.33], [231.51, 192.38]]}, {\"category\": \"car\", \"corners_3d\": [[376.91, 259.48], [296.0, 267.57], [218.04, 253.07], [293.2, 247.19], [376.91, 188.13], [296.0, 189.55], [218.04, 187.0], [293.2, 185.96]]}, {\"category\": \"car\", \"corners_3d\": [[455.48, 230.46], [411.74, 233.77], [341.36, 227.26], [384.08, 224.61], [455.48, 183.29], [411.74, 183.89], [341.36, 182.71], [384.08, 182.23]]}, {\"category\": \"car\", \"corners_3d\": [[477.19, 223.91], [440.37, 226.86], [372.0, 222.58], [409.51, 220.07], [477.19, 177.83], [440.37, 178.12], [372.0, 177.7], [409.51, 177.46]]}, {\"category\": \"car\", \"corners_3d\": [[497.73, 212.32], [467.67, 213.68], [422.15, 210.27], [451.18, 209.12], [497.73, 181.19], [467.67, 181.48], [422.15, 180.76], [451.18, 180.52]]}, {\"category\": \"car\", \"corners_3d\": [[534.13, 208.22], [566.89, 208.27], [559.42, 212.35], [522.92, 212.28], [534.13, 180.68], [566.89, 180.7], [559.42, 181.6], [522.92, 181.58]]}, {\"category\": \"car\", \"corners_3d\": [[529.56, 202.16], [510.59, 203.15], [458.05, 201.53], [477.66, 200.64], [529.56, 175.56], [510.59, 175.65], [458.05, 175.5], [477.66, 175.42]]}]\n```" - } - ] - }, - { - "id": 470, - "data_source": "Mono3DRefer", - "scene_name": "000264", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[640.09, 202.66], [688.9, 202.66], [709.47, 210.45], [647.92, 210.44], [640.09, 137.91], [688.9, 137.91], [709.47, 128.78], [647.92, 128.78]]}]\n```" - } - ] - }, - { - "id": 471, - "data_source": "Mono3DRefer", - "scene_name": "000264", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[819.28, 340.04], [966.79, 339.72], [1283.55, 484.99], [1008.29, 486.09], [819.28, 183.83], [966.79, 183.81], [1283.55, 193.35], [1008.29, 193.42]]}, {\"category\": \"car\", \"corners_3d\": [[654.41, 218.25], [698.28, 218.19], [715.17, 225.66], [664.09, 225.73], [654.41, 179.09], [698.28, 179.08], [715.17, 180.1], [664.09, 180.11]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[740.07, 213.19], [731.19, 212.14], [746.06, 211.89], [755.27, 212.93], [740.07, 167.15], [731.19, 167.3], [746.06, 167.34], [755.27, 167.19]]}, {\"category\": \"car\", \"corners_3d\": [[127.0, 232.33], [71.6, 232.28], [158.69, 222.78], [205.28, 222.81], [127.0, 186.96], [71.6, 186.95], [158.69, 184.69], [205.28, 184.7]]}, {\"category\": \"car\", \"corners_3d\": [[586.8, 205.86], [619.23, 205.85], [620.99, 209.4], [585.07, 209.42], [586.8, 176.34], [619.23, 176.34], [620.99, 176.72], [585.07, 176.72]]}, {\"category\": \"car\", \"corners_3d\": [[320.39, 206.6], [290.82, 206.57], [320.91, 203.52], [347.82, 203.55], [320.39, 178.34], [290.82, 178.34], [320.91, 177.84], [347.82, 177.84]]}, {\"category\": \"car\", \"corners_3d\": [[574.8, 220.59], [624.01, 220.59], [626.2, 228.13], [569.22, 228.13], [574.8, 181.0], [624.01, 181.0], [626.2, 182.28], [569.22, 182.28]]}, {\"category\": \"car\", \"corners_3d\": [[215.53, 363.88], [399.82, 361.54], [167.97, 612.68], [-275.65, 625.61], [215.53, 196.84], [399.82, 196.55], [167.97, 228.08], [-275.65, 229.7]]}, {\"category\": \"car\", \"corners_3d\": [[373.89, 284.15], [478.65, 284.17], [421.2, 332.8], [270.69, 332.78], [373.89, 192.48], [478.65, 192.48], [421.2, 201.05], [270.69, 201.05]]}, {\"category\": \"car\", \"corners_3d\": [[327.85, 287.54], [221.25, 301.48], [134.36, 278.48], [231.51, 268.9], [327.85, 196.17], [221.25, 199.01], [134.36, 194.33], [231.51, 192.38]]}, {\"category\": \"car\", \"corners_3d\": [[376.91, 259.48], [296.0, 267.57], [218.04, 253.07], [293.2, 247.19], [376.91, 188.13], [296.0, 189.55], [218.04, 187.0], [293.2, 185.96]]}, {\"category\": \"car\", \"corners_3d\": [[455.48, 230.46], [411.74, 233.77], [341.36, 227.26], [384.08, 224.61], [455.48, 183.29], [411.74, 183.89], [341.36, 182.71], [384.08, 182.23]]}, {\"category\": \"car\", \"corners_3d\": [[477.19, 223.91], [440.37, 226.86], [372.0, 222.58], [409.51, 220.07], [477.19, 177.83], [440.37, 178.12], [372.0, 177.7], [409.51, 177.46]]}, {\"category\": \"car\", \"corners_3d\": [[497.73, 212.32], [467.67, 213.68], [422.15, 210.27], [451.18, 209.12], [497.73, 181.19], [467.67, 181.48], [422.15, 180.76], [451.18, 180.52]]}, {\"category\": \"car\", \"corners_3d\": [[534.13, 208.22], [566.89, 208.27], [559.42, 212.35], [522.92, 212.28], [534.13, 180.68], [566.89, 180.7], [559.42, 181.6], [522.92, 181.58]]}, {\"category\": \"car\", \"corners_3d\": [[529.56, 202.16], [510.59, 203.15], [458.05, 201.53], [477.66, 200.64], [529.56, 175.56], [510.59, 175.65], [458.05, 175.5], [477.66, 175.42]]}]\n```" - } - ] - }, - { - "id": 472, - "data_source": "Mono3DRefer", - "scene_name": "000264", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[640.09, 202.66], [688.9, 202.66], [709.47, 210.45], [647.92, 210.44], [640.09, 137.91], [688.9, 137.91], [709.47, 128.78], [647.92, 128.78]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[740.07, 213.19], [731.19, 212.14], [746.06, 211.89], [755.27, 212.93], [740.07, 167.15], [731.19, 167.3], [746.06, 167.34], [755.27, 167.19]]}]\n```" - } - ] - }, - { - "id": 473, - "data_source": "Mono3DRefer", - "scene_name": "000264", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[819.28, 340.04], [966.79, 339.72], [1283.55, 484.99], [1008.29, 486.09], [819.28, 183.83], [966.79, 183.81], [1283.55, 193.35], [1008.29, 193.42]]}, {\"category\": \"car\", \"corners_3d\": [[654.41, 218.25], [698.28, 218.19], [715.17, 225.66], [664.09, 225.73], [654.41, 179.09], [698.28, 179.08], [715.17, 180.1], [664.09, 180.11]]}, {\"category\": \"truck\", \"corners_3d\": [[640.09, 202.66], [688.9, 202.66], [709.47, 210.45], [647.92, 210.44], [640.09, 137.91], [688.9, 137.91], [709.47, 128.78], [647.92, 128.78]]}, {\"category\": \"car\", \"corners_3d\": [[127.0, 232.33], [71.6, 232.28], [158.69, 222.78], [205.28, 222.81], [127.0, 186.96], [71.6, 186.95], [158.69, 184.69], [205.28, 184.7]]}, {\"category\": \"car\", \"corners_3d\": [[586.8, 205.86], [619.23, 205.85], [620.99, 209.4], [585.07, 209.42], [586.8, 176.34], [619.23, 176.34], [620.99, 176.72], [585.07, 176.72]]}, {\"category\": \"car\", \"corners_3d\": [[320.39, 206.6], [290.82, 206.57], [320.91, 203.52], [347.82, 203.55], [320.39, 178.34], [290.82, 178.34], [320.91, 177.84], [347.82, 177.84]]}, {\"category\": \"car\", \"corners_3d\": [[574.8, 220.59], [624.01, 220.59], [626.2, 228.13], [569.22, 228.13], [574.8, 181.0], [624.01, 181.0], [626.2, 182.28], [569.22, 182.28]]}, {\"category\": \"car\", \"corners_3d\": [[215.53, 363.88], [399.82, 361.54], [167.97, 612.68], [-275.65, 625.61], [215.53, 196.84], [399.82, 196.55], [167.97, 228.08], [-275.65, 229.7]]}, {\"category\": \"car\", \"corners_3d\": [[373.89, 284.15], [478.65, 284.17], [421.2, 332.8], [270.69, 332.78], [373.89, 192.48], [478.65, 192.48], [421.2, 201.05], [270.69, 201.05]]}, {\"category\": \"car\", \"corners_3d\": [[327.85, 287.54], [221.25, 301.48], [134.36, 278.48], [231.51, 268.9], [327.85, 196.17], [221.25, 199.01], [134.36, 194.33], [231.51, 192.38]]}, {\"category\": \"car\", \"corners_3d\": [[376.91, 259.48], [296.0, 267.57], [218.04, 253.07], [293.2, 247.19], [376.91, 188.13], [296.0, 189.55], [218.04, 187.0], [293.2, 185.96]]}, {\"category\": \"car\", \"corners_3d\": [[455.48, 230.46], [411.74, 233.77], [341.36, 227.26], [384.08, 224.61], [455.48, 183.29], [411.74, 183.89], [341.36, 182.71], [384.08, 182.23]]}, {\"category\": \"car\", \"corners_3d\": [[477.19, 223.91], [440.37, 226.86], [372.0, 222.58], [409.51, 220.07], [477.19, 177.83], [440.37, 178.12], [372.0, 177.7], [409.51, 177.46]]}, {\"category\": \"car\", \"corners_3d\": [[497.73, 212.32], [467.67, 213.68], [422.15, 210.27], [451.18, 209.12], [497.73, 181.19], [467.67, 181.48], [422.15, 180.76], [451.18, 180.52]]}, {\"category\": \"car\", \"corners_3d\": [[534.13, 208.22], [566.89, 208.27], [559.42, 212.35], [522.92, 212.28], [534.13, 180.68], [566.89, 180.7], [559.42, 181.6], [522.92, 181.58]]}, {\"category\": \"car\", \"corners_3d\": [[529.56, 202.16], [510.59, 203.15], [458.05, 201.53], [477.66, 200.64], [529.56, 175.56], [510.59, 175.65], [458.05, 175.5], [477.66, 175.42]]}]\n```" - } - ] - }, - { - "id": 474, - "data_source": "Mono3DRefer", - "scene_name": "000264", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[819.28, 340.04], [966.79, 339.72], [1283.55, 484.99], [1008.29, 486.09], [819.28, 183.83], [966.79, 183.81], [1283.55, 193.35], [1008.29, 193.42]]}, {\"category\": \"car\", \"corners_3d\": [[654.41, 218.25], [698.28, 218.19], [715.17, 225.66], [664.09, 225.73], [654.41, 179.09], [698.28, 179.08], [715.17, 180.1], [664.09, 180.11]]}, {\"category\": \"truck\", \"corners_3d\": [[640.09, 202.66], [688.9, 202.66], [709.47, 210.45], [647.92, 210.44], [640.09, 137.91], [688.9, 137.91], [709.47, 128.78], [647.92, 128.78]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[740.07, 213.19], [731.19, 212.14], [746.06, 211.89], [755.27, 212.93], [740.07, 167.15], [731.19, 167.3], [746.06, 167.34], [755.27, 167.19]]}, {\"category\": \"car\", \"corners_3d\": [[127.0, 232.33], [71.6, 232.28], [158.69, 222.78], [205.28, 222.81], [127.0, 186.96], [71.6, 186.95], [158.69, 184.69], [205.28, 184.7]]}, {\"category\": \"car\", \"corners_3d\": [[586.8, 205.86], [619.23, 205.85], [620.99, 209.4], [585.07, 209.42], [586.8, 176.34], [619.23, 176.34], [620.99, 176.72], [585.07, 176.72]]}, {\"category\": \"car\", \"corners_3d\": [[320.39, 206.6], [290.82, 206.57], [320.91, 203.52], [347.82, 203.55], [320.39, 178.34], [290.82, 178.34], [320.91, 177.84], [347.82, 177.84]]}, {\"category\": \"car\", \"corners_3d\": [[574.8, 220.59], [624.01, 220.59], [626.2, 228.13], [569.22, 228.13], [574.8, 181.0], [624.01, 181.0], [626.2, 182.28], [569.22, 182.28]]}, {\"category\": \"car\", \"corners_3d\": [[215.53, 363.88], [399.82, 361.54], [167.97, 612.68], [-275.65, 625.61], [215.53, 196.84], [399.82, 196.55], [167.97, 228.08], [-275.65, 229.7]]}, {\"category\": \"car\", \"corners_3d\": [[373.89, 284.15], [478.65, 284.17], [421.2, 332.8], [270.69, 332.78], [373.89, 192.48], [478.65, 192.48], [421.2, 201.05], [270.69, 201.05]]}, {\"category\": \"car\", \"corners_3d\": [[327.85, 287.54], [221.25, 301.48], [134.36, 278.48], [231.51, 268.9], [327.85, 196.17], [221.25, 199.01], [134.36, 194.33], [231.51, 192.38]]}, {\"category\": \"car\", \"corners_3d\": [[376.91, 259.48], [296.0, 267.57], [218.04, 253.07], [293.2, 247.19], [376.91, 188.13], [296.0, 189.55], [218.04, 187.0], [293.2, 185.96]]}, {\"category\": \"car\", \"corners_3d\": [[455.48, 230.46], [411.74, 233.77], [341.36, 227.26], [384.08, 224.61], [455.48, 183.29], [411.74, 183.89], [341.36, 182.71], [384.08, 182.23]]}, {\"category\": \"car\", \"corners_3d\": [[477.19, 223.91], [440.37, 226.86], [372.0, 222.58], [409.51, 220.07], [477.19, 177.83], [440.37, 178.12], [372.0, 177.7], [409.51, 177.46]]}, {\"category\": \"car\", \"corners_3d\": [[497.73, 212.32], [467.67, 213.68], [422.15, 210.27], [451.18, 209.12], [497.73, 181.19], [467.67, 181.48], [422.15, 180.76], [451.18, 180.52]]}, {\"category\": \"car\", \"corners_3d\": [[534.13, 208.22], [566.89, 208.27], [559.42, 212.35], [522.92, 212.28], [534.13, 180.68], [566.89, 180.7], [559.42, 181.6], [522.92, 181.58]]}, {\"category\": \"car\", \"corners_3d\": [[529.56, 202.16], [510.59, 203.15], [458.05, 201.53], [477.66, 200.64], [529.56, 175.56], [510.59, 175.65], [458.05, 175.5], [477.66, 175.42]]}]\n```" - } - ] - }, - { - "id": 475, - "data_source": "Mono3DRefer", - "scene_name": "000265", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[904.41, 275.5], [961.94, 276.53], [1022.2, 300.79], [950.64, 299.24], [904.41, 140.96], [961.94, 140.65], [1022.2, 133.11], [950.64, 133.59]]}, {\"category\": \"cyclist\", \"corners_3d\": [[761.27, 217.07], [776.75, 217.15], [784.5, 220.53], [767.82, 220.43], [761.27, 164.35], [776.75, 164.33], [784.5, 163.68], [767.82, 163.7]]}, {\"category\": \"cyclist\", \"corners_3d\": [[720.81, 200.31], [728.72, 200.33], [731.74, 201.43], [723.51, 201.41], [720.81, 171.06], [728.72, 171.06], [731.74, 170.99], [723.51, 170.99]]}]\n```" - } - ] - }, - { - "id": 476, - "data_source": "Mono3DRefer", - "scene_name": "000265", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[780.92, 210.34], [797.79, 210.34], [802.4, 211.22], [785.13, 211.23], [780.92, 168.68], [797.79, 168.68], [802.4, 168.58], [785.13, 168.58]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[800.68, 209.48], [812.37, 209.47], [818.72, 210.58], [806.68, 210.59], [800.68, 167.96], [812.37, 167.96], [818.72, 167.81], [806.68, 167.81]]}]\n```" - } - ] - }, - { - "id": 477, - "data_source": "Mono3DRefer", - "scene_name": "000265", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[904.41, 275.5], [961.94, 276.53], [1022.2, 300.79], [950.64, 299.24], [904.41, 140.96], [961.94, 140.65], [1022.2, 133.11], [950.64, 133.59]]}, {\"category\": \"cyclist\", \"corners_3d\": [[761.27, 217.07], [776.75, 217.15], [784.5, 220.53], [767.82, 220.43], [761.27, 164.35], [776.75, 164.33], [784.5, 163.68], [767.82, 163.7]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[780.92, 210.34], [797.79, 210.34], [802.4, 211.22], [785.13, 211.23], [780.92, 168.68], [797.79, 168.68], [802.4, 168.58], [785.13, 168.58]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[800.68, 209.48], [812.37, 209.47], [818.72, 210.58], [806.68, 210.59], [800.68, 167.96], [812.37, 167.96], [818.72, 167.81], [806.68, 167.81]]}, {\"category\": \"cyclist\", \"corners_3d\": [[720.81, 200.31], [728.72, 200.33], [731.74, 201.43], [723.51, 201.41], [720.81, 171.06], [728.72, 171.06], [731.74, 170.99], [723.51, 170.99]]}]\n```" - } - ] - }, - { - "id": 478, - "data_source": "Mono3DRefer", - "scene_name": "000266", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[250.64, 272.65], [351.46, 270.82], [290.94, 308.59], [148.66, 312.12], [250.64, 195.78], [351.46, 195.36], [290.94, 204.04], [148.66, 204.85]]}, {\"category\": \"car\", \"corners_3d\": [[647.71, 240.34], [714.49, 239.33], [777.23, 258.22], [692.34, 259.9], [647.71, 166.74], [714.49, 166.83], [777.23, 165.12], [692.34, 164.97]]}, {\"category\": \"car\", \"corners_3d\": [[357.61, 263.3], [263.98, 264.56], [327.78, 242.51], [398.46, 241.78], [357.61, 187.74], [263.98, 187.95], [327.78, 184.32], [398.46, 184.2]]}, {\"category\": \"car\", \"corners_3d\": [[581.55, 213.39], [624.34, 213.05], [640.66, 218.71], [591.95, 219.15], [581.55, 175.63], [624.34, 175.61], [640.66, 176.0], [591.95, 176.03]]}, {\"category\": \"car\", \"corners_3d\": [[412.4, 213.51], [452.22, 213.14], [448.4, 217.66], [404.03, 218.11], [412.4, 183.6], [452.22, 183.51], [448.4, 184.7], [404.03, 184.82]]}, {\"category\": \"car\", \"corners_3d\": [[557.99, 203.64], [586.9, 203.45], [594.99, 206.34], [563.39, 206.56], [557.99, 176.04], [586.9, 176.02], [594.99, 176.32], [563.39, 176.34]]}, {\"category\": \"car\", \"corners_3d\": [[748.87, 290.62], [868.67, 289.25], [1034.46, 351.68], [852.27, 354.94], [748.87, 179.71], [868.67, 179.63], [1034.46, 183.27], [852.27, 183.46]]}]\n```" - } - ] - }, - { - "id": 479, - "data_source": "Mono3DRefer", - "scene_name": "000266", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[402.99, 295.63], [355.23, 296.12], [389.88, 275.82], [429.69, 275.48], [402.99, 175.67], [355.23, 175.68], [389.88, 175.21], [429.69, 175.21]]}, {\"category\": \"cyclist\", \"corners_3d\": [[301.82, 331.83], [248.41, 332.3], [311.22, 302.11], [354.4, 301.8], [301.82, 170.06], [248.41, 170.06], [311.22, 170.59], [354.4, 170.59]]}]\n```" - } - ] - }, - { - "id": 480, - "data_source": "Mono3DRefer", - "scene_name": "000266", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[418.95, 203.27], [406.25, 203.38], [406.97, 202.85], [419.44, 202.74], [418.95, 170.87], [406.25, 170.86], [406.97, 170.89], [419.44, 170.9]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[407.17, 207.92], [395.01, 208.01], [396.61, 207.43], [408.56, 207.35], [407.17, 175.38], [395.01, 175.38], [396.61, 175.34], [408.56, 175.34]]}]\n```" - } - ] - }, - { - "id": 481, - "data_source": "Mono3DRefer", - "scene_name": "000266", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[250.64, 272.65], [351.46, 270.82], [290.94, 308.59], [148.66, 312.12], [250.64, 195.78], [351.46, 195.36], [290.94, 204.04], [148.66, 204.85]]}, {\"category\": \"car\", \"corners_3d\": [[647.71, 240.34], [714.49, 239.33], [777.23, 258.22], [692.34, 259.9], [647.71, 166.74], [714.49, 166.83], [777.23, 165.12], [692.34, 164.97]]}, {\"category\": \"car\", \"corners_3d\": [[357.61, 263.3], [263.98, 264.56], [327.78, 242.51], [398.46, 241.78], [357.61, 187.74], [263.98, 187.95], [327.78, 184.32], [398.46, 184.2]]}, {\"category\": \"car\", \"corners_3d\": [[581.55, 213.39], [624.34, 213.05], [640.66, 218.71], [591.95, 219.15], [581.55, 175.63], [624.34, 175.61], [640.66, 176.0], [591.95, 176.03]]}, {\"category\": \"car\", \"corners_3d\": [[412.4, 213.51], [452.22, 213.14], [448.4, 217.66], [404.03, 218.11], [412.4, 183.6], [452.22, 183.51], [448.4, 184.7], [404.03, 184.82]]}, {\"category\": \"car\", \"corners_3d\": [[557.99, 203.64], [586.9, 203.45], [594.99, 206.34], [563.39, 206.56], [557.99, 176.04], [586.9, 176.02], [594.99, 176.32], [563.39, 176.34]]}, {\"category\": \"cyclist\", \"corners_3d\": [[402.99, 295.63], [355.23, 296.12], [389.88, 275.82], [429.69, 275.48], [402.99, 175.67], [355.23, 175.68], [389.88, 175.21], [429.69, 175.21]]}, {\"category\": \"car\", \"corners_3d\": [[748.87, 290.62], [868.67, 289.25], [1034.46, 351.68], [852.27, 354.94], [748.87, 179.71], [868.67, 179.63], [1034.46, 183.27], [852.27, 183.46]]}, {\"category\": \"cyclist\", \"corners_3d\": [[301.82, 331.83], [248.41, 332.3], [311.22, 302.11], [354.4, 301.8], [301.82, 170.06], [248.41, 170.06], [311.22, 170.59], [354.4, 170.59]]}]\n```" - } - ] - }, - { - "id": 482, - "data_source": "Mono3DRefer", - "scene_name": "000266", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[250.64, 272.65], [351.46, 270.82], [290.94, 308.59], [148.66, 312.12], [250.64, 195.78], [351.46, 195.36], [290.94, 204.04], [148.66, 204.85]]}, {\"category\": \"car\", \"corners_3d\": [[647.71, 240.34], [714.49, 239.33], [777.23, 258.22], [692.34, 259.9], [647.71, 166.74], [714.49, 166.83], [777.23, 165.12], [692.34, 164.97]]}, {\"category\": \"car\", \"corners_3d\": [[357.61, 263.3], [263.98, 264.56], [327.78, 242.51], [398.46, 241.78], [357.61, 187.74], [263.98, 187.95], [327.78, 184.32], [398.46, 184.2]]}, {\"category\": \"car\", \"corners_3d\": [[581.55, 213.39], [624.34, 213.05], [640.66, 218.71], [591.95, 219.15], [581.55, 175.63], [624.34, 175.61], [640.66, 176.0], [591.95, 176.03]]}, {\"category\": \"car\", \"corners_3d\": [[412.4, 213.51], [452.22, 213.14], [448.4, 217.66], [404.03, 218.11], [412.4, 183.6], [452.22, 183.51], [448.4, 184.7], [404.03, 184.82]]}, {\"category\": \"car\", \"corners_3d\": [[557.99, 203.64], [586.9, 203.45], [594.99, 206.34], [563.39, 206.56], [557.99, 176.04], [586.9, 176.02], [594.99, 176.32], [563.39, 176.34]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[418.95, 203.27], [406.25, 203.38], [406.97, 202.85], [419.44, 202.74], [418.95, 170.87], [406.25, 170.86], [406.97, 170.89], [419.44, 170.9]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[407.17, 207.92], [395.01, 208.01], [396.61, 207.43], [408.56, 207.35], [407.17, 175.38], [395.01, 175.38], [396.61, 175.34], [408.56, 175.34]]}, {\"category\": \"car\", \"corners_3d\": [[748.87, 290.62], [868.67, 289.25], [1034.46, 351.68], [852.27, 354.94], [748.87, 179.71], [868.67, 179.63], [1034.46, 183.27], [852.27, 183.46]]}]\n```" - } - ] - }, - { - "id": 483, - "data_source": "Mono3DRefer", - "scene_name": "000266", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[418.95, 203.27], [406.25, 203.38], [406.97, 202.85], [419.44, 202.74], [418.95, 170.87], [406.25, 170.86], [406.97, 170.89], [419.44, 170.9]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[407.17, 207.92], [395.01, 208.01], [396.61, 207.43], [408.56, 207.35], [407.17, 175.38], [395.01, 175.38], [396.61, 175.34], [408.56, 175.34]]}, {\"category\": \"cyclist\", \"corners_3d\": [[402.99, 295.63], [355.23, 296.12], [389.88, 275.82], [429.69, 275.48], [402.99, 175.67], [355.23, 175.68], [389.88, 175.21], [429.69, 175.21]]}, {\"category\": \"cyclist\", \"corners_3d\": [[301.82, 331.83], [248.41, 332.3], [311.22, 302.11], [354.4, 301.8], [301.82, 170.06], [248.41, 170.06], [311.22, 170.59], [354.4, 170.59]]}]\n```" - } - ] - }, - { - "id": 484, - "data_source": "Mono3DRefer", - "scene_name": "000266", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[250.64, 272.65], [351.46, 270.82], [290.94, 308.59], [148.66, 312.12], [250.64, 195.78], [351.46, 195.36], [290.94, 204.04], [148.66, 204.85]]}, {\"category\": \"car\", \"corners_3d\": [[647.71, 240.34], [714.49, 239.33], [777.23, 258.22], [692.34, 259.9], [647.71, 166.74], [714.49, 166.83], [777.23, 165.12], [692.34, 164.97]]}, {\"category\": \"car\", \"corners_3d\": [[357.61, 263.3], [263.98, 264.56], [327.78, 242.51], [398.46, 241.78], [357.61, 187.74], [263.98, 187.95], [327.78, 184.32], [398.46, 184.2]]}, {\"category\": \"car\", \"corners_3d\": [[581.55, 213.39], [624.34, 213.05], [640.66, 218.71], [591.95, 219.15], [581.55, 175.63], [624.34, 175.61], [640.66, 176.0], [591.95, 176.03]]}, {\"category\": \"car\", \"corners_3d\": [[412.4, 213.51], [452.22, 213.14], [448.4, 217.66], [404.03, 218.11], [412.4, 183.6], [452.22, 183.51], [448.4, 184.7], [404.03, 184.82]]}, {\"category\": \"car\", \"corners_3d\": [[557.99, 203.64], [586.9, 203.45], [594.99, 206.34], [563.39, 206.56], [557.99, 176.04], [586.9, 176.02], [594.99, 176.32], [563.39, 176.34]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[418.95, 203.27], [406.25, 203.38], [406.97, 202.85], [419.44, 202.74], [418.95, 170.87], [406.25, 170.86], [406.97, 170.89], [419.44, 170.9]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[407.17, 207.92], [395.01, 208.01], [396.61, 207.43], [408.56, 207.35], [407.17, 175.38], [395.01, 175.38], [396.61, 175.34], [408.56, 175.34]]}, {\"category\": \"cyclist\", \"corners_3d\": [[402.99, 295.63], [355.23, 296.12], [389.88, 275.82], [429.69, 275.48], [402.99, 175.67], [355.23, 175.68], [389.88, 175.21], [429.69, 175.21]]}, {\"category\": \"car\", \"corners_3d\": [[748.87, 290.62], [868.67, 289.25], [1034.46, 351.68], [852.27, 354.94], [748.87, 179.71], [868.67, 179.63], [1034.46, 183.27], [852.27, 183.46]]}, {\"category\": \"cyclist\", \"corners_3d\": [[301.82, 331.83], [248.41, 332.3], [311.22, 302.11], [354.4, 301.8], [301.82, 170.06], [248.41, 170.06], [311.22, 170.59], [354.4, 170.59]]}]\n```" - } - ] - }, - { - "id": 485, - "data_source": "Mono3DRefer", - "scene_name": "000268", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[131.5, 221.1], [148.35, 221.06], [116.43, 224.61], [98.31, 224.66], [131.5, 172.3], [148.35, 172.3], [116.43, 172.26], [98.31, 172.25]]}]\n```" - } - ] - }, - { - "id": 486, - "data_source": "Mono3DRefer", - "scene_name": "000268", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[967.58, 201.94], [942.41, 203.22], [722.23, 195.74], [748.49, 195.0], [967.58, 125.89], [942.41, 123.81], [722.23, 135.89], [748.49, 137.08]]}]\n```" - } - ] - }, - { - "id": 487, - "data_source": "Mono3DRefer", - "scene_name": "000268", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[568.75, 284.32], [685.66, 284.33], [715.86, 328.91], [552.21, 328.89], [568.75, 172.15], [685.66, 172.15], [715.86, 171.87], [552.21, 171.87]]}, {\"category\": \"car\", \"corners_3d\": [[305.8, 229.18], [257.02, 229.29], [306.56, 220.83], [347.95, 220.74], [305.8, 182.58], [257.02, 182.6], [306.56, 181.14], [347.95, 181.12]]}]\n```" - } - ] - }, - { - "id": 488, - "data_source": "Mono3DRefer", - "scene_name": "000268", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[967.58, 201.94], [942.41, 203.22], [722.23, 195.74], [748.49, 195.0], [967.58, 125.89], [942.41, 123.81], [722.23, 135.89], [748.49, 137.08]]}, {\"category\": \"cyclist\", \"corners_3d\": [[131.5, 221.1], [148.35, 221.06], [116.43, 224.61], [98.31, 224.66], [131.5, 172.3], [148.35, 172.3], [116.43, 172.26], [98.31, 172.25]]}]\n```" - } - ] - }, - { - "id": 489, - "data_source": "Mono3DRefer", - "scene_name": "000268", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[568.75, 284.32], [685.66, 284.33], [715.86, 328.91], [552.21, 328.89], [568.75, 172.15], [685.66, 172.15], [715.86, 171.87], [552.21, 171.87]]}, {\"category\": \"car\", \"corners_3d\": [[305.8, 229.18], [257.02, 229.29], [306.56, 220.83], [347.95, 220.74], [305.8, 182.58], [257.02, 182.6], [306.56, 181.14], [347.95, 181.12]]}, {\"category\": \"cyclist\", \"corners_3d\": [[131.5, 221.1], [148.35, 221.06], [116.43, 224.61], [98.31, 224.66], [131.5, 172.3], [148.35, 172.3], [116.43, 172.26], [98.31, 172.25]]}]\n```" - } - ] - }, - { - "id": 490, - "data_source": "Mono3DRefer", - "scene_name": "000268", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[568.75, 284.32], [685.66, 284.33], [715.86, 328.91], [552.21, 328.89], [568.75, 172.15], [685.66, 172.15], [715.86, 171.87], [552.21, 171.87]]}, {\"category\": \"tram\", \"corners_3d\": [[967.58, 201.94], [942.41, 203.22], [722.23, 195.74], [748.49, 195.0], [967.58, 125.89], [942.41, 123.81], [722.23, 135.89], [748.49, 137.08]]}, {\"category\": \"car\", \"corners_3d\": [[305.8, 229.18], [257.02, 229.29], [306.56, 220.83], [347.95, 220.74], [305.8, 182.58], [257.02, 182.6], [306.56, 181.14], [347.95, 181.12]]}]\n```" - } - ] - }, - { - "id": 491, - "data_source": "Mono3DRefer", - "scene_name": "000268", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[568.75, 284.32], [685.66, 284.33], [715.86, 328.91], [552.21, 328.89], [568.75, 172.15], [685.66, 172.15], [715.86, 171.87], [552.21, 171.87]]}, {\"category\": \"tram\", \"corners_3d\": [[967.58, 201.94], [942.41, 203.22], [722.23, 195.74], [748.49, 195.0], [967.58, 125.89], [942.41, 123.81], [722.23, 135.89], [748.49, 137.08]]}, {\"category\": \"car\", \"corners_3d\": [[305.8, 229.18], [257.02, 229.29], [306.56, 220.83], [347.95, 220.74], [305.8, 182.58], [257.02, 182.6], [306.56, 181.14], [347.95, 181.12]]}, {\"category\": \"cyclist\", \"corners_3d\": [[131.5, 221.1], [148.35, 221.06], [116.43, 224.61], [98.31, 224.66], [131.5, 172.3], [148.35, 172.3], [116.43, 172.26], [98.31, 172.25]]}]\n```" - } - ] - }, - { - "id": 492, - "data_source": "Mono3DRefer", - "scene_name": "000271", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[913.97, 214.63], [904.1, 213.45], [964.76, 213.24], [976.33, 214.41], [913.97, 189.2], [904.1, 188.74], [964.76, 188.65], [976.33, 189.11]]}, {\"category\": \"car\", \"corners_3d\": [[564.83, 222.73], [611.02, 222.67], [614.3, 233.35], [558.2, 233.44], [564.83, 176.81], [611.02, 176.81], [614.3, 177.65], [558.2, 177.66]]}, {\"category\": \"car\", \"corners_3d\": [[259.92, 204.43], [238.93, 204.39], [273.18, 201.69], [292.4, 201.72], [259.92, 180.54], [238.93, 180.53], [273.18, 179.87], [292.4, 179.88]]}]\n```" - } - ] - }, - { - "id": 493, - "data_source": "Mono3DRefer", - "scene_name": "000271", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[835.12, 216.18], [842.29, 216.15], [855.91, 218.29], [848.39, 218.32], [835.12, 180.82], [842.29, 180.82], [855.91, 181.21], [848.39, 181.22]]}]\n```" - } - ] - }, - { - "id": 494, - "data_source": "Mono3DRefer", - "scene_name": "000271", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[495.45, 239.68], [411.64, 239.84], [459.83, 222.19], [521.52, 222.1], [495.45, 133.04], [411.64, 132.94], [459.83, 143.46], [521.52, 143.51]]}]\n```" - } - ] - }, - { - "id": 495, - "data_source": "Mono3DRefer", - "scene_name": "000271", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[913.97, 214.63], [904.1, 213.45], [964.76, 213.24], [976.33, 214.41], [913.97, 189.2], [904.1, 188.74], [964.76, 188.65], [976.33, 189.11]]}, {\"category\": \"car\", \"corners_3d\": [[564.83, 222.73], [611.02, 222.67], [614.3, 233.35], [558.2, 233.44], [564.83, 176.81], [611.02, 176.81], [614.3, 177.65], [558.2, 177.66]]}, {\"category\": \"car\", \"corners_3d\": [[259.92, 204.43], [238.93, 204.39], [273.18, 201.69], [292.4, 201.72], [259.92, 180.54], [238.93, 180.53], [273.18, 179.87], [292.4, 179.88]]}, {\"category\": \"cyclist\", \"corners_3d\": [[835.12, 216.18], [842.29, 216.15], [855.91, 218.29], [848.39, 218.32], [835.12, 180.82], [842.29, 180.82], [855.91, 181.21], [848.39, 181.22]]}]\n```" - } - ] - }, - { - "id": 496, - "data_source": "Mono3DRefer", - "scene_name": "000271", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[495.45, 239.68], [411.64, 239.84], [459.83, 222.19], [521.52, 222.1], [495.45, 133.04], [411.64, 132.94], [459.83, 143.46], [521.52, 143.51]]}, {\"category\": \"car\", \"corners_3d\": [[913.97, 214.63], [904.1, 213.45], [964.76, 213.24], [976.33, 214.41], [913.97, 189.2], [904.1, 188.74], [964.76, 188.65], [976.33, 189.11]]}, {\"category\": \"car\", \"corners_3d\": [[564.83, 222.73], [611.02, 222.67], [614.3, 233.35], [558.2, 233.44], [564.83, 176.81], [611.02, 176.81], [614.3, 177.65], [558.2, 177.66]]}, {\"category\": \"car\", \"corners_3d\": [[259.92, 204.43], [238.93, 204.39], [273.18, 201.69], [292.4, 201.72], [259.92, 180.54], [238.93, 180.53], [273.18, 179.87], [292.4, 179.88]]}]\n```" - } - ] - }, - { - "id": 497, - "data_source": "Mono3DRefer", - "scene_name": "000271", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[495.45, 239.68], [411.64, 239.84], [459.83, 222.19], [521.52, 222.1], [495.45, 133.04], [411.64, 132.94], [459.83, 143.46], [521.52, 143.51]]}, {\"category\": \"cyclist\", \"corners_3d\": [[835.12, 216.18], [842.29, 216.15], [855.91, 218.29], [848.39, 218.32], [835.12, 180.82], [842.29, 180.82], [855.91, 181.21], [848.39, 181.22]]}]\n```" - } - ] - }, - { - "id": 498, - "data_source": "Mono3DRefer", - "scene_name": "000271", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[495.45, 239.68], [411.64, 239.84], [459.83, 222.19], [521.52, 222.1], [495.45, 133.04], [411.64, 132.94], [459.83, 143.46], [521.52, 143.51]]}, {\"category\": \"car\", \"corners_3d\": [[913.97, 214.63], [904.1, 213.45], [964.76, 213.24], [976.33, 214.41], [913.97, 189.2], [904.1, 188.74], [964.76, 188.65], [976.33, 189.11]]}, {\"category\": \"car\", \"corners_3d\": [[564.83, 222.73], [611.02, 222.67], [614.3, 233.35], [558.2, 233.44], [564.83, 176.81], [611.02, 176.81], [614.3, 177.65], [558.2, 177.66]]}, {\"category\": \"car\", \"corners_3d\": [[259.92, 204.43], [238.93, 204.39], [273.18, 201.69], [292.4, 201.72], [259.92, 180.54], [238.93, 180.53], [273.18, 179.87], [292.4, 179.88]]}, {\"category\": \"cyclist\", \"corners_3d\": [[835.12, 216.18], [842.29, 216.15], [855.91, 218.29], [848.39, 218.32], [835.12, 180.82], [842.29, 180.82], [855.91, 181.21], [848.39, 181.22]]}]\n```" - } - ] - }, - { - "id": 499, - "data_source": "Mono3DRefer", - "scene_name": "000272", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[633.15, 204.0], [603.75, 203.77], [615.0, 201.22], [642.04, 201.42], [633.15, 177.6], [603.75, 177.56], [615.0, 177.18], [642.04, 177.21]]}]\n```" - } - ] - }, - { - "id": 500, - "data_source": "Mono3DRefer", - "scene_name": "000275", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[577.07, 247.41], [656.51, 247.5], [669.31, 271.9], [563.93, 271.74], [577.07, 172.36], [656.51, 172.36], [669.31, 172.19], [563.93, 172.19]]}, {\"category\": \"car\", \"corners_3d\": [[1117.95, 307.38], [1328.48, 331.64], [1282.04, 435.07], [976.51, 374.89], [1117.95, 173.74], [1328.48, 173.9], [1282.04, 174.59], [976.51, 174.19]]}, {\"category\": \"car\", \"corners_3d\": [[961.25, 262.26], [1088.7, 274.44], [945.53, 309.65], [802.96, 288.43], [961.25, 163.18], [1088.7, 161.86], [945.53, 158.05], [802.96, 160.35]]}, {\"category\": \"car\", \"corners_3d\": [[885.45, 242.09], [965.0, 247.98], [863.49, 265.52], [776.79, 256.72], [885.45, 174.31], [965.0, 174.43], [863.49, 174.8], [776.79, 174.62]]}, {\"category\": \"car\", \"corners_3d\": [[77.85, 263.13], [-36.25, 265.08], [105.46, 240.71], [187.65, 239.64], [77.85, 169.86], [-36.25, 169.8], [105.46, 170.61], [187.65, 170.64]]}]\n```" - } - ] - }, - { - "id": 501, - "data_source": "Mono3DRefer", - "scene_name": "000275", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[730.33, 212.67], [744.57, 213.43], [728.65, 214.36], [714.39, 213.58], [730.33, 166.35], [744.57, 166.22], [728.65, 166.07], [714.39, 166.2]]}]\n```" - } - ] - }, - { - "id": 502, - "data_source": "Mono3DRefer", - "scene_name": "000275", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[549.44, 190.91], [527.14, 190.9], [534.22, 189.47], [554.75, 189.47], [549.44, 168.25], [527.14, 168.25], [534.22, 168.62], [554.75, 168.62]]}]\n```" - } - ] - }, - { - "id": 503, - "data_source": "Mono3DRefer", - "scene_name": "000275", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[577.07, 247.41], [656.51, 247.5], [669.31, 271.9], [563.93, 271.74], [577.07, 172.36], [656.51, 172.36], [669.31, 172.19], [563.93, 172.19]]}, {\"category\": \"car\", \"corners_3d\": [[1117.95, 307.38], [1328.48, 331.64], [1282.04, 435.07], [976.51, 374.89], [1117.95, 173.74], [1328.48, 173.9], [1282.04, 174.59], [976.51, 174.19]]}, {\"category\": \"car\", \"corners_3d\": [[961.25, 262.26], [1088.7, 274.44], [945.53, 309.65], [802.96, 288.43], [961.25, 163.18], [1088.7, 161.86], [945.53, 158.05], [802.96, 160.35]]}, {\"category\": \"car\", \"corners_3d\": [[885.45, 242.09], [965.0, 247.98], [863.49, 265.52], [776.79, 256.72], [885.45, 174.31], [965.0, 174.43], [863.49, 174.8], [776.79, 174.62]]}, {\"category\": \"car\", \"corners_3d\": [[77.85, 263.13], [-36.25, 265.08], [105.46, 240.71], [187.65, 239.64], [77.85, 169.86], [-36.25, 169.8], [105.46, 170.61], [187.65, 170.64]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[730.33, 212.67], [744.57, 213.43], [728.65, 214.36], [714.39, 213.58], [730.33, 166.35], [744.57, 166.22], [728.65, 166.07], [714.39, 166.2]]}]\n```" - } - ] - }, - { - "id": 504, - "data_source": "Mono3DRefer", - "scene_name": "000275", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[577.07, 247.41], [656.51, 247.5], [669.31, 271.9], [563.93, 271.74], [577.07, 172.36], [656.51, 172.36], [669.31, 172.19], [563.93, 172.19]]}, {\"category\": \"car\", \"corners_3d\": [[1117.95, 307.38], [1328.48, 331.64], [1282.04, 435.07], [976.51, 374.89], [1117.95, 173.74], [1328.48, 173.9], [1282.04, 174.59], [976.51, 174.19]]}, {\"category\": \"car\", \"corners_3d\": [[961.25, 262.26], [1088.7, 274.44], [945.53, 309.65], [802.96, 288.43], [961.25, 163.18], [1088.7, 161.86], [945.53, 158.05], [802.96, 160.35]]}, {\"category\": \"car\", \"corners_3d\": [[885.45, 242.09], [965.0, 247.98], [863.49, 265.52], [776.79, 256.72], [885.45, 174.31], [965.0, 174.43], [863.49, 174.8], [776.79, 174.62]]}, {\"category\": \"car\", \"corners_3d\": [[77.85, 263.13], [-36.25, 265.08], [105.46, 240.71], [187.65, 239.64], [77.85, 169.86], [-36.25, 169.8], [105.46, 170.61], [187.65, 170.64]]}, {\"category\": \"van\", \"corners_3d\": [[549.44, 190.91], [527.14, 190.9], [534.22, 189.47], [554.75, 189.47], [549.44, 168.25], [527.14, 168.25], [534.22, 168.62], [554.75, 168.62]]}]\n```" - } - ] - }, - { - "id": 505, - "data_source": "Mono3DRefer", - "scene_name": "000275", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[549.44, 190.91], [527.14, 190.9], [534.22, 189.47], [554.75, 189.47], [549.44, 168.25], [527.14, 168.25], [534.22, 168.62], [554.75, 168.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[730.33, 212.67], [744.57, 213.43], [728.65, 214.36], [714.39, 213.58], [730.33, 166.35], [744.57, 166.22], [728.65, 166.07], [714.39, 166.2]]}]\n```" - } - ] - }, - { - "id": 506, - "data_source": "Mono3DRefer", - "scene_name": "000275", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[577.07, 247.41], [656.51, 247.5], [669.31, 271.9], [563.93, 271.74], [577.07, 172.36], [656.51, 172.36], [669.31, 172.19], [563.93, 172.19]]}, {\"category\": \"car\", \"corners_3d\": [[1117.95, 307.38], [1328.48, 331.64], [1282.04, 435.07], [976.51, 374.89], [1117.95, 173.74], [1328.48, 173.9], [1282.04, 174.59], [976.51, 174.19]]}, {\"category\": \"car\", \"corners_3d\": [[961.25, 262.26], [1088.7, 274.44], [945.53, 309.65], [802.96, 288.43], [961.25, 163.18], [1088.7, 161.86], [945.53, 158.05], [802.96, 160.35]]}, {\"category\": \"car\", \"corners_3d\": [[885.45, 242.09], [965.0, 247.98], [863.49, 265.52], [776.79, 256.72], [885.45, 174.31], [965.0, 174.43], [863.49, 174.8], [776.79, 174.62]]}, {\"category\": \"car\", \"corners_3d\": [[77.85, 263.13], [-36.25, 265.08], [105.46, 240.71], [187.65, 239.64], [77.85, 169.86], [-36.25, 169.8], [105.46, 170.61], [187.65, 170.64]]}, {\"category\": \"van\", \"corners_3d\": [[549.44, 190.91], [527.14, 190.9], [534.22, 189.47], [554.75, 189.47], [549.44, 168.25], [527.14, 168.25], [534.22, 168.62], [554.75, 168.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[730.33, 212.67], [744.57, 213.43], [728.65, 214.36], [714.39, 213.58], [730.33, 166.35], [744.57, 166.22], [728.65, 166.07], [714.39, 166.2]]}]\n```" - } - ] - }, - { - "id": 507, - "data_source": "Mono3DRefer", - "scene_name": "000276", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[502.85, 240.06], [441.77, 240.12], [469.17, 228.6], [519.78, 228.56], [502.85, 180.31], [441.77, 180.32], [469.17, 179.04], [519.78, 179.03]]}, {\"category\": \"car\", \"corners_3d\": [[565.27, 204.59], [538.38, 204.58], [543.73, 202.39], [568.78, 202.4], [565.27, 179.99], [538.38, 179.99], [543.73, 179.5], [568.78, 179.5]]}, {\"category\": \"car\", \"corners_3d\": [[583.26, 194.26], [568.89, 194.26], [570.78, 193.25], [584.47, 193.25], [583.26, 179.51], [568.89, 179.51], [570.78, 179.2], [584.47, 179.2]]}, {\"category\": \"car\", \"corners_3d\": [[396.47, 196.26], [379.48, 196.26], [392.61, 194.96], [408.67, 194.97], [396.47, 175.55], [379.48, 175.55], [392.61, 175.4], [408.67, 175.4]]}, {\"category\": \"car\", \"corners_3d\": [[593.09, 212.51], [629.21, 212.51], [632.26, 218.87], [590.35, 218.87], [593.09, 176.61], [629.21, 176.61], [632.26, 177.22], [590.35, 177.22]]}]\n```" - } - ] - }, - { - "id": 508, - "data_source": "Mono3DRefer", - "scene_name": "000277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[1121.01, 305.59], [1076.03, 305.12], [1052.55, 297.79], [1094.96, 298.21], [1121.01, 208.57], [1076.03, 208.44], [1052.55, 206.47], [1094.96, 206.58]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1182.52, 302.21], [1139.33, 301.91], [1111.3, 294.7], [1152.01, 294.96], [1182.52, 213.99], [1139.33, 213.9], [1111.3, 211.6], [1152.01, 211.69]]}]\n```" - } - ] - }, - { - "id": 509, - "data_source": "Mono3DRefer", - "scene_name": "000277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[755.08, 244.72], [747.96, 241.55], [770.01, 241.5], [778.15, 244.67], [755.08, 159.39], [747.96, 159.98], [770.01, 159.99], [778.15, 159.4]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1200.77, 291.86], [1139.27, 287.05], [1140.51, 282.23], [1199.36, 286.63], [1200.77, 150.7], [1139.27, 151.59], [1140.51, 152.49], [1199.36, 151.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[957.32, 253.93], [979.1, 257.3], [955.43, 258.07], [934.4, 254.64], [957.32, 165.04], [979.1, 164.71], [955.43, 164.64], [934.4, 164.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[944.15, 248.12], [924.05, 247.86], [916.71, 245.37], [936.12, 245.61], [944.15, 164.96], [924.05, 164.99], [916.71, 165.25], [936.12, 165.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[313.18, 236.37], [331.04, 236.37], [313.75, 240.3], [294.79, 240.3], [313.18, 168.87], [331.04, 168.87], [313.75, 168.62], [294.79, 168.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[402.41, 235.66], [394.79, 238.06], [371.93, 238.08], [380.39, 235.68], [402.41, 167.54], [394.79, 167.34], [371.93, 167.34], [380.39, 167.54]]}]\n```" - } - ] - }, - { - "id": 510, - "data_source": "Mono3DRefer", - "scene_name": "000277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[583.24, 196.32], [615.28, 196.56], [593.35, 199.74], [557.33, 199.42], [583.24, 156.5], [615.28, 156.33], [593.35, 154.12], [557.33, 154.34]]}]\n```" - } - ] - }, - { - "id": 511, - "data_source": "Mono3DRefer", - "scene_name": "000277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[875.5, 266.56], [910.27, 269.16], [879.02, 282.46], [840.55, 279.1], [875.5, 160.45], [910.27, 160.11], [879.02, 158.35], [840.55, 158.79]]}]\n```" - } - ] - }, - { - "id": 512, - "data_source": "Mono3DRefer", - "scene_name": "000277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[755.08, 244.72], [747.96, 241.55], [770.01, 241.5], [778.15, 244.67], [755.08, 159.39], [747.96, 159.98], [770.01, 159.99], [778.15, 159.4]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1121.01, 305.59], [1076.03, 305.12], [1052.55, 297.79], [1094.96, 298.21], [1121.01, 208.57], [1076.03, 208.44], [1052.55, 206.47], [1094.96, 206.58]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1182.52, 302.21], [1139.33, 301.91], [1111.3, 294.7], [1152.01, 294.96], [1182.52, 213.99], [1139.33, 213.9], [1111.3, 211.6], [1152.01, 211.69]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1200.77, 291.86], [1139.27, 287.05], [1140.51, 282.23], [1199.36, 286.63], [1200.77, 150.7], [1139.27, 151.59], [1140.51, 152.49], [1199.36, 151.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[957.32, 253.93], [979.1, 257.3], [955.43, 258.07], [934.4, 254.64], [957.32, 165.04], [979.1, 164.71], [955.43, 164.64], [934.4, 164.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[944.15, 248.12], [924.05, 247.86], [916.71, 245.37], [936.12, 245.61], [944.15, 164.96], [924.05, 164.99], [916.71, 165.25], [936.12, 165.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[313.18, 236.37], [331.04, 236.37], [313.75, 240.3], [294.79, 240.3], [313.18, 168.87], [331.04, 168.87], [313.75, 168.62], [294.79, 168.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[402.41, 235.66], [394.79, 238.06], [371.93, 238.08], [380.39, 235.68], [402.41, 167.54], [394.79, 167.34], [371.93, 167.34], [380.39, 167.54]]}]\n```" - } - ] - }, - { - "id": 513, - "data_source": "Mono3DRefer", - "scene_name": "000277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[1121.01, 305.59], [1076.03, 305.12], [1052.55, 297.79], [1094.96, 298.21], [1121.01, 208.57], [1076.03, 208.44], [1052.55, 206.47], [1094.96, 206.58]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1182.52, 302.21], [1139.33, 301.91], [1111.3, 294.7], [1152.01, 294.96], [1182.52, 213.99], [1139.33, 213.9], [1111.3, 211.6], [1152.01, 211.69]]}, {\"category\": \"van\", \"corners_3d\": [[583.24, 196.32], [615.28, 196.56], [593.35, 199.74], [557.33, 199.42], [583.24, 156.5], [615.28, 156.33], [593.35, 154.12], [557.33, 154.34]]}]\n```" - } - ] - }, - { - "id": 514, - "data_source": "Mono3DRefer", - "scene_name": "000277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[875.5, 266.56], [910.27, 269.16], [879.02, 282.46], [840.55, 279.1], [875.5, 160.45], [910.27, 160.11], [879.02, 158.35], [840.55, 158.79]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1121.01, 305.59], [1076.03, 305.12], [1052.55, 297.79], [1094.96, 298.21], [1121.01, 208.57], [1076.03, 208.44], [1052.55, 206.47], [1094.96, 206.58]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1182.52, 302.21], [1139.33, 301.91], [1111.3, 294.7], [1152.01, 294.96], [1182.52, 213.99], [1139.33, 213.9], [1111.3, 211.6], [1152.01, 211.69]]}]\n```" - } - ] - }, - { - "id": 515, - "data_source": "Mono3DRefer", - "scene_name": "000277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[755.08, 244.72], [747.96, 241.55], [770.01, 241.5], [778.15, 244.67], [755.08, 159.39], [747.96, 159.98], [770.01, 159.99], [778.15, 159.4]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1200.77, 291.86], [1139.27, 287.05], [1140.51, 282.23], [1199.36, 286.63], [1200.77, 150.7], [1139.27, 151.59], [1140.51, 152.49], [1199.36, 151.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[957.32, 253.93], [979.1, 257.3], [955.43, 258.07], [934.4, 254.64], [957.32, 165.04], [979.1, 164.71], [955.43, 164.64], [934.4, 164.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[944.15, 248.12], [924.05, 247.86], [916.71, 245.37], [936.12, 245.61], [944.15, 164.96], [924.05, 164.99], [916.71, 165.25], [936.12, 165.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[313.18, 236.37], [331.04, 236.37], [313.75, 240.3], [294.79, 240.3], [313.18, 168.87], [331.04, 168.87], [313.75, 168.62], [294.79, 168.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[402.41, 235.66], [394.79, 238.06], [371.93, 238.08], [380.39, 235.68], [402.41, 167.54], [394.79, 167.34], [371.93, 167.34], [380.39, 167.54]]}, {\"category\": \"van\", \"corners_3d\": [[583.24, 196.32], [615.28, 196.56], [593.35, 199.74], [557.33, 199.42], [583.24, 156.5], [615.28, 156.33], [593.35, 154.12], [557.33, 154.34]]}]\n```" - } - ] - }, - { - "id": 516, - "data_source": "Mono3DRefer", - "scene_name": "000277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[875.5, 266.56], [910.27, 269.16], [879.02, 282.46], [840.55, 279.1], [875.5, 160.45], [910.27, 160.11], [879.02, 158.35], [840.55, 158.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[755.08, 244.72], [747.96, 241.55], [770.01, 241.5], [778.15, 244.67], [755.08, 159.39], [747.96, 159.98], [770.01, 159.99], [778.15, 159.4]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1200.77, 291.86], [1139.27, 287.05], [1140.51, 282.23], [1199.36, 286.63], [1200.77, 150.7], [1139.27, 151.59], [1140.51, 152.49], [1199.36, 151.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[957.32, 253.93], [979.1, 257.3], [955.43, 258.07], [934.4, 254.64], [957.32, 165.04], [979.1, 164.71], [955.43, 164.64], [934.4, 164.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[944.15, 248.12], [924.05, 247.86], [916.71, 245.37], [936.12, 245.61], [944.15, 164.96], [924.05, 164.99], [916.71, 165.25], [936.12, 165.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[313.18, 236.37], [331.04, 236.37], [313.75, 240.3], [294.79, 240.3], [313.18, 168.87], [331.04, 168.87], [313.75, 168.62], [294.79, 168.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[402.41, 235.66], [394.79, 238.06], [371.93, 238.08], [380.39, 235.68], [402.41, 167.54], [394.79, 167.34], [371.93, 167.34], [380.39, 167.54]]}]\n```" - } - ] - }, - { - "id": 517, - "data_source": "Mono3DRefer", - "scene_name": "000277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[875.5, 266.56], [910.27, 269.16], [879.02, 282.46], [840.55, 279.1], [875.5, 160.45], [910.27, 160.11], [879.02, 158.35], [840.55, 158.79]]}, {\"category\": \"van\", \"corners_3d\": [[583.24, 196.32], [615.28, 196.56], [593.35, 199.74], [557.33, 199.42], [583.24, 156.5], [615.28, 156.33], [593.35, 154.12], [557.33, 154.34]]}]\n```" - } - ] - }, - { - "id": 518, - "data_source": "Mono3DRefer", - "scene_name": "000277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[755.08, 244.72], [747.96, 241.55], [770.01, 241.5], [778.15, 244.67], [755.08, 159.39], [747.96, 159.98], [770.01, 159.99], [778.15, 159.4]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1121.01, 305.59], [1076.03, 305.12], [1052.55, 297.79], [1094.96, 298.21], [1121.01, 208.57], [1076.03, 208.44], [1052.55, 206.47], [1094.96, 206.58]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1182.52, 302.21], [1139.33, 301.91], [1111.3, 294.7], [1152.01, 294.96], [1182.52, 213.99], [1139.33, 213.9], [1111.3, 211.6], [1152.01, 211.69]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1200.77, 291.86], [1139.27, 287.05], [1140.51, 282.23], [1199.36, 286.63], [1200.77, 150.7], [1139.27, 151.59], [1140.51, 152.49], [1199.36, 151.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[957.32, 253.93], [979.1, 257.3], [955.43, 258.07], [934.4, 254.64], [957.32, 165.04], [979.1, 164.71], [955.43, 164.64], [934.4, 164.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[944.15, 248.12], [924.05, 247.86], [916.71, 245.37], [936.12, 245.61], [944.15, 164.96], [924.05, 164.99], [916.71, 165.25], [936.12, 165.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[313.18, 236.37], [331.04, 236.37], [313.75, 240.3], [294.79, 240.3], [313.18, 168.87], [331.04, 168.87], [313.75, 168.62], [294.79, 168.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[402.41, 235.66], [394.79, 238.06], [371.93, 238.08], [380.39, 235.68], [402.41, 167.54], [394.79, 167.34], [371.93, 167.34], [380.39, 167.54]]}, {\"category\": \"van\", \"corners_3d\": [[583.24, 196.32], [615.28, 196.56], [593.35, 199.74], [557.33, 199.42], [583.24, 156.5], [615.28, 156.33], [593.35, 154.12], [557.33, 154.34]]}]\n```" - } - ] - }, - { - "id": 519, - "data_source": "Mono3DRefer", - "scene_name": "000277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[875.5, 266.56], [910.27, 269.16], [879.02, 282.46], [840.55, 279.1], [875.5, 160.45], [910.27, 160.11], [879.02, 158.35], [840.55, 158.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[755.08, 244.72], [747.96, 241.55], [770.01, 241.5], [778.15, 244.67], [755.08, 159.39], [747.96, 159.98], [770.01, 159.99], [778.15, 159.4]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1121.01, 305.59], [1076.03, 305.12], [1052.55, 297.79], [1094.96, 298.21], [1121.01, 208.57], [1076.03, 208.44], [1052.55, 206.47], [1094.96, 206.58]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1182.52, 302.21], [1139.33, 301.91], [1111.3, 294.7], [1152.01, 294.96], [1182.52, 213.99], [1139.33, 213.9], [1111.3, 211.6], [1152.01, 211.69]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1200.77, 291.86], [1139.27, 287.05], [1140.51, 282.23], [1199.36, 286.63], [1200.77, 150.7], [1139.27, 151.59], [1140.51, 152.49], [1199.36, 151.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[957.32, 253.93], [979.1, 257.3], [955.43, 258.07], [934.4, 254.64], [957.32, 165.04], [979.1, 164.71], [955.43, 164.64], [934.4, 164.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[944.15, 248.12], [924.05, 247.86], [916.71, 245.37], [936.12, 245.61], [944.15, 164.96], [924.05, 164.99], [916.71, 165.25], [936.12, 165.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[313.18, 236.37], [331.04, 236.37], [313.75, 240.3], [294.79, 240.3], [313.18, 168.87], [331.04, 168.87], [313.75, 168.62], [294.79, 168.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[402.41, 235.66], [394.79, 238.06], [371.93, 238.08], [380.39, 235.68], [402.41, 167.54], [394.79, 167.34], [371.93, 167.34], [380.39, 167.54]]}]\n```" - } - ] - }, - { - "id": 520, - "data_source": "Mono3DRefer", - "scene_name": "000277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[875.5, 266.56], [910.27, 269.16], [879.02, 282.46], [840.55, 279.1], [875.5, 160.45], [910.27, 160.11], [879.02, 158.35], [840.55, 158.79]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1121.01, 305.59], [1076.03, 305.12], [1052.55, 297.79], [1094.96, 298.21], [1121.01, 208.57], [1076.03, 208.44], [1052.55, 206.47], [1094.96, 206.58]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1182.52, 302.21], [1139.33, 301.91], [1111.3, 294.7], [1152.01, 294.96], [1182.52, 213.99], [1139.33, 213.9], [1111.3, 211.6], [1152.01, 211.69]]}, {\"category\": \"van\", \"corners_3d\": [[583.24, 196.32], [615.28, 196.56], [593.35, 199.74], [557.33, 199.42], [583.24, 156.5], [615.28, 156.33], [593.35, 154.12], [557.33, 154.34]]}]\n```" - } - ] - }, - { - "id": 521, - "data_source": "Mono3DRefer", - "scene_name": "000277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[875.5, 266.56], [910.27, 269.16], [879.02, 282.46], [840.55, 279.1], [875.5, 160.45], [910.27, 160.11], [879.02, 158.35], [840.55, 158.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[755.08, 244.72], [747.96, 241.55], [770.01, 241.5], [778.15, 244.67], [755.08, 159.39], [747.96, 159.98], [770.01, 159.99], [778.15, 159.4]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1200.77, 291.86], [1139.27, 287.05], [1140.51, 282.23], [1199.36, 286.63], [1200.77, 150.7], [1139.27, 151.59], [1140.51, 152.49], [1199.36, 151.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[957.32, 253.93], [979.1, 257.3], [955.43, 258.07], [934.4, 254.64], [957.32, 165.04], [979.1, 164.71], [955.43, 164.64], [934.4, 164.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[944.15, 248.12], [924.05, 247.86], [916.71, 245.37], [936.12, 245.61], [944.15, 164.96], [924.05, 164.99], [916.71, 165.25], [936.12, 165.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[313.18, 236.37], [331.04, 236.37], [313.75, 240.3], [294.79, 240.3], [313.18, 168.87], [331.04, 168.87], [313.75, 168.62], [294.79, 168.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[402.41, 235.66], [394.79, 238.06], [371.93, 238.08], [380.39, 235.68], [402.41, 167.54], [394.79, 167.34], [371.93, 167.34], [380.39, 167.54]]}, {\"category\": \"van\", \"corners_3d\": [[583.24, 196.32], [615.28, 196.56], [593.35, 199.74], [557.33, 199.42], [583.24, 156.5], [615.28, 156.33], [593.35, 154.12], [557.33, 154.34]]}]\n```" - } - ] - }, - { - "id": 522, - "data_source": "Mono3DRefer", - "scene_name": "000277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, pedestrian, van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[875.5, 266.56], [910.27, 269.16], [879.02, 282.46], [840.55, 279.1], [875.5, 160.45], [910.27, 160.11], [879.02, 158.35], [840.55, 158.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[755.08, 244.72], [747.96, 241.55], [770.01, 241.5], [778.15, 244.67], [755.08, 159.39], [747.96, 159.98], [770.01, 159.99], [778.15, 159.4]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1121.01, 305.59], [1076.03, 305.12], [1052.55, 297.79], [1094.96, 298.21], [1121.01, 208.57], [1076.03, 208.44], [1052.55, 206.47], [1094.96, 206.58]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1182.52, 302.21], [1139.33, 301.91], [1111.3, 294.7], [1152.01, 294.96], [1182.52, 213.99], [1139.33, 213.9], [1111.3, 211.6], [1152.01, 211.69]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1200.77, 291.86], [1139.27, 287.05], [1140.51, 282.23], [1199.36, 286.63], [1200.77, 150.7], [1139.27, 151.59], [1140.51, 152.49], [1199.36, 151.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[957.32, 253.93], [979.1, 257.3], [955.43, 258.07], [934.4, 254.64], [957.32, 165.04], [979.1, 164.71], [955.43, 164.64], [934.4, 164.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[944.15, 248.12], [924.05, 247.86], [916.71, 245.37], [936.12, 245.61], [944.15, 164.96], [924.05, 164.99], [916.71, 165.25], [936.12, 165.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[313.18, 236.37], [331.04, 236.37], [313.75, 240.3], [294.79, 240.3], [313.18, 168.87], [331.04, 168.87], [313.75, 168.62], [294.79, 168.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[402.41, 235.66], [394.79, 238.06], [371.93, 238.08], [380.39, 235.68], [402.41, 167.54], [394.79, 167.34], [371.93, 167.34], [380.39, 167.54]]}, {\"category\": \"van\", \"corners_3d\": [[583.24, 196.32], [615.28, 196.56], [593.35, 199.74], [557.33, 199.42], [583.24, 156.5], [615.28, 156.33], [593.35, 154.12], [557.33, 154.34]]}]\n```" - } - ] - }, - { - "id": 523, - "data_source": "Mono3DRefer", - "scene_name": "000280", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[591.2, 199.63], [629.9, 199.63], [634.87, 206.36], [586.45, 206.35], [591.2, 152.49], [629.9, 152.49], [634.87, 147.38], [586.45, 147.38]]}]\n```" - } - ] - }, - { - "id": 524, - "data_source": "Mono3DRefer", - "scene_name": "000280", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[299.62, 223.66], [265.28, 223.66], [299.42, 218.62], [330.35, 218.61], [299.62, 195.28], [265.28, 195.28], [299.42, 193.06], [330.35, 193.05]]}]\n```" - } - ] - }, - { - "id": 525, - "data_source": "Mono3DRefer", - "scene_name": "000280", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[591.2, 199.63], [629.9, 199.63], [634.87, 206.36], [586.45, 206.35], [591.2, 152.49], [629.9, 152.49], [634.87, 147.38], [586.45, 147.38]]}, {\"category\": \"car\", \"corners_3d\": [[299.62, 223.66], [265.28, 223.66], [299.42, 218.62], [330.35, 218.61], [299.62, 195.28], [265.28, 195.28], [299.42, 193.06], [330.35, 193.05]]}]\n```" - } - ] - }, - { - "id": 526, - "data_source": "Mono3DRefer", - "scene_name": "000281", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[594.54, 199.28], [624.62, 199.28], [625.77, 201.37], [593.3, 201.37], [594.54, 170.41], [624.62, 170.41], [625.77, 170.22], [593.3, 170.22]]}]\n```" - } - ] - }, - { - "id": 527, - "data_source": "Mono3DRefer", - "scene_name": "000281", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[678.52, 187.07], [701.87, 186.99], [742.62, 189.57], [715.26, 189.69], [678.52, 160.75], [701.87, 160.82], [742.62, 158.62], [715.26, 158.52]]}]\n```" - } - ] - }, - { - "id": 528, - "data_source": "Mono3DRefer", - "scene_name": "000281", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[380.86, 212.82], [325.99, 212.89], [378.85, 205.01], [422.87, 204.97], [380.86, 142.51], [325.99, 142.46], [378.85, 148.45], [422.87, 148.48]]}]\n```" - } - ] - }, - { - "id": 529, - "data_source": "Mono3DRefer", - "scene_name": "000281", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[678.52, 187.07], [701.87, 186.99], [742.62, 189.57], [715.26, 189.69], [678.52, 160.75], [701.87, 160.82], [742.62, 158.62], [715.26, 158.52]]}, {\"category\": \"car\", \"corners_3d\": [[594.54, 199.28], [624.62, 199.28], [625.77, 201.37], [593.3, 201.37], [594.54, 170.41], [624.62, 170.41], [625.77, 170.22], [593.3, 170.22]]}]\n```" - } - ] - }, - { - "id": 530, - "data_source": "Mono3DRefer", - "scene_name": "000281", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[380.86, 212.82], [325.99, 212.89], [378.85, 205.01], [422.87, 204.97], [380.86, 142.51], [325.99, 142.46], [378.85, 148.45], [422.87, 148.48]]}, {\"category\": \"car\", \"corners_3d\": [[594.54, 199.28], [624.62, 199.28], [625.77, 201.37], [593.3, 201.37], [594.54, 170.41], [624.62, 170.41], [625.77, 170.22], [593.3, 170.22]]}]\n```" - } - ] - }, - { - "id": 531, - "data_source": "Mono3DRefer", - "scene_name": "000281", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[678.52, 187.07], [701.87, 186.99], [742.62, 189.57], [715.26, 189.69], [678.52, 160.75], [701.87, 160.82], [742.62, 158.62], [715.26, 158.52]]}, {\"category\": \"truck\", \"corners_3d\": [[380.86, 212.82], [325.99, 212.89], [378.85, 205.01], [422.87, 204.97], [380.86, 142.51], [325.99, 142.46], [378.85, 148.45], [422.87, 148.48]]}]\n```" - } - ] - }, - { - "id": 532, - "data_source": "Mono3DRefer", - "scene_name": "000281", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[678.52, 187.07], [701.87, 186.99], [742.62, 189.57], [715.26, 189.69], [678.52, 160.75], [701.87, 160.82], [742.62, 158.62], [715.26, 158.52]]}, {\"category\": \"truck\", \"corners_3d\": [[380.86, 212.82], [325.99, 212.89], [378.85, 205.01], [422.87, 204.97], [380.86, 142.51], [325.99, 142.46], [378.85, 148.45], [422.87, 148.48]]}, {\"category\": \"car\", \"corners_3d\": [[594.54, 199.28], [624.62, 199.28], [625.77, 201.37], [593.3, 201.37], [594.54, 170.41], [624.62, 170.41], [625.77, 170.22], [593.3, 170.22]]}]\n```" - } - ] - }, - { - "id": 533, - "data_source": "Mono3DRefer", - "scene_name": "000285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[597.34, 232.29], [651.5, 232.24], [661.29, 244.22], [596.21, 244.28], [597.34, 181.05], [651.5, 181.05], [661.29, 182.7], [596.21, 182.71]]}, {\"category\": \"car\", \"corners_3d\": [[953.12, 266.06], [1076.98, 276.84], [988.61, 313.96], [838.5, 294.83], [953.12, 170.98], [1076.98, 170.76], [988.61, 170.02], [838.5, 170.4]]}, {\"category\": \"car\", \"corners_3d\": [[783.01, 227.47], [827.18, 230.14], [762.68, 237.57], [716.46, 234.18], [783.01, 178.63], [827.18, 178.91], [762.68, 179.69], [716.46, 179.34]]}, {\"category\": \"car\", \"corners_3d\": [[737.07, 214.94], [771.23, 216.68], [713.07, 221.05], [678.17, 218.96], [737.07, 180.52], [771.23, 180.84], [713.07, 181.63], [678.17, 181.25]]}, {\"category\": \"car\", \"corners_3d\": [[410.78, 243.68], [353.45, 243.4], [399.97, 232.14], [448.29, 232.34], [410.78, 185.26], [353.45, 185.21], [399.97, 183.24], [448.29, 183.27]]}, {\"category\": \"car\", \"corners_3d\": [[517.44, 213.52], [484.3, 213.33], [503.61, 209.36], [533.57, 209.51], [517.44, 181.42], [484.3, 181.38], [503.61, 180.54], [533.57, 180.58]]}, {\"category\": \"car\", \"corners_3d\": [[229.27, 253.96], [164.87, 253.75], [259.33, 237.34], [310.84, 237.48], [229.27, 196.3], [164.87, 196.24], [259.33, 191.5], [310.84, 191.54]]}, {\"category\": \"car\", \"corners_3d\": [[327.18, 234.34], [280.82, 234.3], [326.93, 225.85], [366.93, 225.88], [327.18, 181.34], [280.82, 181.33], [326.93, 180.17], [366.93, 180.17]]}, {\"category\": \"car\", \"corners_3d\": [[579.04, 199.63], [555.6, 199.51], [567.29, 197.5], [589.0, 197.6], [579.04, 180.78], [555.6, 180.74], [567.29, 180.15], [589.0, 180.18]]}, {\"category\": \"car\", \"corners_3d\": [[420.76, 215.34], [388.66, 215.12], [420.44, 210.65], [449.27, 210.82], [420.76, 183.16], [388.66, 183.11], [420.44, 182.02], [449.27, 182.06]]}, {\"category\": \"car\", \"corners_3d\": [[495.39, 205.7], [472.19, 205.39], [500.31, 202.67], [521.79, 202.93], [495.39, 184.1], [472.19, 184.0], [500.31, 183.07], [521.79, 183.15]]}, {\"category\": \"car\", \"corners_3d\": [[10.28, 252.68], [-44.47, 251.87], [69.95, 239.66], [117.16, 240.24], [10.28, 191.23], [-44.47, 191.05], [69.95, 188.23], [117.16, 188.37]]}, {\"category\": \"car\", \"corners_3d\": [[45.26, 226.31], [11.76, 225.25], [120.36, 218.38], [151.29, 219.18], [45.26, 186.82], [11.76, 186.54], [120.36, 184.75], [151.29, 184.95]]}, {\"category\": \"car\", \"corners_3d\": [[667.53, 190.89], [685.13, 191.0], [679.57, 191.99], [661.05, 191.88], [667.53, 176.87], [685.13, 176.89], [679.57, 177.12], [661.05, 177.09]]}]\n```" - } - ] - }, - { - "id": 534, - "data_source": "Mono3DRefer", - "scene_name": "000286", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[425.77, 212.6], [419.91, 214.1], [347.64, 214.27], [356.17, 212.75], [425.77, 186.1], [419.91, 186.6], [347.64, 186.65], [356.17, 186.15]]}, {\"category\": \"car\", \"corners_3d\": [[331.65, 214.29], [320.45, 215.92], [243.78, 215.87], [257.87, 214.25], [331.65, 186.66], [320.45, 187.2], [243.78, 187.19], [257.87, 186.65]]}, {\"category\": \"car\", \"corners_3d\": [[627.69, 196.88], [647.45, 196.86], [652.39, 198.65], [631.16, 198.68], [627.69, 177.83], [647.45, 177.83], [652.39, 178.2], [631.16, 178.21]]}, {\"category\": \"car\", \"corners_3d\": [[590.61, 197.9], [613.29, 197.89], [614.08, 199.8], [589.67, 199.81], [590.61, 176.47], [613.29, 176.47], [614.08, 176.74], [589.67, 176.74]]}, {\"category\": \"car\", \"corners_3d\": [[491.4, 266.77], [406.52, 266.78], [449.01, 247.07], [516.08, 247.07], [491.4, 189.38], [406.52, 189.38], [449.01, 185.91], [516.08, 185.91]]}, {\"category\": \"car\", \"corners_3d\": [[958.58, 424.91], [1186.13, 422.56], [3655.79, 1454.41], [2564.98, 1518.97], [958.58, 190.24], [1186.13, 190.08], [3655.79, 261.25], [2564.98, 265.7]]}, {\"category\": \"car\", \"corners_3d\": [[763.16, 307.84], [893.17, 307.37], [1116.78, 408.49], [889.81, 409.93], [763.16, 176.82], [893.17, 176.81], [1116.78, 179.78], [889.81, 179.82]]}, {\"category\": \"car\", \"corners_3d\": [[720.57, 272.28], [800.47, 271.62], [863.73, 298.36], [762.56, 299.43], [720.57, 173.42], [800.47, 173.41], [863.73, 173.57], [762.56, 173.57]]}, {\"category\": \"car\", \"corners_3d\": [[664.89, 226.86], [715.38, 226.71], [742.32, 237.53], [681.75, 237.74], [664.89, 178.92], [715.38, 178.9], [742.32, 180.11], [681.75, 180.14]]}, {\"category\": \"car\", \"corners_3d\": [[639.82, 206.2], [669.6, 206.21], [675.85, 210.21], [642.51, 210.19], [639.82, 177.96], [669.6, 177.97], [675.85, 178.58], [642.51, 178.58]]}, {\"category\": \"car\", \"corners_3d\": [[655.33, 218.15], [696.68, 218.13], [711.34, 225.21], [663.53, 225.24], [655.33, 177.27], [696.68, 177.27], [711.34, 177.96], [663.53, 177.96]]}]\n```" - } - ] - }, - { - "id": 535, - "data_source": "Mono3DRefer", - "scene_name": "000286", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[820.77, 256.43], [858.98, 256.43], [872.68, 261.04], [832.37, 261.03], [820.77, 155.67], [858.98, 155.67], [872.68, 154.73], [832.37, 154.73]]}]\n```" - } - ] - }, - { - "id": 536, - "data_source": "Mono3DRefer", - "scene_name": "000286", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[425.77, 212.6], [419.91, 214.1], [347.64, 214.27], [356.17, 212.75], [425.77, 186.1], [419.91, 186.6], [347.64, 186.65], [356.17, 186.15]]}, {\"category\": \"car\", \"corners_3d\": [[331.65, 214.29], [320.45, 215.92], [243.78, 215.87], [257.87, 214.25], [331.65, 186.66], [320.45, 187.2], [243.78, 187.19], [257.87, 186.65]]}, {\"category\": \"car\", \"corners_3d\": [[627.69, 196.88], [647.45, 196.86], [652.39, 198.65], [631.16, 198.68], [627.69, 177.83], [647.45, 177.83], [652.39, 178.2], [631.16, 178.21]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[820.77, 256.43], [858.98, 256.43], [872.68, 261.04], [832.37, 261.03], [820.77, 155.67], [858.98, 155.67], [872.68, 154.73], [832.37, 154.73]]}, {\"category\": \"car\", \"corners_3d\": [[590.61, 197.9], [613.29, 197.89], [614.08, 199.8], [589.67, 199.81], [590.61, 176.47], [613.29, 176.47], [614.08, 176.74], [589.67, 176.74]]}, {\"category\": \"car\", \"corners_3d\": [[491.4, 266.77], [406.52, 266.78], [449.01, 247.07], [516.08, 247.07], [491.4, 189.38], [406.52, 189.38], [449.01, 185.91], [516.08, 185.91]]}, {\"category\": \"car\", \"corners_3d\": [[958.58, 424.91], [1186.13, 422.56], [3655.79, 1454.41], [2564.98, 1518.97], [958.58, 190.24], [1186.13, 190.08], [3655.79, 261.25], [2564.98, 265.7]]}, {\"category\": \"car\", \"corners_3d\": [[763.16, 307.84], [893.17, 307.37], [1116.78, 408.49], [889.81, 409.93], [763.16, 176.82], [893.17, 176.81], [1116.78, 179.78], [889.81, 179.82]]}, {\"category\": \"car\", \"corners_3d\": [[720.57, 272.28], [800.47, 271.62], [863.73, 298.36], [762.56, 299.43], [720.57, 173.42], [800.47, 173.41], [863.73, 173.57], [762.56, 173.57]]}, {\"category\": \"car\", \"corners_3d\": [[664.89, 226.86], [715.38, 226.71], [742.32, 237.53], [681.75, 237.74], [664.89, 178.92], [715.38, 178.9], [742.32, 180.11], [681.75, 180.14]]}, {\"category\": \"car\", \"corners_3d\": [[639.82, 206.2], [669.6, 206.21], [675.85, 210.21], [642.51, 210.19], [639.82, 177.96], [669.6, 177.97], [675.85, 178.58], [642.51, 178.58]]}, {\"category\": \"car\", \"corners_3d\": [[655.33, 218.15], [696.68, 218.13], [711.34, 225.21], [663.53, 225.24], [655.33, 177.27], [696.68, 177.27], [711.34, 177.96], [663.53, 177.96]]}]\n```" - } - ] - }, - { - "id": 537, - "data_source": "Mono3DRefer", - "scene_name": "000287", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[855.65, 264.59], [798.93, 261.54], [804.71, 258.47], [859.6, 261.31], [855.65, 160.32], [798.93, 160.74], [804.71, 161.16], [859.6, 160.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[774.31, 251.89], [737.73, 250.74], [740.75, 248.24], [776.19, 249.33], [774.31, 175.32], [737.73, 175.29], [740.75, 175.21], [776.19, 175.24]]}]\n```" - } - ] - }, - { - "id": 538, - "data_source": "Mono3DRefer", - "scene_name": "000291", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[529.32, 223.66], [480.39, 223.35], [508.66, 215.98], [550.56, 216.21], [529.32, 181.16], [480.39, 181.11], [508.66, 179.91], [550.56, 179.95]]}, {\"category\": \"car\", \"corners_3d\": [[556.96, 211.1], [520.27, 211.13], [530.06, 206.53], [562.34, 206.51], [556.96, 177.46], [520.27, 177.46], [530.06, 176.9], [562.34, 176.9]]}, {\"category\": \"car\", \"corners_3d\": [[570.56, 197.94], [543.61, 197.9], [552.32, 195.75], [576.97, 195.79], [570.56, 174.86], [543.61, 174.85], [552.32, 174.68], [576.97, 174.68]]}, {\"category\": \"car\", \"corners_3d\": [[619.65, 210.19], [655.44, 210.24], [657.59, 213.64], [618.55, 213.57], [619.65, 174.22], [655.44, 174.22], [657.59, 174.35], [618.55, 174.34]]}]\n```" - } - ] - }, - { - "id": 539, - "data_source": "Mono3DRefer", - "scene_name": "000292", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[749.63, 191.21], [719.15, 191.14], [709.18, 187.1], [732.86, 187.14], [749.63, 140.71], [719.15, 140.83], [709.18, 147.92], [732.86, 147.85]]}]\n```" - } - ] - }, - { - "id": 540, - "data_source": "Mono3DRefer", - "scene_name": "000292", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[425.97, 205.12], [416.67, 205.08], [422.33, 204.37], [431.44, 204.41], [425.97, 175.1], [416.67, 175.1], [422.33, 175.05], [431.44, 175.05]]}]\n```" - } - ] - }, - { - "id": 541, - "data_source": "Mono3DRefer", - "scene_name": "000292", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[602.55, 229.83], [658.93, 229.93], [665.02, 240.05], [598.65, 239.92], [602.55, 175.22], [658.93, 175.22], [665.02, 175.64], [598.65, 175.63]]}]\n```" - } - ] - }, - { - "id": 542, - "data_source": "Mono3DRefer", - "scene_name": "000292", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[749.63, 191.21], [719.15, 191.14], [709.18, 187.1], [732.86, 187.14], [749.63, 140.71], [719.15, 140.83], [709.18, 147.92], [732.86, 147.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[425.97, 205.12], [416.67, 205.08], [422.33, 204.37], [431.44, 204.41], [425.97, 175.1], [416.67, 175.1], [422.33, 175.05], [431.44, 175.05]]}]\n```" - } - ] - }, - { - "id": 543, - "data_source": "Mono3DRefer", - "scene_name": "000292", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[602.55, 229.83], [658.93, 229.93], [665.02, 240.05], [598.65, 239.92], [602.55, 175.22], [658.93, 175.22], [665.02, 175.64], [598.65, 175.63]]}, {\"category\": \"tram\", \"corners_3d\": [[749.63, 191.21], [719.15, 191.14], [709.18, 187.1], [732.86, 187.14], [749.63, 140.71], [719.15, 140.83], [709.18, 147.92], [732.86, 147.85]]}]\n```" - } - ] - }, - { - "id": 544, - "data_source": "Mono3DRefer", - "scene_name": "000292", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[602.55, 229.83], [658.93, 229.93], [665.02, 240.05], [598.65, 239.92], [602.55, 175.22], [658.93, 175.22], [665.02, 175.64], [598.65, 175.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[425.97, 205.12], [416.67, 205.08], [422.33, 204.37], [431.44, 204.41], [425.97, 175.1], [416.67, 175.1], [422.33, 175.05], [431.44, 175.05]]}]\n```" - } - ] - }, - { - "id": 545, - "data_source": "Mono3DRefer", - "scene_name": "000292", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[602.55, 229.83], [658.93, 229.93], [665.02, 240.05], [598.65, 239.92], [602.55, 175.22], [658.93, 175.22], [665.02, 175.64], [598.65, 175.63]]}, {\"category\": \"tram\", \"corners_3d\": [[749.63, 191.21], [719.15, 191.14], [709.18, 187.1], [732.86, 187.14], [749.63, 140.71], [719.15, 140.83], [709.18, 147.92], [732.86, 147.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[425.97, 205.12], [416.67, 205.08], [422.33, 204.37], [431.44, 204.41], [425.97, 175.1], [416.67, 175.1], [422.33, 175.05], [431.44, 175.05]]}]\n```" - } - ] - }, - { - "id": 546, - "data_source": "Mono3DRefer", - "scene_name": "000293", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[237.23, 419.44], [-2.53, 417.85], [264.56, 313.31], [402.63, 313.84], [237.23, 204.35], [-2.53, 204.15], [264.56, 190.79], [402.63, 190.86]]}, {\"category\": \"car\", \"corners_3d\": [[500.57, 211.04], [529.46, 210.94], [526.67, 214.47], [495.09, 214.59], [500.57, 181.93], [529.46, 181.91], [526.67, 182.75], [495.09, 182.78]]}, {\"category\": \"car\", \"corners_3d\": [[331.2, 218.49], [338.27, 216.02], [444.95, 214.81], [444.0, 217.13], [331.2, 175.52], [338.27, 175.37], [444.95, 175.3], [444.0, 175.44]]}]\n```" - } - ] - }, - { - "id": 547, - "data_source": "Mono3DRefer", - "scene_name": "000295", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[816.57, 281.27], [850.71, 281.02], [905.44, 302.41], [864.67, 302.78], [816.57, 143.63], [850.71, 143.7], [905.44, 137.94], [864.67, 137.84]]}]\n```" - } - ] - }, - { - "id": 548, - "data_source": "Mono3DRefer", - "scene_name": "000295", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[564.72, 190.07], [549.16, 190.1], [549.92, 189.3], [564.76, 189.28], [564.72, 177.14], [549.16, 177.14], [549.92, 176.94], [564.76, 176.94]]}, {\"category\": \"car\", \"corners_3d\": [[395.15, 281.49], [286.66, 284.32], [339.85, 253.28], [417.7, 251.8], [395.15, 190.54], [286.66, 191.0], [339.85, 185.94], [417.7, 185.7]]}]\n```" - } - ] - }, - { - "id": 549, - "data_source": "Mono3DRefer", - "scene_name": "000295", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[816.57, 281.27], [850.71, 281.02], [905.44, 302.41], [864.67, 302.78], [816.57, 143.63], [850.71, 143.7], [905.44, 137.94], [864.67, 137.84]]}, {\"category\": \"car\", \"corners_3d\": [[564.72, 190.07], [549.16, 190.1], [549.92, 189.3], [564.76, 189.28], [564.72, 177.14], [549.16, 177.14], [549.92, 176.94], [564.76, 176.94]]}, {\"category\": \"car\", \"corners_3d\": [[395.15, 281.49], [286.66, 284.32], [339.85, 253.28], [417.7, 251.8], [395.15, 190.54], [286.66, 191.0], [339.85, 185.94], [417.7, 185.7]]}]\n```" - } - ] - }, - { - "id": 550, - "data_source": "Mono3DRefer", - "scene_name": "000296", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[704.48, 204.01], [739.35, 204.47], [725.56, 208.75], [686.29, 208.15], [704.48, 166.57], [739.35, 166.48], [725.56, 165.62], [686.29, 165.74]]}, {\"category\": \"car\", \"corners_3d\": [[267.71, 239.65], [225.22, 238.82], [298.49, 229.57], [335.74, 230.18], [267.71, 188.31], [225.22, 188.12], [298.49, 185.98], [335.74, 186.12]]}, {\"category\": \"car\", \"corners_3d\": [[117.99, 294.9], [193.35, 301.11], [-290.18, 392.28], [-376.08, 374.7], [117.99, 187.56], [193.35, 188.31], [-290.18, 199.3], [-376.08, 197.18]]}, {\"category\": \"car\", \"corners_3d\": [[464.76, 211.94], [430.19, 211.71], [462.64, 206.96], [493.14, 207.14], [464.76, 182.03], [430.19, 181.98], [462.64, 180.86], [493.14, 180.9]]}]\n```" - } - ] - }, - { - "id": 551, - "data_source": "Mono3DRefer", - "scene_name": "000296", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[439.9, 271.17], [541.35, 277.08], [358.66, 331.69], [223.88, 318.35], [439.9, 161.23], [541.35, 160.53], [358.66, 154.07], [223.88, 155.65]]}]\n```" - } - ] - }, - { - "id": 552, - "data_source": "Mono3DRefer", - "scene_name": "000296", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[664.27, 189.26], [642.72, 189.45], [597.16, 186.98], [615.97, 186.84], [664.27, 157.71], [642.72, 157.54], [597.16, 159.82], [615.97, 159.95]]}]\n```" - } - ] - }, - { - "id": 553, - "data_source": "Mono3DRefer", - "scene_name": "000296", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[704.48, 204.01], [739.35, 204.47], [725.56, 208.75], [686.29, 208.15], [704.48, 166.57], [739.35, 166.48], [725.56, 165.62], [686.29, 165.74]]}, {\"category\": \"van\", \"corners_3d\": [[439.9, 271.17], [541.35, 277.08], [358.66, 331.69], [223.88, 318.35], [439.9, 161.23], [541.35, 160.53], [358.66, 154.07], [223.88, 155.65]]}, {\"category\": \"car\", \"corners_3d\": [[267.71, 239.65], [225.22, 238.82], [298.49, 229.57], [335.74, 230.18], [267.71, 188.31], [225.22, 188.12], [298.49, 185.98], [335.74, 186.12]]}, {\"category\": \"car\", \"corners_3d\": [[117.99, 294.9], [193.35, 301.11], [-290.18, 392.28], [-376.08, 374.7], [117.99, 187.56], [193.35, 188.31], [-290.18, 199.3], [-376.08, 197.18]]}, {\"category\": \"car\", \"corners_3d\": [[464.76, 211.94], [430.19, 211.71], [462.64, 206.96], [493.14, 207.14], [464.76, 182.03], [430.19, 181.98], [462.64, 180.86], [493.14, 180.9]]}]\n```" - } - ] - }, - { - "id": 554, - "data_source": "Mono3DRefer", - "scene_name": "000296", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[704.48, 204.01], [739.35, 204.47], [725.56, 208.75], [686.29, 208.15], [704.48, 166.57], [739.35, 166.48], [725.56, 165.62], [686.29, 165.74]]}, {\"category\": \"car\", \"corners_3d\": [[267.71, 239.65], [225.22, 238.82], [298.49, 229.57], [335.74, 230.18], [267.71, 188.31], [225.22, 188.12], [298.49, 185.98], [335.74, 186.12]]}, {\"category\": \"car\", \"corners_3d\": [[117.99, 294.9], [193.35, 301.11], [-290.18, 392.28], [-376.08, 374.7], [117.99, 187.56], [193.35, 188.31], [-290.18, 199.3], [-376.08, 197.18]]}, {\"category\": \"car\", \"corners_3d\": [[464.76, 211.94], [430.19, 211.71], [462.64, 206.96], [493.14, 207.14], [464.76, 182.03], [430.19, 181.98], [462.64, 180.86], [493.14, 180.9]]}, {\"category\": \"tram\", \"corners_3d\": [[664.27, 189.26], [642.72, 189.45], [597.16, 186.98], [615.97, 186.84], [664.27, 157.71], [642.72, 157.54], [597.16, 159.82], [615.97, 159.95]]}]\n```" - } - ] - }, - { - "id": 555, - "data_source": "Mono3DRefer", - "scene_name": "000296", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[439.9, 271.17], [541.35, 277.08], [358.66, 331.69], [223.88, 318.35], [439.9, 161.23], [541.35, 160.53], [358.66, 154.07], [223.88, 155.65]]}, {\"category\": \"tram\", \"corners_3d\": [[664.27, 189.26], [642.72, 189.45], [597.16, 186.98], [615.97, 186.84], [664.27, 157.71], [642.72, 157.54], [597.16, 159.82], [615.97, 159.95]]}]\n```" - } - ] - }, - { - "id": 556, - "data_source": "Mono3DRefer", - "scene_name": "000296", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[704.48, 204.01], [739.35, 204.47], [725.56, 208.75], [686.29, 208.15], [704.48, 166.57], [739.35, 166.48], [725.56, 165.62], [686.29, 165.74]]}, {\"category\": \"van\", \"corners_3d\": [[439.9, 271.17], [541.35, 277.08], [358.66, 331.69], [223.88, 318.35], [439.9, 161.23], [541.35, 160.53], [358.66, 154.07], [223.88, 155.65]]}, {\"category\": \"car\", \"corners_3d\": [[267.71, 239.65], [225.22, 238.82], [298.49, 229.57], [335.74, 230.18], [267.71, 188.31], [225.22, 188.12], [298.49, 185.98], [335.74, 186.12]]}, {\"category\": \"car\", \"corners_3d\": [[117.99, 294.9], [193.35, 301.11], [-290.18, 392.28], [-376.08, 374.7], [117.99, 187.56], [193.35, 188.31], [-290.18, 199.3], [-376.08, 197.18]]}, {\"category\": \"car\", \"corners_3d\": [[464.76, 211.94], [430.19, 211.71], [462.64, 206.96], [493.14, 207.14], [464.76, 182.03], [430.19, 181.98], [462.64, 180.86], [493.14, 180.9]]}, {\"category\": \"tram\", \"corners_3d\": [[664.27, 189.26], [642.72, 189.45], [597.16, 186.98], [615.97, 186.84], [664.27, 157.71], [642.72, 157.54], [597.16, 159.82], [615.97, 159.95]]}]\n```" - } - ] - }, - { - "id": 557, - "data_source": "Mono3DRefer", - "scene_name": "000297", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[335.01, 243.0], [269.19, 243.13], [323.57, 231.39], [378.33, 231.3], [335.01, 184.53], [269.19, 184.56], [323.57, 182.6], [378.33, 182.59]]}, {\"category\": \"van\", \"corners_3d\": [[918.06, 305.84], [1114.87, 305.14], [1927.08, 512.1], [1429.31, 516.75], [918.06, 100.78], [1114.87, 101.16], [1927.08, -11.01], [1429.31, -13.54]]}]\n```" - } - ] - }, - { - "id": 558, - "data_source": "Mono3DRefer", - "scene_name": "000297", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[515.89, 200.4], [485.7, 200.44], [494.6, 197.84], [521.94, 197.8], [515.89, 163.77], [485.7, 163.76], [494.6, 164.62], [521.94, 164.63]]}]\n```" - } - ] - }, - { - "id": 559, - "data_source": "Mono3DRefer", - "scene_name": "000297", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[446.57, 217.53], [412.77, 217.61], [430.33, 212.91], [460.56, 212.85], [446.57, 187.35], [412.77, 187.38], [430.33, 185.86], [460.56, 185.83]]}, {\"category\": \"car\", \"corners_3d\": [[482.9, 206.49], [459.24, 206.5], [466.28, 204.84], [488.76, 204.83], [482.9, 183.87], [459.24, 183.88], [466.28, 183.33], [488.76, 183.33]]}, {\"category\": \"car\", \"corners_3d\": [[602.05, 193.54], [618.68, 193.53], [620.31, 194.64], [602.79, 194.66], [602.05, 180.44], [618.68, 180.43], [620.31, 180.84], [602.79, 180.85]]}]\n```" - } - ] - }, - { - "id": 560, - "data_source": "Mono3DRefer", - "scene_name": "000297", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[335.01, 243.0], [269.19, 243.13], [323.57, 231.39], [378.33, 231.3], [335.01, 184.53], [269.19, 184.56], [323.57, 182.6], [378.33, 182.59]]}, {\"category\": \"van\", \"corners_3d\": [[918.06, 305.84], [1114.87, 305.14], [1927.08, 512.1], [1429.31, 516.75], [918.06, 100.78], [1114.87, 101.16], [1927.08, -11.01], [1429.31, -13.54]]}, {\"category\": \"truck\", \"corners_3d\": [[515.89, 200.4], [485.7, 200.44], [494.6, 197.84], [521.94, 197.8], [515.89, 163.77], [485.7, 163.76], [494.6, 164.62], [521.94, 164.63]]}]\n```" - } - ] - }, - { - "id": 561, - "data_source": "Mono3DRefer", - "scene_name": "000297", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[335.01, 243.0], [269.19, 243.13], [323.57, 231.39], [378.33, 231.3], [335.01, 184.53], [269.19, 184.56], [323.57, 182.6], [378.33, 182.59]]}, {\"category\": \"van\", \"corners_3d\": [[918.06, 305.84], [1114.87, 305.14], [1927.08, 512.1], [1429.31, 516.75], [918.06, 100.78], [1114.87, 101.16], [1927.08, -11.01], [1429.31, -13.54]]}, {\"category\": \"car\", \"corners_3d\": [[446.57, 217.53], [412.77, 217.61], [430.33, 212.91], [460.56, 212.85], [446.57, 187.35], [412.77, 187.38], [430.33, 185.86], [460.56, 185.83]]}, {\"category\": \"car\", \"corners_3d\": [[482.9, 206.49], [459.24, 206.5], [466.28, 204.84], [488.76, 204.83], [482.9, 183.87], [459.24, 183.88], [466.28, 183.33], [488.76, 183.33]]}, {\"category\": \"car\", \"corners_3d\": [[602.05, 193.54], [618.68, 193.53], [620.31, 194.64], [602.79, 194.66], [602.05, 180.44], [618.68, 180.43], [620.31, 180.84], [602.79, 180.85]]}]\n```" - } - ] - }, - { - "id": 562, - "data_source": "Mono3DRefer", - "scene_name": "000297", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[446.57, 217.53], [412.77, 217.61], [430.33, 212.91], [460.56, 212.85], [446.57, 187.35], [412.77, 187.38], [430.33, 185.86], [460.56, 185.83]]}, {\"category\": \"car\", \"corners_3d\": [[482.9, 206.49], [459.24, 206.5], [466.28, 204.84], [488.76, 204.83], [482.9, 183.87], [459.24, 183.88], [466.28, 183.33], [488.76, 183.33]]}, {\"category\": \"truck\", \"corners_3d\": [[515.89, 200.4], [485.7, 200.44], [494.6, 197.84], [521.94, 197.8], [515.89, 163.77], [485.7, 163.76], [494.6, 164.62], [521.94, 164.63]]}, {\"category\": \"car\", \"corners_3d\": [[602.05, 193.54], [618.68, 193.53], [620.31, 194.64], [602.79, 194.66], [602.05, 180.44], [618.68, 180.43], [620.31, 180.84], [602.79, 180.85]]}]\n```" - } - ] - }, - { - "id": 563, - "data_source": "Mono3DRefer", - "scene_name": "000297", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[335.01, 243.0], [269.19, 243.13], [323.57, 231.39], [378.33, 231.3], [335.01, 184.53], [269.19, 184.56], [323.57, 182.6], [378.33, 182.59]]}, {\"category\": \"van\", \"corners_3d\": [[918.06, 305.84], [1114.87, 305.14], [1927.08, 512.1], [1429.31, 516.75], [918.06, 100.78], [1114.87, 101.16], [1927.08, -11.01], [1429.31, -13.54]]}, {\"category\": \"car\", \"corners_3d\": [[446.57, 217.53], [412.77, 217.61], [430.33, 212.91], [460.56, 212.85], [446.57, 187.35], [412.77, 187.38], [430.33, 185.86], [460.56, 185.83]]}, {\"category\": \"car\", \"corners_3d\": [[482.9, 206.49], [459.24, 206.5], [466.28, 204.84], [488.76, 204.83], [482.9, 183.87], [459.24, 183.88], [466.28, 183.33], [488.76, 183.33]]}, {\"category\": \"truck\", \"corners_3d\": [[515.89, 200.4], [485.7, 200.44], [494.6, 197.84], [521.94, 197.8], [515.89, 163.77], [485.7, 163.76], [494.6, 164.62], [521.94, 164.63]]}, {\"category\": \"car\", \"corners_3d\": [[602.05, 193.54], [618.68, 193.53], [620.31, 194.64], [602.79, 194.66], [602.05, 180.44], [618.68, 180.43], [620.31, 180.84], [602.79, 180.85]]}]\n```" - } - ] - }, - { - "id": 564, - "data_source": "Mono3DRefer", - "scene_name": "000300", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[785.26, 200.79], [801.0, 200.32], [848.62, 202.29], [832.61, 202.84], [785.26, 176.03], [801.0, 175.97], [848.62, 176.2], [832.61, 176.26]]}]\n```" - } - ] - }, - { - "id": 565, - "data_source": "Mono3DRefer", - "scene_name": "000300", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[684.94, 196.44], [710.97, 196.41], [723.62, 198.7], [695.07, 198.74], [684.94, 166.63], [710.97, 166.63], [723.62, 166.03], [695.07, 166.02]]}]\n```" - } - ] - }, - { - "id": 566, - "data_source": "Mono3DRefer", - "scene_name": "000300", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[684.94, 196.44], [710.97, 196.41], [723.62, 198.7], [695.07, 198.74], [684.94, 166.63], [710.97, 166.63], [723.62, 166.03], [695.07, 166.02]]}, {\"category\": \"car\", \"corners_3d\": [[785.26, 200.79], [801.0, 200.32], [848.62, 202.29], [832.61, 202.84], [785.26, 176.03], [801.0, 175.97], [848.62, 176.2], [832.61, 176.26]]}]\n```" - } - ] - }, - { - "id": 567, - "data_source": "Mono3DRefer", - "scene_name": "000301", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[561.4, 230.39], [621.66, 230.32], [627.97, 240.4], [554.24, 240.51], [561.4, 175.59], [621.66, 175.6], [627.97, 173.45], [554.24, 173.43]]}, {\"category\": \"car\", \"corners_3d\": [[423.71, 250.42], [496.36, 250.3], [459.7, 274.89], [359.41, 275.13], [423.71, 191.92], [496.36, 191.9], [459.7, 194.43], [359.41, 194.45]]}, {\"category\": \"car\", \"corners_3d\": [[472.74, 226.63], [525.31, 226.6], [507.81, 236.22], [442.98, 236.26], [472.74, 180.43], [525.31, 180.44], [507.81, 179.33], [442.98, 179.32]]}, {\"category\": \"car\", \"corners_3d\": [[501.66, 215.56], [543.34, 215.55], [534.57, 220.19], [486.51, 220.21], [501.66, 175.14], [543.34, 175.15], [534.57, 173.61], [486.51, 173.6]]}, {\"category\": \"car\", \"corners_3d\": [[430.27, 209.03], [462.56, 209.01], [446.55, 211.92], [410.29, 211.94], [430.27, 176.06], [462.56, 176.06], [446.55, 174.94], [410.29, 174.93]]}, {\"category\": \"car\", \"corners_3d\": [[526.72, 206.76], [558.71, 206.75], [553.56, 209.4], [517.63, 209.41], [526.72, 176.27], [558.71, 176.28], [553.56, 175.18], [517.63, 175.17]]}]\n```" - } - ] - }, - { - "id": 568, - "data_source": "Mono3DRefer", - "scene_name": "000302", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[477.37, 213.14], [505.06, 214.26], [428.21, 218.04], [400.36, 216.66], [477.37, 175.65], [505.06, 175.44], [428.21, 174.74], [400.36, 175.0]]}, {\"category\": \"car\", \"corners_3d\": [[360.05, 202.53], [329.3, 202.37], [369.53, 200.2], [397.32, 200.32], [360.05, 169.17], [329.3, 169.26], [369.53, 170.54], [397.32, 170.47]]}, {\"category\": \"car\", \"corners_3d\": [[573.53, 194.58], [598.34, 194.63], [590.56, 195.87], [563.44, 195.81], [573.53, 173.07], [598.34, 173.03], [590.56, 172.23], [563.44, 172.27]]}]\n```" - } - ] - }, - { - "id": 569, - "data_source": "Mono3DRefer", - "scene_name": "000304", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[625.22, 207.44], [608.56, 207.44], [609.02, 205.33], [624.67, 205.33], [625.22, 192.66], [608.56, 192.66], [609.02, 191.45], [624.67, 191.45]]}]\n```" - } - ] - }, - { - "id": 570, - "data_source": "Mono3DRefer", - "scene_name": "000306", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[563.36, 224.3], [613.85, 224.19], [619.91, 236.46], [557.34, 236.62], [563.36, 174.08], [613.85, 174.08], [619.91, 174.37], [557.34, 174.38]]}, {\"category\": \"car\", \"corners_3d\": [[49.76, 206.79], [14.0, 206.8], [60.26, 204.13], [93.19, 204.12], [49.76, 170.04], [14.0, 170.03], [60.26, 170.26], [93.19, 170.26]]}]\n```" - } - ] - }, - { - "id": 571, - "data_source": "Mono3DRefer", - "scene_name": "000306", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[781.47, 207.8], [787.62, 207.8], [795.24, 209.24], [788.85, 209.24], [781.47, 178.1], [787.62, 178.1], [795.24, 178.32], [788.85, 178.32]]}]\n```" - } - ] - }, - { - "id": 572, - "data_source": "Mono3DRefer", - "scene_name": "000306", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[563.36, 224.3], [613.85, 224.19], [619.91, 236.46], [557.34, 236.62], [563.36, 174.08], [613.85, 174.08], [619.91, 174.37], [557.34, 174.38]]}, {\"category\": \"car\", \"corners_3d\": [[49.76, 206.79], [14.0, 206.8], [60.26, 204.13], [93.19, 204.12], [49.76, 170.04], [14.0, 170.03], [60.26, 170.26], [93.19, 170.26]]}, {\"category\": \"cyclist\", \"corners_3d\": [[781.47, 207.8], [787.62, 207.8], [795.24, 209.24], [788.85, 209.24], [781.47, 178.1], [787.62, 178.1], [795.24, 178.32], [788.85, 178.32]]}]\n```" - } - ] - }, - { - "id": 573, - "data_source": "Mono3DRefer", - "scene_name": "000307", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[365.75, 213.71], [336.79, 213.7], [357.35, 210.69], [384.18, 210.7], [365.75, 187.76], [336.79, 187.76], [357.35, 186.66], [384.18, 186.67]]}]\n```" - } - ] - }, - { - "id": 574, - "data_source": "Mono3DRefer", - "scene_name": "000307", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[600.23, 191.19], [625.48, 191.19], [629.91, 194.79], [599.7, 194.79], [600.23, 163.82], [625.48, 163.83], [629.91, 162.05], [599.7, 162.05]]}]\n```" - } - ] - }, - { - "id": 575, - "data_source": "Mono3DRefer", - "scene_name": "000307", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[300.18, 210.67], [292.15, 210.62], [303.6, 209.51], [311.4, 209.55], [300.18, 186.11], [292.15, 186.09], [303.6, 185.7], [311.4, 185.72]]}, {\"category\": \"cyclist\", \"corners_3d\": [[681.17, 197.32], [691.1, 197.32], [695.03, 198.51], [684.62, 198.51], [681.17, 166.56], [691.1, 166.56], [695.03, 166.26], [684.62, 166.26]]}]\n```" - } - ] - }, - { - "id": 576, - "data_source": "Mono3DRefer", - "scene_name": "000307", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[600.23, 191.19], [625.48, 191.19], [629.91, 194.79], [599.7, 194.79], [600.23, 163.82], [625.48, 163.83], [629.91, 162.05], [599.7, 162.05]]}, {\"category\": \"car\", \"corners_3d\": [[365.75, 213.71], [336.79, 213.7], [357.35, 210.69], [384.18, 210.7], [365.75, 187.76], [336.79, 187.76], [357.35, 186.66], [384.18, 186.67]]}]\n```" - } - ] - }, - { - "id": 577, - "data_source": "Mono3DRefer", - "scene_name": "000307", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[300.18, 210.67], [292.15, 210.62], [303.6, 209.51], [311.4, 209.55], [300.18, 186.11], [292.15, 186.09], [303.6, 185.7], [311.4, 185.72]]}, {\"category\": \"car\", \"corners_3d\": [[365.75, 213.71], [336.79, 213.7], [357.35, 210.69], [384.18, 210.7], [365.75, 187.76], [336.79, 187.76], [357.35, 186.66], [384.18, 186.67]]}, {\"category\": \"cyclist\", \"corners_3d\": [[681.17, 197.32], [691.1, 197.32], [695.03, 198.51], [684.62, 198.51], [681.17, 166.56], [691.1, 166.56], [695.03, 166.26], [684.62, 166.26]]}]\n```" - } - ] - }, - { - "id": 578, - "data_source": "Mono3DRefer", - "scene_name": "000307", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[600.23, 191.19], [625.48, 191.19], [629.91, 194.79], [599.7, 194.79], [600.23, 163.82], [625.48, 163.83], [629.91, 162.05], [599.7, 162.05]]}, {\"category\": \"cyclist\", \"corners_3d\": [[300.18, 210.67], [292.15, 210.62], [303.6, 209.51], [311.4, 209.55], [300.18, 186.11], [292.15, 186.09], [303.6, 185.7], [311.4, 185.72]]}, {\"category\": \"cyclist\", \"corners_3d\": [[681.17, 197.32], [691.1, 197.32], [695.03, 198.51], [684.62, 198.51], [681.17, 166.56], [691.1, 166.56], [695.03, 166.26], [684.62, 166.26]]}]\n```" - } - ] - }, - { - "id": 579, - "data_source": "Mono3DRefer", - "scene_name": "000307", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[600.23, 191.19], [625.48, 191.19], [629.91, 194.79], [599.7, 194.79], [600.23, 163.82], [625.48, 163.83], [629.91, 162.05], [599.7, 162.05]]}, {\"category\": \"cyclist\", \"corners_3d\": [[300.18, 210.67], [292.15, 210.62], [303.6, 209.51], [311.4, 209.55], [300.18, 186.11], [292.15, 186.09], [303.6, 185.7], [311.4, 185.72]]}, {\"category\": \"car\", \"corners_3d\": [[365.75, 213.71], [336.79, 213.7], [357.35, 210.69], [384.18, 210.7], [365.75, 187.76], [336.79, 187.76], [357.35, 186.66], [384.18, 186.67]]}, {\"category\": \"cyclist\", \"corners_3d\": [[681.17, 197.32], [691.1, 197.32], [695.03, 198.51], [684.62, 198.51], [681.17, 166.56], [691.1, 166.56], [695.03, 166.26], [684.62, 166.26]]}]\n```" - } - ] - }, - { - "id": 580, - "data_source": "Mono3DRefer", - "scene_name": "000310", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[505.63, 262.52], [546.68, 261.37], [561.09, 273.45], [514.56, 274.94], [505.63, 174.85], [546.68, 174.83], [561.09, 175.09], [514.56, 175.13]]}]\n```" - } - ] - }, - { - "id": 581, - "data_source": "Mono3DRefer", - "scene_name": "000310", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[550.96, 232.78], [610.02, 231.66], [644.16, 243.55], [573.67, 245.18], [550.96, 175.7], [610.02, 175.64], [644.16, 176.21], [573.67, 176.28]]}, {\"category\": \"car\", \"corners_3d\": [[524.72, 223.09], [573.37, 222.22], [600.12, 230.73], [543.46, 231.93], [524.72, 179.45], [573.37, 179.33], [600.12, 180.45], [543.46, 180.61]]}, {\"category\": \"car\", \"corners_3d\": [[332.47, 241.09], [275.12, 242.6], [291.05, 232.55], [340.0, 231.45], [332.47, 193.84], [275.12, 194.3], [291.05, 191.22], [340.0, 190.88]]}, {\"category\": \"car\", \"corners_3d\": [[503.27, 215.06], [538.98, 214.52], [553.91, 219.68], [513.94, 220.35], [503.27, 182.38], [538.98, 182.26], [553.91, 183.43], [513.94, 183.58]]}, {\"category\": \"car\", \"corners_3d\": [[847.1, 286.74], [983.31, 295.33], [1010.48, 374.28], [793.15, 352.07], [847.1, 178.56], [983.31, 178.99], [1010.48, 182.95], [793.15, 181.83]]}, {\"category\": \"car\", \"corners_3d\": [[614.14, 257.91], [692.97, 255.64], [789.96, 286.6], [684.23, 290.93], [614.14, 180.76], [692.97, 180.55], [789.96, 183.43], [684.23, 183.83]]}, {\"category\": \"car\", \"corners_3d\": [[223.25, 293.88], [107.08, 297.29], [201.23, 265.03], [285.96, 263.14], [223.25, 200.77], [107.08, 201.55], [201.23, 194.11], [285.96, 193.68]]}]\n```" - } - ] - }, - { - "id": 582, - "data_source": "Mono3DRefer", - "scene_name": "000310", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[296.62, 262.01], [215.72, 264.14], [261.5, 245.5], [325.34, 244.15], [296.62, 185.76], [215.72, 186.07], [261.5, 183.37], [325.34, 183.18]]}, {\"category\": \"van\", \"corners_3d\": [[430.01, 205.87], [462.28, 205.54], [465.56, 209.27], [429.6, 209.69], [430.01, 171.11], [462.28, 171.13], [465.56, 170.93], [429.6, 170.91]]}]\n```" - } - ] - }, - { - "id": 583, - "data_source": "Mono3DRefer", - "scene_name": "000310", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[550.96, 232.78], [610.02, 231.66], [644.16, 243.55], [573.67, 245.18], [550.96, 175.7], [610.02, 175.64], [644.16, 176.21], [573.67, 176.28]]}, {\"category\": \"car\", \"corners_3d\": [[524.72, 223.09], [573.37, 222.22], [600.12, 230.73], [543.46, 231.93], [524.72, 179.45], [573.37, 179.33], [600.12, 180.45], [543.46, 180.61]]}, {\"category\": \"car\", \"corners_3d\": [[332.47, 241.09], [275.12, 242.6], [291.05, 232.55], [340.0, 231.45], [332.47, 193.84], [275.12, 194.3], [291.05, 191.22], [340.0, 190.88]]}, {\"category\": \"car\", \"corners_3d\": [[503.27, 215.06], [538.98, 214.52], [553.91, 219.68], [513.94, 220.35], [503.27, 182.38], [538.98, 182.26], [553.91, 183.43], [513.94, 183.58]]}, {\"category\": \"cyclist\", \"corners_3d\": [[505.63, 262.52], [546.68, 261.37], [561.09, 273.45], [514.56, 274.94], [505.63, 174.85], [546.68, 174.83], [561.09, 175.09], [514.56, 175.13]]}, {\"category\": \"car\", \"corners_3d\": [[847.1, 286.74], [983.31, 295.33], [1010.48, 374.28], [793.15, 352.07], [847.1, 178.56], [983.31, 178.99], [1010.48, 182.95], [793.15, 181.83]]}, {\"category\": \"car\", \"corners_3d\": [[614.14, 257.91], [692.97, 255.64], [789.96, 286.6], [684.23, 290.93], [614.14, 180.76], [692.97, 180.55], [789.96, 183.43], [684.23, 183.83]]}, {\"category\": \"car\", \"corners_3d\": [[223.25, 293.88], [107.08, 297.29], [201.23, 265.03], [285.96, 263.14], [223.25, 200.77], [107.08, 201.55], [201.23, 194.11], [285.96, 193.68]]}]\n```" - } - ] - }, - { - "id": 584, - "data_source": "Mono3DRefer", - "scene_name": "000310", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[296.62, 262.01], [215.72, 264.14], [261.5, 245.5], [325.34, 244.15], [296.62, 185.76], [215.72, 186.07], [261.5, 183.37], [325.34, 183.18]]}, {\"category\": \"van\", \"corners_3d\": [[430.01, 205.87], [462.28, 205.54], [465.56, 209.27], [429.6, 209.69], [430.01, 171.11], [462.28, 171.13], [465.56, 170.93], [429.6, 170.91]]}, {\"category\": \"cyclist\", \"corners_3d\": [[505.63, 262.52], [546.68, 261.37], [561.09, 273.45], [514.56, 274.94], [505.63, 174.85], [546.68, 174.83], [561.09, 175.09], [514.56, 175.13]]}]\n```" - } - ] - }, - { - "id": 585, - "data_source": "Mono3DRefer", - "scene_name": "000310", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[296.62, 262.01], [215.72, 264.14], [261.5, 245.5], [325.34, 244.15], [296.62, 185.76], [215.72, 186.07], [261.5, 183.37], [325.34, 183.18]]}, {\"category\": \"car\", \"corners_3d\": [[550.96, 232.78], [610.02, 231.66], [644.16, 243.55], [573.67, 245.18], [550.96, 175.7], [610.02, 175.64], [644.16, 176.21], [573.67, 176.28]]}, {\"category\": \"car\", \"corners_3d\": [[524.72, 223.09], [573.37, 222.22], [600.12, 230.73], [543.46, 231.93], [524.72, 179.45], [573.37, 179.33], [600.12, 180.45], [543.46, 180.61]]}, {\"category\": \"car\", \"corners_3d\": [[332.47, 241.09], [275.12, 242.6], [291.05, 232.55], [340.0, 231.45], [332.47, 193.84], [275.12, 194.3], [291.05, 191.22], [340.0, 190.88]]}, {\"category\": \"car\", \"corners_3d\": [[503.27, 215.06], [538.98, 214.52], [553.91, 219.68], [513.94, 220.35], [503.27, 182.38], [538.98, 182.26], [553.91, 183.43], [513.94, 183.58]]}, {\"category\": \"van\", \"corners_3d\": [[430.01, 205.87], [462.28, 205.54], [465.56, 209.27], [429.6, 209.69], [430.01, 171.11], [462.28, 171.13], [465.56, 170.93], [429.6, 170.91]]}, {\"category\": \"car\", \"corners_3d\": [[847.1, 286.74], [983.31, 295.33], [1010.48, 374.28], [793.15, 352.07], [847.1, 178.56], [983.31, 178.99], [1010.48, 182.95], [793.15, 181.83]]}, {\"category\": \"car\", \"corners_3d\": [[614.14, 257.91], [692.97, 255.64], [789.96, 286.6], [684.23, 290.93], [614.14, 180.76], [692.97, 180.55], [789.96, 183.43], [684.23, 183.83]]}, {\"category\": \"car\", \"corners_3d\": [[223.25, 293.88], [107.08, 297.29], [201.23, 265.03], [285.96, 263.14], [223.25, 200.77], [107.08, 201.55], [201.23, 194.11], [285.96, 193.68]]}]\n```" - } - ] - }, - { - "id": 586, - "data_source": "Mono3DRefer", - "scene_name": "000310", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[296.62, 262.01], [215.72, 264.14], [261.5, 245.5], [325.34, 244.15], [296.62, 185.76], [215.72, 186.07], [261.5, 183.37], [325.34, 183.18]]}, {\"category\": \"car\", \"corners_3d\": [[550.96, 232.78], [610.02, 231.66], [644.16, 243.55], [573.67, 245.18], [550.96, 175.7], [610.02, 175.64], [644.16, 176.21], [573.67, 176.28]]}, {\"category\": \"car\", \"corners_3d\": [[524.72, 223.09], [573.37, 222.22], [600.12, 230.73], [543.46, 231.93], [524.72, 179.45], [573.37, 179.33], [600.12, 180.45], [543.46, 180.61]]}, {\"category\": \"car\", \"corners_3d\": [[332.47, 241.09], [275.12, 242.6], [291.05, 232.55], [340.0, 231.45], [332.47, 193.84], [275.12, 194.3], [291.05, 191.22], [340.0, 190.88]]}, {\"category\": \"car\", \"corners_3d\": [[503.27, 215.06], [538.98, 214.52], [553.91, 219.68], [513.94, 220.35], [503.27, 182.38], [538.98, 182.26], [553.91, 183.43], [513.94, 183.58]]}, {\"category\": \"van\", \"corners_3d\": [[430.01, 205.87], [462.28, 205.54], [465.56, 209.27], [429.6, 209.69], [430.01, 171.11], [462.28, 171.13], [465.56, 170.93], [429.6, 170.91]]}, {\"category\": \"cyclist\", \"corners_3d\": [[505.63, 262.52], [546.68, 261.37], [561.09, 273.45], [514.56, 274.94], [505.63, 174.85], [546.68, 174.83], [561.09, 175.09], [514.56, 175.13]]}, {\"category\": \"car\", \"corners_3d\": [[847.1, 286.74], [983.31, 295.33], [1010.48, 374.28], [793.15, 352.07], [847.1, 178.56], [983.31, 178.99], [1010.48, 182.95], [793.15, 181.83]]}, {\"category\": \"car\", \"corners_3d\": [[614.14, 257.91], [692.97, 255.64], [789.96, 286.6], [684.23, 290.93], [614.14, 180.76], [692.97, 180.55], [789.96, 183.43], [684.23, 183.83]]}, {\"category\": \"car\", \"corners_3d\": [[223.25, 293.88], [107.08, 297.29], [201.23, 265.03], [285.96, 263.14], [223.25, 200.77], [107.08, 201.55], [201.23, 194.11], [285.96, 193.68]]}]\n```" - } - ] - }, - { - "id": 587, - "data_source": "Mono3DRefer", - "scene_name": "000311", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[379.0, 277.13], [487.85, 276.68], [417.35, 349.41], [231.14, 350.73], [379.0, 177.57], [487.85, 177.55], [417.35, 180.84], [231.14, 180.9]]}, {\"category\": \"car\", \"corners_3d\": [[765.44, 323.42], [927.97, 324.12], [1249.92, 484.64], [913.5, 481.67], [765.44, 168.05], [927.97, 168.03], [1249.92, 162.91], [913.5, 163.01]]}, {\"category\": \"car\", \"corners_3d\": [[464.81, 251.63], [541.96, 252.24], [499.2, 280.82], [394.99, 279.7], [464.81, 183.96], [541.96, 184.05], [499.2, 188.08], [394.99, 187.92]]}, {\"category\": \"car\", \"corners_3d\": [[662.58, 216.69], [703.15, 216.66], [719.89, 223.98], [672.55, 224.01], [662.58, 179.78], [703.15, 179.78], [719.89, 180.93], [672.55, 180.94]]}, {\"category\": \"car\", \"corners_3d\": [[713.9, 255.81], [799.07, 255.81], [880.18, 291.43], [758.44, 291.41], [713.9, 178.84], [799.07, 178.84], [880.18, 181.41], [758.44, 181.41]]}, {\"category\": \"car\", \"corners_3d\": [[557.21, 203.77], [586.95, 203.73], [586.8, 206.84], [554.06, 206.88], [557.21, 176.59], [586.95, 176.58], [586.8, 176.96], [554.06, 176.96]]}, {\"category\": \"car\", \"corners_3d\": [[648.98, 205.45], [677.51, 205.42], [686.32, 208.93], [654.72, 208.96], [648.98, 177.73], [677.51, 177.73], [686.32, 178.25], [654.72, 178.25]]}]\n```" - } - ] - }, - { - "id": 588, - "data_source": "Mono3DRefer", - "scene_name": "000311", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[407.36, 249.71], [434.26, 249.74], [429.7, 251.65], [402.13, 251.62], [407.36, 167.44], [434.26, 167.44], [429.7, 167.31], [402.13, 167.31]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[428.26, 238.05], [441.07, 238.05], [438.2, 239.16], [425.17, 239.16], [428.26, 169.83], [441.07, 169.83], [438.2, 169.77], [425.17, 169.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[433.92, 240.07], [449.46, 240.07], [445.65, 241.66], [429.74, 241.66], [433.92, 170.9], [449.46, 170.9], [445.65, 170.85], [429.74, 170.85]]}]\n```" - } - ] - }, - { - "id": 589, - "data_source": "Mono3DRefer", - "scene_name": "000311", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[379.0, 277.13], [487.85, 276.68], [417.35, 349.41], [231.14, 350.73], [379.0, 177.57], [487.85, 177.55], [417.35, 180.84], [231.14, 180.9]]}, {\"category\": \"car\", \"corners_3d\": [[765.44, 323.42], [927.97, 324.12], [1249.92, 484.64], [913.5, 481.67], [765.44, 168.05], [927.97, 168.03], [1249.92, 162.91], [913.5, 163.01]]}, {\"category\": \"car\", \"corners_3d\": [[464.81, 251.63], [541.96, 252.24], [499.2, 280.82], [394.99, 279.7], [464.81, 183.96], [541.96, 184.05], [499.2, 188.08], [394.99, 187.92]]}, {\"category\": \"car\", \"corners_3d\": [[662.58, 216.69], [703.15, 216.66], [719.89, 223.98], [672.55, 224.01], [662.58, 179.78], [703.15, 179.78], [719.89, 180.93], [672.55, 180.94]]}, {\"category\": \"car\", \"corners_3d\": [[713.9, 255.81], [799.07, 255.81], [880.18, 291.43], [758.44, 291.41], [713.9, 178.84], [799.07, 178.84], [880.18, 181.41], [758.44, 181.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[407.36, 249.71], [434.26, 249.74], [429.7, 251.65], [402.13, 251.62], [407.36, 167.44], [434.26, 167.44], [429.7, 167.31], [402.13, 167.31]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[428.26, 238.05], [441.07, 238.05], [438.2, 239.16], [425.17, 239.16], [428.26, 169.83], [441.07, 169.83], [438.2, 169.77], [425.17, 169.77]]}, {\"category\": \"car\", \"corners_3d\": [[557.21, 203.77], [586.95, 203.73], [586.8, 206.84], [554.06, 206.88], [557.21, 176.59], [586.95, 176.58], [586.8, 176.96], [554.06, 176.96]]}, {\"category\": \"car\", \"corners_3d\": [[648.98, 205.45], [677.51, 205.42], [686.32, 208.93], [654.72, 208.96], [648.98, 177.73], [677.51, 177.73], [686.32, 178.25], [654.72, 178.25]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[433.92, 240.07], [449.46, 240.07], [445.65, 241.66], [429.74, 241.66], [433.92, 170.9], [449.46, 170.9], [445.65, 170.85], [429.74, 170.85]]}]\n```" - } - ] - }, - { - "id": 590, - "data_source": "Mono3DRefer", - "scene_name": "000313", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[426.51, 226.64], [410.19, 230.52], [275.09, 229.86], [300.22, 226.07], [426.51, 187.31], [410.19, 188.35], [275.09, 188.17], [300.22, 187.15]]}]\n```" - } - ] - }, - { - "id": 591, - "data_source": "Mono3DRefer", - "scene_name": "000314", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1023.81, 168.55], [1066.21, 168.44], [1071.06, 167.94], [1023.85, 168.08], [1023.81, 124.29], [1066.21, 123.05], [1071.06, 117.4], [1023.85, 118.92]]}]\n```" - } - ] - }, - { - "id": 592, - "data_source": "Mono3DRefer", - "scene_name": "000314", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[108.29, 326.43], [-51.11, 339.65], [24.53, 293.79], [138.23, 286.68], [108.29, 215.84], [-51.11, 219.54], [24.53, 206.7], [138.23, 204.71]]}, {\"category\": \"car\", \"corners_3d\": [[231.1, 240.83], [193.27, 242.26], [197.61, 236.13], [232.07, 234.94], [231.1, 210.83], [193.27, 211.63], [197.61, 208.21], [232.07, 207.54]]}, {\"category\": \"car\", \"corners_3d\": [[187.46, 281.12], [99.43, 285.81], [131.11, 266.69], [203.67, 263.44], [187.46, 213.85], [99.43, 215.62], [131.11, 208.38], [203.67, 207.15]]}]\n```" - } - ] - }, - { - "id": 593, - "data_source": "Mono3DRefer", - "scene_name": "000314", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[108.29, 326.43], [-51.11, 339.65], [24.53, 293.79], [138.23, 286.68], [108.29, 215.84], [-51.11, 219.54], [24.53, 206.7], [138.23, 204.71]]}, {\"category\": \"car\", \"corners_3d\": [[231.1, 240.83], [193.27, 242.26], [197.61, 236.13], [232.07, 234.94], [231.1, 210.83], [193.27, 211.63], [197.61, 208.21], [232.07, 207.54]]}, {\"category\": \"car\", \"corners_3d\": [[187.46, 281.12], [99.43, 285.81], [131.11, 266.69], [203.67, 263.44], [187.46, 213.85], [99.43, 215.62], [131.11, 208.38], [203.67, 207.15]]}, {\"category\": \"van\", \"corners_3d\": [[1023.81, 168.55], [1066.21, 168.44], [1071.06, 167.94], [1023.85, 168.08], [1023.81, 124.29], [1066.21, 123.05], [1071.06, 117.4], [1023.85, 118.92]]}]\n```" - } - ] - }, - { - "id": 594, - "data_source": "Mono3DRefer", - "scene_name": "000315", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[653.75, 197.55], [631.1, 197.35], [641.5, 195.78], [662.77, 195.95], [653.75, 177.08], [631.1, 177.05], [641.5, 176.78], [662.77, 176.81]]}, {\"category\": \"car\", \"corners_3d\": [[724.23, 270.04], [817.41, 271.05], [858.12, 297.8], [739.35, 296.18], [724.23, 178.1], [817.41, 178.16], [858.12, 179.6], [739.35, 179.52]]}, {\"category\": \"car\", \"corners_3d\": [[429.39, 321.21], [309.07, 318.15], [426.53, 274.04], [511.5, 275.51], [429.39, 197.0], [309.07, 196.5], [426.53, 189.32], [511.5, 189.56]]}]\n```" - } - ] - }, - { - "id": 595, - "data_source": "Mono3DRefer", - "scene_name": "000316", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[788.2, 241.0], [819.77, 241.0], [829.22, 244.07], [796.22, 244.07], [788.2, 157.74], [819.77, 157.74], [829.22, 157.06], [796.22, 157.06]]}]\n```" - } - ] - }, - { - "id": 596, - "data_source": "Mono3DRefer", - "scene_name": "000316", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[589.69, 197.1], [612.15, 197.1], [612.84, 198.92], [588.69, 198.93], [589.69, 175.88], [612.15, 175.88], [612.84, 176.11], [588.69, 176.11]]}, {\"category\": \"car\", \"corners_3d\": [[525.49, 237.46], [466.43, 237.47], [488.66, 227.39], [538.51, 227.39], [525.49, 183.61], [466.43, 183.61], [488.66, 181.93], [538.51, 181.93]]}, {\"category\": \"car\", \"corners_3d\": [[815.81, 322.74], [951.58, 321.91], [1283.57, 458.19], [1025.88, 461.26], [815.81, 184.06], [951.58, 184.0], [1283.57, 194.2], [1025.88, 194.42]]}, {\"category\": \"car\", \"corners_3d\": [[720.02, 271.95], [815.36, 271.56], [919.6, 316.92], [780.8, 317.74], [720.02, 175.77], [815.36, 175.75], [919.6, 177.09], [780.8, 177.11]]}, {\"category\": \"car\", \"corners_3d\": [[696.41, 251.12], [759.72, 250.64], [800.33, 266.43], [724.38, 267.12], [696.41, 172.84], [759.72, 172.84], [800.33, 172.83], [724.38, 172.83]]}, {\"category\": \"car\", \"corners_3d\": [[657.69, 220.1], [701.89, 219.98], [722.58, 228.07], [670.83, 228.22], [657.69, 178.16], [701.89, 178.14], [722.58, 179.05], [670.83, 179.07]]}, {\"category\": \"car\", \"corners_3d\": [[637.11, 203.42], [664.56, 203.44], [669.73, 206.78], [639.27, 206.77], [637.11, 177.39], [664.56, 177.39], [669.73, 177.89], [639.27, 177.89]]}, {\"category\": \"car\", \"corners_3d\": [[648.49, 213.46], [685.56, 213.44], [696.97, 219.04], [654.78, 219.07], [648.49, 176.81], [685.56, 176.81], [696.97, 177.36], [654.78, 177.36]]}, {\"category\": \"car\", \"corners_3d\": [[437.56, 207.47], [432.81, 208.68], [365.8, 208.85], [372.85, 207.63], [437.56, 182.88], [432.81, 183.23], [365.8, 183.28], [372.85, 182.92]]}, {\"category\": \"car\", \"corners_3d\": [[349.57, 208.74], [340.11, 210.04], [268.75, 210.04], [280.72, 208.74], [349.57, 183.08], [340.11, 183.45], [268.75, 183.45], [280.72, 183.08]]}, {\"category\": \"car\", \"corners_3d\": [[626.35, 194.25], [645.04, 194.23], [649.52, 195.73], [629.52, 195.76], [626.35, 176.24], [645.04, 176.23], [649.52, 176.47], [629.52, 176.47]]}]\n```" - } - ] - }, - { - "id": 597, - "data_source": "Mono3DRefer", - "scene_name": "000316", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[589.69, 197.1], [612.15, 197.1], [612.84, 198.92], [588.69, 198.93], [589.69, 175.88], [612.15, 175.88], [612.84, 176.11], [588.69, 176.11]]}, {\"category\": \"car\", \"corners_3d\": [[525.49, 237.46], [466.43, 237.47], [488.66, 227.39], [538.51, 227.39], [525.49, 183.61], [466.43, 183.61], [488.66, 181.93], [538.51, 181.93]]}, {\"category\": \"car\", \"corners_3d\": [[815.81, 322.74], [951.58, 321.91], [1283.57, 458.19], [1025.88, 461.26], [815.81, 184.06], [951.58, 184.0], [1283.57, 194.2], [1025.88, 194.42]]}, {\"category\": \"car\", \"corners_3d\": [[720.02, 271.95], [815.36, 271.56], [919.6, 316.92], [780.8, 317.74], [720.02, 175.77], [815.36, 175.75], [919.6, 177.09], [780.8, 177.11]]}, {\"category\": \"car\", \"corners_3d\": [[696.41, 251.12], [759.72, 250.64], [800.33, 266.43], [724.38, 267.12], [696.41, 172.84], [759.72, 172.84], [800.33, 172.83], [724.38, 172.83]]}, {\"category\": \"car\", \"corners_3d\": [[657.69, 220.1], [701.89, 219.98], [722.58, 228.07], [670.83, 228.22], [657.69, 178.16], [701.89, 178.14], [722.58, 179.05], [670.83, 179.07]]}, {\"category\": \"car\", \"corners_3d\": [[637.11, 203.42], [664.56, 203.44], [669.73, 206.78], [639.27, 206.77], [637.11, 177.39], [664.56, 177.39], [669.73, 177.89], [639.27, 177.89]]}, {\"category\": \"car\", \"corners_3d\": [[648.49, 213.46], [685.56, 213.44], [696.97, 219.04], [654.78, 219.07], [648.49, 176.81], [685.56, 176.81], [696.97, 177.36], [654.78, 177.36]]}, {\"category\": \"car\", \"corners_3d\": [[437.56, 207.47], [432.81, 208.68], [365.8, 208.85], [372.85, 207.63], [437.56, 182.88], [432.81, 183.23], [365.8, 183.28], [372.85, 182.92]]}, {\"category\": \"car\", \"corners_3d\": [[349.57, 208.74], [340.11, 210.04], [268.75, 210.04], [280.72, 208.74], [349.57, 183.08], [340.11, 183.45], [268.75, 183.45], [280.72, 183.08]]}, {\"category\": \"car\", \"corners_3d\": [[626.35, 194.25], [645.04, 194.23], [649.52, 195.73], [629.52, 195.76], [626.35, 176.24], [645.04, 176.23], [649.52, 176.47], [629.52, 176.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[788.2, 241.0], [819.77, 241.0], [829.22, 244.07], [796.22, 244.07], [788.2, 157.74], [819.77, 157.74], [829.22, 157.06], [796.22, 157.06]]}]\n```" - } - ] - }, - { - "id": 598, - "data_source": "Mono3DRefer", - "scene_name": "000317", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[474.75, 222.57], [469.73, 221.63], [488.84, 221.12], [494.23, 222.05], [474.75, 177.02], [469.73, 177.09], [488.84, 177.14], [494.23, 177.06]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[460.95, 219.91], [453.68, 219.15], [471.65, 218.53], [479.14, 219.27], [460.95, 181.65], [453.68, 181.63], [471.65, 181.61], [479.14, 181.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[712.36, 218.34], [714.19, 219.21], [698.54, 219.17], [697.05, 218.31], [712.36, 179.82], [714.19, 179.8], [698.54, 179.8], [697.05, 179.82]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[722.54, 227.7], [720.58, 226.58], [739.85, 226.64], [742.28, 227.77], [722.54, 188.11], [720.58, 187.93], [739.85, 187.94], [742.28, 188.13]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[335.05, 224.28], [316.68, 224.82], [314.88, 224.14], [333.0, 223.62], [335.05, 175.48], [316.68, 175.42], [314.88, 175.5], [333.0, 175.56]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[327.87, 221.36], [342.4, 220.8], [350.34, 221.66], [335.6, 222.24], [327.87, 179.7], [342.4, 179.72], [350.34, 179.7], [335.6, 179.69]]}]\n```" - } - ] - }, - { - "id": 599, - "data_source": "Mono3DRefer", - "scene_name": "000317", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[901.5, 261.94], [932.54, 263.91], [916.16, 271.62], [882.75, 269.28], [901.5, 182.19], [932.54, 182.23], [916.16, 182.39], [882.75, 182.34]]}, {\"category\": \"cyclist\", \"corners_3d\": [[355.05, 498.78], [266.56, 481.55], [441.49, 394.22], [510.28, 402.76], [355.05, 172.49], [266.56, 172.93], [441.49, 175.12], [510.28, 174.91]]}, {\"category\": \"cyclist\", \"corners_3d\": [[732.6, 262.18], [706.04, 261.01], [722.99, 253.75], [747.34, 254.72], [732.6, 185.4], [706.04, 185.33], [722.99, 184.89], [747.34, 184.95]]}, {\"category\": \"cyclist\", \"corners_3d\": [[691.81, 219.07], [702.06, 219.65], [678.1, 220.91], [667.89, 220.3], [691.81, 182.85], [702.06, 182.88], [678.1, 182.96], [667.89, 182.92]]}]\n```" - } - ] - }, - { - "id": 600, - "data_source": "Mono3DRefer", - "scene_name": "000317", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[474.75, 222.57], [469.73, 221.63], [488.84, 221.12], [494.23, 222.05], [474.75, 177.02], [469.73, 177.09], [488.84, 177.14], [494.23, 177.06]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[460.95, 219.91], [453.68, 219.15], [471.65, 218.53], [479.14, 219.27], [460.95, 181.65], [453.68, 181.63], [471.65, 181.61], [479.14, 181.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[712.36, 218.34], [714.19, 219.21], [698.54, 219.17], [697.05, 218.31], [712.36, 179.82], [714.19, 179.8], [698.54, 179.8], [697.05, 179.82]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[722.54, 227.7], [720.58, 226.58], [739.85, 226.64], [742.28, 227.77], [722.54, 188.11], [720.58, 187.93], [739.85, 187.94], [742.28, 188.13]]}, {\"category\": \"cyclist\", \"corners_3d\": [[901.5, 261.94], [932.54, 263.91], [916.16, 271.62], [882.75, 269.28], [901.5, 182.19], [932.54, 182.23], [916.16, 182.39], [882.75, 182.34]]}, {\"category\": \"cyclist\", \"corners_3d\": [[355.05, 498.78], [266.56, 481.55], [441.49, 394.22], [510.28, 402.76], [355.05, 172.49], [266.56, 172.93], [441.49, 175.12], [510.28, 174.91]]}, {\"category\": \"cyclist\", \"corners_3d\": [[732.6, 262.18], [706.04, 261.01], [722.99, 253.75], [747.34, 254.72], [732.6, 185.4], [706.04, 185.33], [722.99, 184.89], [747.34, 184.95]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[335.05, 224.28], [316.68, 224.82], [314.88, 224.14], [333.0, 223.62], [335.05, 175.48], [316.68, 175.42], [314.88, 175.5], [333.0, 175.56]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[327.87, 221.36], [342.4, 220.8], [350.34, 221.66], [335.6, 222.24], [327.87, 179.7], [342.4, 179.72], [350.34, 179.7], [335.6, 179.69]]}, {\"category\": \"cyclist\", \"corners_3d\": [[691.81, 219.07], [702.06, 219.65], [678.1, 220.91], [667.89, 220.3], [691.81, 182.85], [702.06, 182.88], [678.1, 182.96], [667.89, 182.92]]}]\n```" - } - ] - }, - { - "id": 601, - "data_source": "Mono3DRefer", - "scene_name": "000319", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[600.17, 202.08], [625.72, 202.08], [627.03, 204.52], [599.33, 204.52], [600.17, 176.83], [625.72, 176.83], [627.03, 177.16], [599.33, 177.16]]}, {\"category\": \"car\", \"corners_3d\": [[233.12, 245.45], [179.08, 245.61], [246.5, 233.6], [291.5, 233.48], [233.12, 197.16], [179.08, 197.21], [246.5, 193.19], [291.5, 193.15]]}, {\"category\": \"car\", \"corners_3d\": [[295.88, 249.61], [237.27, 249.61], [300.07, 236.64], [348.77, 236.64], [295.88, 193.85], [237.27, 193.85], [300.07, 190.3], [348.77, 190.3]]}, {\"category\": \"car\", \"corners_3d\": [[476.47, 207.73], [449.96, 207.79], [461.23, 204.76], [485.43, 204.72], [476.47, 183.5], [449.96, 183.52], [461.23, 182.59], [485.43, 182.58]]}]\n```" - } - ] - }, - { - "id": 602, - "data_source": "Mono3DRefer", - "scene_name": "000321", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[550.94, 195.9], [528.88, 195.92], [533.6, 194.26], [554.07, 194.25], [550.94, 175.23], [528.88, 175.23], [533.6, 175.06], [554.07, 175.06]]}, {\"category\": \"car\", \"corners_3d\": [[325.21, 371.13], [133.48, 371.7], [311.96, 295.92], [430.43, 295.7], [325.21, 194.35], [133.48, 194.41], [311.96, 186.19], [430.43, 186.17]]}, {\"category\": \"car\", \"corners_3d\": [[506.55, 232.55], [450.96, 232.56], [476.94, 222.74], [523.39, 222.74], [506.55, 187.13], [450.96, 187.14], [476.94, 184.79], [523.39, 184.79]]}]\n```" - } - ] - }, - { - "id": 603, - "data_source": "Mono3DRefer", - "scene_name": "000321", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[521.63, 214.48], [483.36, 214.5], [496.51, 209.88], [530.53, 209.86], [521.63, 174.9], [483.36, 174.91], [496.51, 174.68], [530.53, 174.68]]}]\n```" - } - ] - }, - { - "id": 604, - "data_source": "Mono3DRefer", - "scene_name": "000321", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[550.94, 195.9], [528.88, 195.92], [533.6, 194.26], [554.07, 194.25], [550.94, 175.23], [528.88, 175.23], [533.6, 175.06], [554.07, 175.06]]}, {\"category\": \"car\", \"corners_3d\": [[325.21, 371.13], [133.48, 371.7], [311.96, 295.92], [430.43, 295.7], [325.21, 194.35], [133.48, 194.41], [311.96, 186.19], [430.43, 186.17]]}, {\"category\": \"car\", \"corners_3d\": [[506.55, 232.55], [450.96, 232.56], [476.94, 222.74], [523.39, 222.74], [506.55, 187.13], [450.96, 187.14], [476.94, 184.79], [523.39, 184.79]]}, {\"category\": \"van\", \"corners_3d\": [[521.63, 214.48], [483.36, 214.5], [496.51, 209.88], [530.53, 209.86], [521.63, 174.9], [483.36, 174.91], [496.51, 174.68], [530.53, 174.68]]}]\n```" - } - ] - }, - { - "id": 605, - "data_source": "Mono3DRefer", - "scene_name": "000323", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[4.72, 334.5], [131.03, 311.24], [330.15, 330.52], [221.84, 361.44], [4.72, 213.86], [131.03, 207.96], [330.15, 212.85], [221.84, 220.69]]}, {\"category\": \"car\", \"corners_3d\": [[732.41, 236.39], [803.42, 238.3], [793.5, 255.42], [704.95, 252.41], [732.41, 173.62], [803.42, 173.65], [793.5, 173.86], [704.95, 173.82]]}, {\"category\": \"car\", \"corners_3d\": [[961.45, 235.94], [1044.85, 237.35], [1142.12, 259.69], [1027.84, 257.15], [961.45, 162.7], [1044.85, 162.48], [1142.12, 158.88], [1027.84, 159.29]]}, {\"category\": \"car\", \"corners_3d\": [[933.01, 215.44], [998.89, 216.53], [1038.18, 227.42], [955.15, 225.73], [933.01, 166.29], [998.89, 166.12], [1038.18, 164.44], [955.15, 164.7]]}, {\"category\": \"car\", \"corners_3d\": [[776.32, 206.36], [813.74, 206.89], [809.83, 210.64], [768.42, 209.98], [776.32, 172.61], [813.74, 172.61], [809.83, 172.58], [768.42, 172.59]]}, {\"category\": \"car\", \"corners_3d\": [[905.84, 199.13], [949.22, 199.68], [957.03, 203.8], [906.95, 203.08], [905.84, 158.02], [949.22, 157.71], [957.03, 155.38], [906.95, 155.79]]}]\n```" - } - ] - }, - { - "id": 606, - "data_source": "Mono3DRefer", - "scene_name": "000326", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[688.1, 241.59], [750.03, 241.18], [808.47, 262.15], [727.86, 262.85], [688.1, 184.09], [750.03, 184.02], [808.47, 187.45], [727.86, 187.56]]}, {\"category\": \"car\", \"corners_3d\": [[913.36, 339.2], [1087.78, 337.18], [1662.3, 520.94], [1302.85, 530.14], [913.36, 165.2], [1087.78, 165.3], [1662.3, 156.84], [1302.85, 156.42]]}]\n```" - } - ] - }, - { - "id": 607, - "data_source": "Mono3DRefer", - "scene_name": "000327", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[788.74, 300.74], [777.09, 293.48], [857.43, 293.33], [874.22, 300.57], [788.74, 143.32], [777.09, 145.56], [857.43, 145.61], [874.22, 143.37]]}]\n```" - } - ] - }, - { - "id": 608, - "data_source": "Mono3DRefer", - "scene_name": "000329", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[601.33, 204.88], [626.68, 204.87], [628.65, 207.53], [601.2, 207.54], [601.33, 179.83], [626.68, 179.83], [628.65, 180.41], [601.2, 180.41]]}, {\"category\": \"car\", \"corners_3d\": [[518.7, 201.61], [501.35, 201.6], [507.88, 199.97], [524.25, 199.98], [518.7, 185.07], [501.35, 185.07], [507.88, 184.38], [524.25, 184.38]]}]\n```" - } - ] - }, - { - "id": 609, - "data_source": "Mono3DRefer", - "scene_name": "000329", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[896.94, 251.84], [925.33, 251.85], [936.14, 254.55], [906.78, 254.55], [896.94, 162.82], [925.33, 162.82], [936.14, 162.47], [906.78, 162.47]]}]\n```" - } - ] - }, - { - "id": 610, - "data_source": "Mono3DRefer", - "scene_name": "000329", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[601.33, 204.88], [626.68, 204.87], [628.65, 207.53], [601.2, 207.54], [601.33, 179.83], [626.68, 179.83], [628.65, 180.41], [601.2, 180.41]]}, {\"category\": \"car\", \"corners_3d\": [[518.7, 201.61], [501.35, 201.6], [507.88, 199.97], [524.25, 199.98], [518.7, 185.07], [501.35, 185.07], [507.88, 184.38], [524.25, 184.38]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[896.94, 251.84], [925.33, 251.85], [936.14, 254.55], [906.78, 254.55], [896.94, 162.82], [925.33, 162.82], [936.14, 162.47], [906.78, 162.47]]}]\n```" - } - ] - }, - { - "id": 611, - "data_source": "Mono3DRefer", - "scene_name": "000334", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[500.62, 197.94], [477.91, 198.01], [482.88, 196.11], [503.87, 196.05], [500.62, 172.44], [477.91, 172.44], [482.88, 172.47], [503.87, 172.47]]}]\n```" - } - ] - }, - { - "id": 612, - "data_source": "Mono3DRefer", - "scene_name": "000334", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[689.47, 200.27], [705.86, 200.1], [723.83, 201.85], [706.49, 202.04], [689.47, 181.22], [705.86, 181.17], [723.83, 181.7], [706.49, 181.76]]}]\n```" - } - ] - }, - { - "id": 613, - "data_source": "Mono3DRefer", - "scene_name": "000334", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[689.47, 200.27], [705.86, 200.1], [723.83, 201.85], [706.49, 202.04], [689.47, 181.22], [705.86, 181.17], [723.83, 181.7], [706.49, 181.76]]}, {\"category\": \"van\", \"corners_3d\": [[500.62, 197.94], [477.91, 198.01], [482.88, 196.11], [503.87, 196.05], [500.62, 172.44], [477.91, 172.44], [482.88, 172.47], [503.87, 172.47]]}]\n```" - } - ] - }, - { - "id": 614, - "data_source": "Mono3DRefer", - "scene_name": "000335", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[523.47, 242.69], [456.94, 242.5], [489.6, 229.41], [543.67, 229.53], [523.47, 172.84], [456.94, 172.84], [489.6, 172.84], [543.67, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[754.71, 315.37], [890.68, 315.92], [1000.18, 374.79], [807.96, 373.69], [754.71, 180.46], [890.68, 180.49], [1000.18, 183.63], [807.96, 183.57]]}, {\"category\": \"car\", \"corners_3d\": [[707.91, 260.63], [792.79, 261.35], [833.57, 288.69], [722.32, 287.45], [707.91, 188.61], [792.79, 188.74], [833.57, 193.64], [722.32, 193.42]]}, {\"category\": \"car\", \"corners_3d\": [[528.26, 215.58], [569.05, 215.73], [556.79, 221.95], [510.14, 221.76], [528.26, 181.2], [569.05, 181.22], [556.79, 182.44], [510.14, 182.4]]}, {\"category\": \"car\", \"corners_3d\": [[670.86, 203.53], [643.57, 203.53], [640.78, 201.05], [665.87, 201.05], [670.86, 178.27], [643.57, 178.27], [640.78, 177.84], [665.87, 177.83]]}, {\"category\": \"car\", \"corners_3d\": [[73.19, 259.69], [135.83, 249.71], [360.44, 249.99], [326.84, 260.05], [73.19, 168.23], [135.83, 168.76], [360.44, 168.75], [326.84, 168.21]]}]\n```" - } - ] - }, - { - "id": 615, - "data_source": "Mono3DRefer", - "scene_name": "000335", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[598.01, 237.51], [577.58, 237.49], [579.81, 233.76], [599.07, 233.77], [598.01, 170.24], [577.58, 170.24], [579.81, 170.39], [599.07, 170.39]]}]\n```" - } - ] - }, - { - "id": 616, - "data_source": "Mono3DRefer", - "scene_name": "000335", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[523.47, 242.69], [456.94, 242.5], [489.6, 229.41], [543.67, 229.53], [523.47, 172.84], [456.94, 172.84], [489.6, 172.84], [543.67, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[754.71, 315.37], [890.68, 315.92], [1000.18, 374.79], [807.96, 373.69], [754.71, 180.46], [890.68, 180.49], [1000.18, 183.63], [807.96, 183.57]]}, {\"category\": \"cyclist\", \"corners_3d\": [[598.01, 237.51], [577.58, 237.49], [579.81, 233.76], [599.07, 233.77], [598.01, 170.24], [577.58, 170.24], [579.81, 170.39], [599.07, 170.39]]}, {\"category\": \"car\", \"corners_3d\": [[707.91, 260.63], [792.79, 261.35], [833.57, 288.69], [722.32, 287.45], [707.91, 188.61], [792.79, 188.74], [833.57, 193.64], [722.32, 193.42]]}, {\"category\": \"car\", \"corners_3d\": [[528.26, 215.58], [569.05, 215.73], [556.79, 221.95], [510.14, 221.76], [528.26, 181.2], [569.05, 181.22], [556.79, 182.44], [510.14, 182.4]]}, {\"category\": \"car\", \"corners_3d\": [[670.86, 203.53], [643.57, 203.53], [640.78, 201.05], [665.87, 201.05], [670.86, 178.27], [643.57, 178.27], [640.78, 177.84], [665.87, 177.83]]}, {\"category\": \"car\", \"corners_3d\": [[73.19, 259.69], [135.83, 249.71], [360.44, 249.99], [326.84, 260.05], [73.19, 168.23], [135.83, 168.76], [360.44, 168.75], [326.84, 168.21]]}]\n```" - } - ] - }, - { - "id": 617, - "data_source": "Mono3DRefer", - "scene_name": "000339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[405.16, 241.22], [346.79, 241.71], [378.78, 230.54], [427.54, 230.19], [405.16, 187.87], [346.79, 187.98], [378.78, 185.52], [427.54, 185.45]]}, {\"category\": \"car\", \"corners_3d\": [[634.42, 227.11], [686.02, 226.81], [712.58, 237.52], [650.85, 237.97], [634.42, 179.76], [686.02, 179.72], [712.58, 181.08], [650.85, 181.14]]}, {\"category\": \"car\", \"corners_3d\": [[-465.75, 603.51], [-870.69, 603.71], [-22.1, 356.61], [150.49, 356.58], [-465.75, 242.55], [-870.69, 242.58], [-22.1, 202.59], [150.49, 202.59]]}, {\"category\": \"car\", \"corners_3d\": [[724.69, 281.67], [840.97, 280.98], [974.53, 336.58], [799.17, 338.17], [724.69, 170.75], [840.97, 170.76], [974.53, 169.69], [799.17, 169.66]]}, {\"category\": \"car\", \"corners_3d\": [[662.11, 245.8], [739.9, 245.1], [802.81, 268.42], [700.39, 269.66], [662.11, 176.05], [739.9, 176.02], [802.81, 177.04], [700.39, 177.1]]}, {\"category\": \"car\", \"corners_3d\": [[355.47, 272.39], [266.01, 273.48], [327.56, 251.19], [396.86, 250.53], [355.47, 194.9], [266.01, 195.14], [327.56, 190.2], [396.86, 190.06]]}]\n```" - } - ] - }, - { - "id": 618, - "data_source": "Mono3DRefer", - "scene_name": "000339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[558.12, 198.81], [587.11, 198.75], [589.3, 201.54], [557.18, 201.62], [558.12, 166.95], [587.11, 166.97], [589.3, 166.33], [557.18, 166.32]]}, {\"category\": \"van\", \"corners_3d\": [[207.53, 337.61], [51.38, 339.03], [235.28, 281.23], [336.39, 280.62], [207.53, 180.8], [51.38, 180.86], [235.28, 178.08], [336.39, 178.05]]}]\n```" - } - ] - }, - { - "id": 619, - "data_source": "Mono3DRefer", - "scene_name": "000339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[570.51, 257.41], [611.93, 257.08], [619.4, 269.09], [572.08, 269.53], [570.51, 169.29], [611.93, 169.3], [619.4, 168.8], [572.08, 168.78]]}]\n```" - } - ] - }, - { - "id": 620, - "data_source": "Mono3DRefer", - "scene_name": "000339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1188.2, 312.65], [1114.4, 313.41], [1057.56, 298.65], [1123.92, 298.05], [1188.2, 150.79], [1114.4, 150.67], [1057.56, 153.0], [1123.92, 153.1]]}]\n```" - } - ] - }, - { - "id": 621, - "data_source": "Mono3DRefer", - "scene_name": "000339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[405.16, 241.22], [346.79, 241.71], [378.78, 230.54], [427.54, 230.19], [405.16, 187.87], [346.79, 187.98], [378.78, 185.52], [427.54, 185.45]]}, {\"category\": \"car\", \"corners_3d\": [[634.42, 227.11], [686.02, 226.81], [712.58, 237.52], [650.85, 237.97], [634.42, 179.76], [686.02, 179.72], [712.58, 181.08], [650.85, 181.14]]}, {\"category\": \"van\", \"corners_3d\": [[558.12, 198.81], [587.11, 198.75], [589.3, 201.54], [557.18, 201.62], [558.12, 166.95], [587.11, 166.97], [589.3, 166.33], [557.18, 166.32]]}, {\"category\": \"car\", \"corners_3d\": [[-465.75, 603.51], [-870.69, 603.71], [-22.1, 356.61], [150.49, 356.58], [-465.75, 242.55], [-870.69, 242.58], [-22.1, 202.59], [150.49, 202.59]]}, {\"category\": \"van\", \"corners_3d\": [[207.53, 337.61], [51.38, 339.03], [235.28, 281.23], [336.39, 280.62], [207.53, 180.8], [51.38, 180.86], [235.28, 178.08], [336.39, 178.05]]}, {\"category\": \"car\", \"corners_3d\": [[724.69, 281.67], [840.97, 280.98], [974.53, 336.58], [799.17, 338.17], [724.69, 170.75], [840.97, 170.76], [974.53, 169.69], [799.17, 169.66]]}, {\"category\": \"car\", \"corners_3d\": [[662.11, 245.8], [739.9, 245.1], [802.81, 268.42], [700.39, 269.66], [662.11, 176.05], [739.9, 176.02], [802.81, 177.04], [700.39, 177.1]]}, {\"category\": \"car\", \"corners_3d\": [[355.47, 272.39], [266.01, 273.48], [327.56, 251.19], [396.86, 250.53], [355.47, 194.9], [266.01, 195.14], [327.56, 190.2], [396.86, 190.06]]}]\n```" - } - ] - }, - { - "id": 622, - "data_source": "Mono3DRefer", - "scene_name": "000339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[405.16, 241.22], [346.79, 241.71], [378.78, 230.54], [427.54, 230.19], [405.16, 187.87], [346.79, 187.98], [378.78, 185.52], [427.54, 185.45]]}, {\"category\": \"car\", \"corners_3d\": [[634.42, 227.11], [686.02, 226.81], [712.58, 237.52], [650.85, 237.97], [634.42, 179.76], [686.02, 179.72], [712.58, 181.08], [650.85, 181.14]]}, {\"category\": \"cyclist\", \"corners_3d\": [[570.51, 257.41], [611.93, 257.08], [619.4, 269.09], [572.08, 269.53], [570.51, 169.29], [611.93, 169.3], [619.4, 168.8], [572.08, 168.78]]}, {\"category\": \"car\", \"corners_3d\": [[-465.75, 603.51], [-870.69, 603.71], [-22.1, 356.61], [150.49, 356.58], [-465.75, 242.55], [-870.69, 242.58], [-22.1, 202.59], [150.49, 202.59]]}, {\"category\": \"car\", \"corners_3d\": [[724.69, 281.67], [840.97, 280.98], [974.53, 336.58], [799.17, 338.17], [724.69, 170.75], [840.97, 170.76], [974.53, 169.69], [799.17, 169.66]]}, {\"category\": \"car\", \"corners_3d\": [[662.11, 245.8], [739.9, 245.1], [802.81, 268.42], [700.39, 269.66], [662.11, 176.05], [739.9, 176.02], [802.81, 177.04], [700.39, 177.1]]}, {\"category\": \"car\", \"corners_3d\": [[355.47, 272.39], [266.01, 273.48], [327.56, 251.19], [396.86, 250.53], [355.47, 194.9], [266.01, 195.14], [327.56, 190.2], [396.86, 190.06]]}]\n```" - } - ] - }, - { - "id": 623, - "data_source": "Mono3DRefer", - "scene_name": "000339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[405.16, 241.22], [346.79, 241.71], [378.78, 230.54], [427.54, 230.19], [405.16, 187.87], [346.79, 187.98], [378.78, 185.52], [427.54, 185.45]]}, {\"category\": \"car\", \"corners_3d\": [[634.42, 227.11], [686.02, 226.81], [712.58, 237.52], [650.85, 237.97], [634.42, 179.76], [686.02, 179.72], [712.58, 181.08], [650.85, 181.14]]}, {\"category\": \"car\", \"corners_3d\": [[-465.75, 603.51], [-870.69, 603.71], [-22.1, 356.61], [150.49, 356.58], [-465.75, 242.55], [-870.69, 242.58], [-22.1, 202.59], [150.49, 202.59]]}, {\"category\": \"car\", \"corners_3d\": [[724.69, 281.67], [840.97, 280.98], [974.53, 336.58], [799.17, 338.17], [724.69, 170.75], [840.97, 170.76], [974.53, 169.69], [799.17, 169.66]]}, {\"category\": \"car\", \"corners_3d\": [[662.11, 245.8], [739.9, 245.1], [802.81, 268.42], [700.39, 269.66], [662.11, 176.05], [739.9, 176.02], [802.81, 177.04], [700.39, 177.1]]}, {\"category\": \"car\", \"corners_3d\": [[355.47, 272.39], [266.01, 273.48], [327.56, 251.19], [396.86, 250.53], [355.47, 194.9], [266.01, 195.14], [327.56, 190.2], [396.86, 190.06]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1188.2, 312.65], [1114.4, 313.41], [1057.56, 298.65], [1123.92, 298.05], [1188.2, 150.79], [1114.4, 150.67], [1057.56, 153.0], [1123.92, 153.1]]}]\n```" - } - ] - }, - { - "id": 624, - "data_source": "Mono3DRefer", - "scene_name": "000339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[558.12, 198.81], [587.11, 198.75], [589.3, 201.54], [557.18, 201.62], [558.12, 166.95], [587.11, 166.97], [589.3, 166.33], [557.18, 166.32]]}, {\"category\": \"cyclist\", \"corners_3d\": [[570.51, 257.41], [611.93, 257.08], [619.4, 269.09], [572.08, 269.53], [570.51, 169.29], [611.93, 169.3], [619.4, 168.8], [572.08, 168.78]]}, {\"category\": \"van\", \"corners_3d\": [[207.53, 337.61], [51.38, 339.03], [235.28, 281.23], [336.39, 280.62], [207.53, 180.8], [51.38, 180.86], [235.28, 178.08], [336.39, 178.05]]}]\n```" - } - ] - }, - { - "id": 625, - "data_source": "Mono3DRefer", - "scene_name": "000339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[558.12, 198.81], [587.11, 198.75], [589.3, 201.54], [557.18, 201.62], [558.12, 166.95], [587.11, 166.97], [589.3, 166.33], [557.18, 166.32]]}, {\"category\": \"van\", \"corners_3d\": [[207.53, 337.61], [51.38, 339.03], [235.28, 281.23], [336.39, 280.62], [207.53, 180.8], [51.38, 180.86], [235.28, 178.08], [336.39, 178.05]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1188.2, 312.65], [1114.4, 313.41], [1057.56, 298.65], [1123.92, 298.05], [1188.2, 150.79], [1114.4, 150.67], [1057.56, 153.0], [1123.92, 153.1]]}]\n```" - } - ] - }, - { - "id": 626, - "data_source": "Mono3DRefer", - "scene_name": "000339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[570.51, 257.41], [611.93, 257.08], [619.4, 269.09], [572.08, 269.53], [570.51, 169.29], [611.93, 169.3], [619.4, 168.8], [572.08, 168.78]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1188.2, 312.65], [1114.4, 313.41], [1057.56, 298.65], [1123.92, 298.05], [1188.2, 150.79], [1114.4, 150.67], [1057.56, 153.0], [1123.92, 153.1]]}]\n```" - } - ] - }, - { - "id": 627, - "data_source": "Mono3DRefer", - "scene_name": "000339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[405.16, 241.22], [346.79, 241.71], [378.78, 230.54], [427.54, 230.19], [405.16, 187.87], [346.79, 187.98], [378.78, 185.52], [427.54, 185.45]]}, {\"category\": \"car\", \"corners_3d\": [[634.42, 227.11], [686.02, 226.81], [712.58, 237.52], [650.85, 237.97], [634.42, 179.76], [686.02, 179.72], [712.58, 181.08], [650.85, 181.14]]}, {\"category\": \"van\", \"corners_3d\": [[558.12, 198.81], [587.11, 198.75], [589.3, 201.54], [557.18, 201.62], [558.12, 166.95], [587.11, 166.97], [589.3, 166.33], [557.18, 166.32]]}, {\"category\": \"cyclist\", \"corners_3d\": [[570.51, 257.41], [611.93, 257.08], [619.4, 269.09], [572.08, 269.53], [570.51, 169.29], [611.93, 169.3], [619.4, 168.8], [572.08, 168.78]]}, {\"category\": \"car\", \"corners_3d\": [[-465.75, 603.51], [-870.69, 603.71], [-22.1, 356.61], [150.49, 356.58], [-465.75, 242.55], [-870.69, 242.58], [-22.1, 202.59], [150.49, 202.59]]}, {\"category\": \"van\", \"corners_3d\": [[207.53, 337.61], [51.38, 339.03], [235.28, 281.23], [336.39, 280.62], [207.53, 180.8], [51.38, 180.86], [235.28, 178.08], [336.39, 178.05]]}, {\"category\": \"car\", \"corners_3d\": [[724.69, 281.67], [840.97, 280.98], [974.53, 336.58], [799.17, 338.17], [724.69, 170.75], [840.97, 170.76], [974.53, 169.69], [799.17, 169.66]]}, {\"category\": \"car\", \"corners_3d\": [[662.11, 245.8], [739.9, 245.1], [802.81, 268.42], [700.39, 269.66], [662.11, 176.05], [739.9, 176.02], [802.81, 177.04], [700.39, 177.1]]}, {\"category\": \"car\", \"corners_3d\": [[355.47, 272.39], [266.01, 273.48], [327.56, 251.19], [396.86, 250.53], [355.47, 194.9], [266.01, 195.14], [327.56, 190.2], [396.86, 190.06]]}]\n```" - } - ] - }, - { - "id": 628, - "data_source": "Mono3DRefer", - "scene_name": "000339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[405.16, 241.22], [346.79, 241.71], [378.78, 230.54], [427.54, 230.19], [405.16, 187.87], [346.79, 187.98], [378.78, 185.52], [427.54, 185.45]]}, {\"category\": \"car\", \"corners_3d\": [[634.42, 227.11], [686.02, 226.81], [712.58, 237.52], [650.85, 237.97], [634.42, 179.76], [686.02, 179.72], [712.58, 181.08], [650.85, 181.14]]}, {\"category\": \"van\", \"corners_3d\": [[558.12, 198.81], [587.11, 198.75], [589.3, 201.54], [557.18, 201.62], [558.12, 166.95], [587.11, 166.97], [589.3, 166.33], [557.18, 166.32]]}, {\"category\": \"car\", \"corners_3d\": [[-465.75, 603.51], [-870.69, 603.71], [-22.1, 356.61], [150.49, 356.58], [-465.75, 242.55], [-870.69, 242.58], [-22.1, 202.59], [150.49, 202.59]]}, {\"category\": \"van\", \"corners_3d\": [[207.53, 337.61], [51.38, 339.03], [235.28, 281.23], [336.39, 280.62], [207.53, 180.8], [51.38, 180.86], [235.28, 178.08], [336.39, 178.05]]}, {\"category\": \"car\", \"corners_3d\": [[724.69, 281.67], [840.97, 280.98], [974.53, 336.58], [799.17, 338.17], [724.69, 170.75], [840.97, 170.76], [974.53, 169.69], [799.17, 169.66]]}, {\"category\": \"car\", \"corners_3d\": [[662.11, 245.8], [739.9, 245.1], [802.81, 268.42], [700.39, 269.66], [662.11, 176.05], [739.9, 176.02], [802.81, 177.04], [700.39, 177.1]]}, {\"category\": \"car\", \"corners_3d\": [[355.47, 272.39], [266.01, 273.48], [327.56, 251.19], [396.86, 250.53], [355.47, 194.9], [266.01, 195.14], [327.56, 190.2], [396.86, 190.06]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1188.2, 312.65], [1114.4, 313.41], [1057.56, 298.65], [1123.92, 298.05], [1188.2, 150.79], [1114.4, 150.67], [1057.56, 153.0], [1123.92, 153.1]]}]\n```" - } - ] - }, - { - "id": 629, - "data_source": "Mono3DRefer", - "scene_name": "000339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[405.16, 241.22], [346.79, 241.71], [378.78, 230.54], [427.54, 230.19], [405.16, 187.87], [346.79, 187.98], [378.78, 185.52], [427.54, 185.45]]}, {\"category\": \"car\", \"corners_3d\": [[634.42, 227.11], [686.02, 226.81], [712.58, 237.52], [650.85, 237.97], [634.42, 179.76], [686.02, 179.72], [712.58, 181.08], [650.85, 181.14]]}, {\"category\": \"cyclist\", \"corners_3d\": [[570.51, 257.41], [611.93, 257.08], [619.4, 269.09], [572.08, 269.53], [570.51, 169.29], [611.93, 169.3], [619.4, 168.8], [572.08, 168.78]]}, {\"category\": \"car\", \"corners_3d\": [[-465.75, 603.51], [-870.69, 603.71], [-22.1, 356.61], [150.49, 356.58], [-465.75, 242.55], [-870.69, 242.58], [-22.1, 202.59], [150.49, 202.59]]}, {\"category\": \"car\", \"corners_3d\": [[724.69, 281.67], [840.97, 280.98], [974.53, 336.58], [799.17, 338.17], [724.69, 170.75], [840.97, 170.76], [974.53, 169.69], [799.17, 169.66]]}, {\"category\": \"car\", \"corners_3d\": [[662.11, 245.8], [739.9, 245.1], [802.81, 268.42], [700.39, 269.66], [662.11, 176.05], [739.9, 176.02], [802.81, 177.04], [700.39, 177.1]]}, {\"category\": \"car\", \"corners_3d\": [[355.47, 272.39], [266.01, 273.48], [327.56, 251.19], [396.86, 250.53], [355.47, 194.9], [266.01, 195.14], [327.56, 190.2], [396.86, 190.06]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1188.2, 312.65], [1114.4, 313.41], [1057.56, 298.65], [1123.92, 298.05], [1188.2, 150.79], [1114.4, 150.67], [1057.56, 153.0], [1123.92, 153.1]]}]\n```" - } - ] - }, - { - "id": 630, - "data_source": "Mono3DRefer", - "scene_name": "000339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[558.12, 198.81], [587.11, 198.75], [589.3, 201.54], [557.18, 201.62], [558.12, 166.95], [587.11, 166.97], [589.3, 166.33], [557.18, 166.32]]}, {\"category\": \"cyclist\", \"corners_3d\": [[570.51, 257.41], [611.93, 257.08], [619.4, 269.09], [572.08, 269.53], [570.51, 169.29], [611.93, 169.3], [619.4, 168.8], [572.08, 168.78]]}, {\"category\": \"van\", \"corners_3d\": [[207.53, 337.61], [51.38, 339.03], [235.28, 281.23], [336.39, 280.62], [207.53, 180.8], [51.38, 180.86], [235.28, 178.08], [336.39, 178.05]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1188.2, 312.65], [1114.4, 313.41], [1057.56, 298.65], [1123.92, 298.05], [1188.2, 150.79], [1114.4, 150.67], [1057.56, 153.0], [1123.92, 153.1]]}]\n```" - } - ] - }, - { - "id": 631, - "data_source": "Mono3DRefer", - "scene_name": "000339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[405.16, 241.22], [346.79, 241.71], [378.78, 230.54], [427.54, 230.19], [405.16, 187.87], [346.79, 187.98], [378.78, 185.52], [427.54, 185.45]]}, {\"category\": \"car\", \"corners_3d\": [[634.42, 227.11], [686.02, 226.81], [712.58, 237.52], [650.85, 237.97], [634.42, 179.76], [686.02, 179.72], [712.58, 181.08], [650.85, 181.14]]}, {\"category\": \"van\", \"corners_3d\": [[558.12, 198.81], [587.11, 198.75], [589.3, 201.54], [557.18, 201.62], [558.12, 166.95], [587.11, 166.97], [589.3, 166.33], [557.18, 166.32]]}, {\"category\": \"cyclist\", \"corners_3d\": [[570.51, 257.41], [611.93, 257.08], [619.4, 269.09], [572.08, 269.53], [570.51, 169.29], [611.93, 169.3], [619.4, 168.8], [572.08, 168.78]]}, {\"category\": \"car\", \"corners_3d\": [[-465.75, 603.51], [-870.69, 603.71], [-22.1, 356.61], [150.49, 356.58], [-465.75, 242.55], [-870.69, 242.58], [-22.1, 202.59], [150.49, 202.59]]}, {\"category\": \"van\", \"corners_3d\": [[207.53, 337.61], [51.38, 339.03], [235.28, 281.23], [336.39, 280.62], [207.53, 180.8], [51.38, 180.86], [235.28, 178.08], [336.39, 178.05]]}, {\"category\": \"car\", \"corners_3d\": [[724.69, 281.67], [840.97, 280.98], [974.53, 336.58], [799.17, 338.17], [724.69, 170.75], [840.97, 170.76], [974.53, 169.69], [799.17, 169.66]]}, {\"category\": \"car\", \"corners_3d\": [[662.11, 245.8], [739.9, 245.1], [802.81, 268.42], [700.39, 269.66], [662.11, 176.05], [739.9, 176.02], [802.81, 177.04], [700.39, 177.1]]}, {\"category\": \"car\", \"corners_3d\": [[355.47, 272.39], [266.01, 273.48], [327.56, 251.19], [396.86, 250.53], [355.47, 194.9], [266.01, 195.14], [327.56, 190.2], [396.86, 190.06]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1188.2, 312.65], [1114.4, 313.41], [1057.56, 298.65], [1123.92, 298.05], [1188.2, 150.79], [1114.4, 150.67], [1057.56, 153.0], [1123.92, 153.1]]}]\n```" - } - ] - }, - { - "id": 632, - "data_source": "Mono3DRefer", - "scene_name": "000341", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[243.78, 291.82], [292.97, 290.44], [283.26, 296.95], [230.96, 298.51], [243.78, 160.48], [292.97, 160.74], [283.26, 159.5], [230.96, 159.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[270.73, 286.86], [231.71, 286.69], [250.32, 281.67], [287.5, 281.82], [270.73, 158.05], [231.71, 158.09], [250.32, 159.21], [287.5, 159.17]]}]\n```" - } - ] - }, - { - "id": 633, - "data_source": "Mono3DRefer", - "scene_name": "000341", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[422.04, 287.55], [472.58, 287.56], [462.11, 296.21], [407.45, 296.2], [422.04, 192.24], [472.58, 192.24], [462.11, 193.11], [407.45, 193.11]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[376.16, 285.95], [425.18, 285.96], [410.66, 294.58], [357.6, 294.58], [376.16, 192.22], [425.18, 192.22], [410.66, 193.1], [357.6, 193.1]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[435.6, 282.89], [394.67, 282.72], [407.95, 276.79], [446.49, 276.94], [435.6, 192.61], [394.67, 192.59], [407.95, 191.94], [446.49, 191.95]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[465.73, 261.6], [432.69, 261.56], [440.62, 257.92], [472.16, 257.95], [465.73, 185.49], [432.69, 185.49], [440.62, 185.31], [472.16, 185.31]]}]\n```" - } - ] - }, - { - "id": 634, - "data_source": "Mono3DRefer", - "scene_name": "000341", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[530.47, 192.71], [541.89, 193.3], [345.34, 194.55], [343.58, 193.83], [530.47, 141.73], [541.89, 139.63], [345.34, 135.2], [343.58, 137.75]]}]\n```" - } - ] - }, - { - "id": 635, - "data_source": "Mono3DRefer", - "scene_name": "000341", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[801.95, 307.64], [947.67, 306.88], [1242.89, 407.39], [982.29, 409.85], [801.95, 148.52], [947.67, 148.72], [1242.89, 122.43], [982.29, 121.79]]}]\n```" - } - ] - }, - { - "id": 636, - "data_source": "Mono3DRefer", - "scene_name": "000341", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[422.04, 287.55], [472.58, 287.56], [462.11, 296.21], [407.45, 296.2], [422.04, 192.24], [472.58, 192.24], [462.11, 193.11], [407.45, 193.11]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[376.16, 285.95], [425.18, 285.96], [410.66, 294.58], [357.6, 294.58], [376.16, 192.22], [425.18, 192.22], [410.66, 193.1], [357.6, 193.1]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[435.6, 282.89], [394.67, 282.72], [407.95, 276.79], [446.49, 276.94], [435.6, 192.61], [394.67, 192.59], [407.95, 191.94], [446.49, 191.95]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[465.73, 261.6], [432.69, 261.56], [440.62, 257.92], [472.16, 257.95], [465.73, 185.49], [432.69, 185.49], [440.62, 185.31], [472.16, 185.31]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[243.78, 291.82], [292.97, 290.44], [283.26, 296.95], [230.96, 298.51], [243.78, 160.48], [292.97, 160.74], [283.26, 159.5], [230.96, 159.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[270.73, 286.86], [231.71, 286.69], [250.32, 281.67], [287.5, 281.82], [270.73, 158.05], [231.71, 158.09], [250.32, 159.21], [287.5, 159.17]]}]\n```" - } - ] - }, - { - "id": 637, - "data_source": "Mono3DRefer", - "scene_name": "000341", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[530.47, 192.71], [541.89, 193.3], [345.34, 194.55], [343.58, 193.83], [530.47, 141.73], [541.89, 139.63], [345.34, 135.2], [343.58, 137.75]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[243.78, 291.82], [292.97, 290.44], [283.26, 296.95], [230.96, 298.51], [243.78, 160.48], [292.97, 160.74], [283.26, 159.5], [230.96, 159.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[270.73, 286.86], [231.71, 286.69], [250.32, 281.67], [287.5, 281.82], [270.73, 158.05], [231.71, 158.09], [250.32, 159.21], [287.5, 159.17]]}]\n```" - } - ] - }, - { - "id": 638, - "data_source": "Mono3DRefer", - "scene_name": "000341", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[801.95, 307.64], [947.67, 306.88], [1242.89, 407.39], [982.29, 409.85], [801.95, 148.52], [947.67, 148.72], [1242.89, 122.43], [982.29, 121.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[243.78, 291.82], [292.97, 290.44], [283.26, 296.95], [230.96, 298.51], [243.78, 160.48], [292.97, 160.74], [283.26, 159.5], [230.96, 159.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[270.73, 286.86], [231.71, 286.69], [250.32, 281.67], [287.5, 281.82], [270.73, 158.05], [231.71, 158.09], [250.32, 159.21], [287.5, 159.17]]}]\n```" - } - ] - }, - { - "id": 639, - "data_source": "Mono3DRefer", - "scene_name": "000341", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[530.47, 192.71], [541.89, 193.3], [345.34, 194.55], [343.58, 193.83], [530.47, 141.73], [541.89, 139.63], [345.34, 135.2], [343.58, 137.75]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[422.04, 287.55], [472.58, 287.56], [462.11, 296.21], [407.45, 296.2], [422.04, 192.24], [472.58, 192.24], [462.11, 193.11], [407.45, 193.11]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[376.16, 285.95], [425.18, 285.96], [410.66, 294.58], [357.6, 294.58], [376.16, 192.22], [425.18, 192.22], [410.66, 193.1], [357.6, 193.1]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[435.6, 282.89], [394.67, 282.72], [407.95, 276.79], [446.49, 276.94], [435.6, 192.61], [394.67, 192.59], [407.95, 191.94], [446.49, 191.95]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[465.73, 261.6], [432.69, 261.56], [440.62, 257.92], [472.16, 257.95], [465.73, 185.49], [432.69, 185.49], [440.62, 185.31], [472.16, 185.31]]}]\n```" - } - ] - }, - { - "id": 640, - "data_source": "Mono3DRefer", - "scene_name": "000341", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[801.95, 307.64], [947.67, 306.88], [1242.89, 407.39], [982.29, 409.85], [801.95, 148.52], [947.67, 148.72], [1242.89, 122.43], [982.29, 121.79]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[422.04, 287.55], [472.58, 287.56], [462.11, 296.21], [407.45, 296.2], [422.04, 192.24], [472.58, 192.24], [462.11, 193.11], [407.45, 193.11]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[376.16, 285.95], [425.18, 285.96], [410.66, 294.58], [357.6, 294.58], [376.16, 192.22], [425.18, 192.22], [410.66, 193.1], [357.6, 193.1]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[435.6, 282.89], [394.67, 282.72], [407.95, 276.79], [446.49, 276.94], [435.6, 192.61], [394.67, 192.59], [407.95, 191.94], [446.49, 191.95]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[465.73, 261.6], [432.69, 261.56], [440.62, 257.92], [472.16, 257.95], [465.73, 185.49], [432.69, 185.49], [440.62, 185.31], [472.16, 185.31]]}]\n```" - } - ] - }, - { - "id": 641, - "data_source": "Mono3DRefer", - "scene_name": "000341", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[801.95, 307.64], [947.67, 306.88], [1242.89, 407.39], [982.29, 409.85], [801.95, 148.52], [947.67, 148.72], [1242.89, 122.43], [982.29, 121.79]]}, {\"category\": \"tram\", \"corners_3d\": [[530.47, 192.71], [541.89, 193.3], [345.34, 194.55], [343.58, 193.83], [530.47, 141.73], [541.89, 139.63], [345.34, 135.2], [343.58, 137.75]]}]\n```" - } - ] - }, - { - "id": 642, - "data_source": "Mono3DRefer", - "scene_name": "000341", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, person_sitting, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[530.47, 192.71], [541.89, 193.3], [345.34, 194.55], [343.58, 193.83], [530.47, 141.73], [541.89, 139.63], [345.34, 135.2], [343.58, 137.75]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[422.04, 287.55], [472.58, 287.56], [462.11, 296.21], [407.45, 296.2], [422.04, 192.24], [472.58, 192.24], [462.11, 193.11], [407.45, 193.11]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[376.16, 285.95], [425.18, 285.96], [410.66, 294.58], [357.6, 294.58], [376.16, 192.22], [425.18, 192.22], [410.66, 193.1], [357.6, 193.1]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[435.6, 282.89], [394.67, 282.72], [407.95, 276.79], [446.49, 276.94], [435.6, 192.61], [394.67, 192.59], [407.95, 191.94], [446.49, 191.95]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[465.73, 261.6], [432.69, 261.56], [440.62, 257.92], [472.16, 257.95], [465.73, 185.49], [432.69, 185.49], [440.62, 185.31], [472.16, 185.31]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[243.78, 291.82], [292.97, 290.44], [283.26, 296.95], [230.96, 298.51], [243.78, 160.48], [292.97, 160.74], [283.26, 159.5], [230.96, 159.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[270.73, 286.86], [231.71, 286.69], [250.32, 281.67], [287.5, 281.82], [270.73, 158.05], [231.71, 158.09], [250.32, 159.21], [287.5, 159.17]]}]\n```" - } - ] - }, - { - "id": 643, - "data_source": "Mono3DRefer", - "scene_name": "000341", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, person_sitting, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[801.95, 307.64], [947.67, 306.88], [1242.89, 407.39], [982.29, 409.85], [801.95, 148.52], [947.67, 148.72], [1242.89, 122.43], [982.29, 121.79]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[422.04, 287.55], [472.58, 287.56], [462.11, 296.21], [407.45, 296.2], [422.04, 192.24], [472.58, 192.24], [462.11, 193.11], [407.45, 193.11]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[376.16, 285.95], [425.18, 285.96], [410.66, 294.58], [357.6, 294.58], [376.16, 192.22], [425.18, 192.22], [410.66, 193.1], [357.6, 193.1]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[435.6, 282.89], [394.67, 282.72], [407.95, 276.79], [446.49, 276.94], [435.6, 192.61], [394.67, 192.59], [407.95, 191.94], [446.49, 191.95]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[465.73, 261.6], [432.69, 261.56], [440.62, 257.92], [472.16, 257.95], [465.73, 185.49], [432.69, 185.49], [440.62, 185.31], [472.16, 185.31]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[243.78, 291.82], [292.97, 290.44], [283.26, 296.95], [230.96, 298.51], [243.78, 160.48], [292.97, 160.74], [283.26, 159.5], [230.96, 159.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[270.73, 286.86], [231.71, 286.69], [250.32, 281.67], [287.5, 281.82], [270.73, 158.05], [231.71, 158.09], [250.32, 159.21], [287.5, 159.17]]}]\n```" - } - ] - }, - { - "id": 644, - "data_source": "Mono3DRefer", - "scene_name": "000341", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, tram, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[801.95, 307.64], [947.67, 306.88], [1242.89, 407.39], [982.29, 409.85], [801.95, 148.52], [947.67, 148.72], [1242.89, 122.43], [982.29, 121.79]]}, {\"category\": \"tram\", \"corners_3d\": [[530.47, 192.71], [541.89, 193.3], [345.34, 194.55], [343.58, 193.83], [530.47, 141.73], [541.89, 139.63], [345.34, 135.2], [343.58, 137.75]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[243.78, 291.82], [292.97, 290.44], [283.26, 296.95], [230.96, 298.51], [243.78, 160.48], [292.97, 160.74], [283.26, 159.5], [230.96, 159.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[270.73, 286.86], [231.71, 286.69], [250.32, 281.67], [287.5, 281.82], [270.73, 158.05], [231.71, 158.09], [250.32, 159.21], [287.5, 159.17]]}]\n```" - } - ] - }, - { - "id": 645, - "data_source": "Mono3DRefer", - "scene_name": "000341", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, tram, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[801.95, 307.64], [947.67, 306.88], [1242.89, 407.39], [982.29, 409.85], [801.95, 148.52], [947.67, 148.72], [1242.89, 122.43], [982.29, 121.79]]}, {\"category\": \"tram\", \"corners_3d\": [[530.47, 192.71], [541.89, 193.3], [345.34, 194.55], [343.58, 193.83], [530.47, 141.73], [541.89, 139.63], [345.34, 135.2], [343.58, 137.75]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[422.04, 287.55], [472.58, 287.56], [462.11, 296.21], [407.45, 296.2], [422.04, 192.24], [472.58, 192.24], [462.11, 193.11], [407.45, 193.11]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[376.16, 285.95], [425.18, 285.96], [410.66, 294.58], [357.6, 294.58], [376.16, 192.22], [425.18, 192.22], [410.66, 193.1], [357.6, 193.1]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[435.6, 282.89], [394.67, 282.72], [407.95, 276.79], [446.49, 276.94], [435.6, 192.61], [394.67, 192.59], [407.95, 191.94], [446.49, 191.95]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[465.73, 261.6], [432.69, 261.56], [440.62, 257.92], [472.16, 257.95], [465.73, 185.49], [432.69, 185.49], [440.62, 185.31], [472.16, 185.31]]}]\n```" - } - ] - }, - { - "id": 646, - "data_source": "Mono3DRefer", - "scene_name": "000341", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, person_sitting, tram, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[801.95, 307.64], [947.67, 306.88], [1242.89, 407.39], [982.29, 409.85], [801.95, 148.52], [947.67, 148.72], [1242.89, 122.43], [982.29, 121.79]]}, {\"category\": \"tram\", \"corners_3d\": [[530.47, 192.71], [541.89, 193.3], [345.34, 194.55], [343.58, 193.83], [530.47, 141.73], [541.89, 139.63], [345.34, 135.2], [343.58, 137.75]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[422.04, 287.55], [472.58, 287.56], [462.11, 296.21], [407.45, 296.2], [422.04, 192.24], [472.58, 192.24], [462.11, 193.11], [407.45, 193.11]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[376.16, 285.95], [425.18, 285.96], [410.66, 294.58], [357.6, 294.58], [376.16, 192.22], [425.18, 192.22], [410.66, 193.1], [357.6, 193.1]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[435.6, 282.89], [394.67, 282.72], [407.95, 276.79], [446.49, 276.94], [435.6, 192.61], [394.67, 192.59], [407.95, 191.94], [446.49, 191.95]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[465.73, 261.6], [432.69, 261.56], [440.62, 257.92], [472.16, 257.95], [465.73, 185.49], [432.69, 185.49], [440.62, 185.31], [472.16, 185.31]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[243.78, 291.82], [292.97, 290.44], [283.26, 296.95], [230.96, 298.51], [243.78, 160.48], [292.97, 160.74], [283.26, 159.5], [230.96, 159.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[270.73, 286.86], [231.71, 286.69], [250.32, 281.67], [287.5, 281.82], [270.73, 158.05], [231.71, 158.09], [250.32, 159.21], [287.5, 159.17]]}]\n```" - } - ] - }, - { - "id": 647, - "data_source": "Mono3DRefer", - "scene_name": "000343", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[795.99, 203.36], [830.31, 203.67], [836.84, 207.1], [798.68, 206.72], [795.99, 167.06], [830.31, 167.0], [836.84, 166.35], [798.68, 166.43]]}, {\"category\": \"car\", \"corners_3d\": [[871.8, 247.38], [957.53, 248.82], [1051.48, 280.75], [928.17, 277.87], [871.8, 164.66], [957.53, 164.5], [1051.48, 160.99], [928.17, 161.31]]}, {\"category\": \"car\", \"corners_3d\": [[675.91, 240.26], [609.56, 239.16], [633.36, 227.08], [687.79, 227.82], [675.91, 178.02], [609.56, 177.94], [633.36, 177.01], [687.79, 177.07]]}, {\"category\": \"car\", \"corners_3d\": [[687.33, 201.82], [717.6, 201.98], [719.27, 205.67], [685.18, 205.47], [687.33, 172.85], [717.6, 172.85], [719.27, 172.85], [685.18, 172.85]]}]\n```" - } - ] - }, - { - "id": 648, - "data_source": "Mono3DRefer", - "scene_name": "000345", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[573.43, 218.78], [618.54, 218.73], [622.52, 224.05], [570.26, 224.1], [573.43, 177.75], [618.54, 177.76], [622.52, 176.58], [570.26, 176.57]]}, {\"category\": \"car\", \"corners_3d\": [[353.6, 269.9], [444.3, 269.7], [377.74, 307.42], [246.17, 307.85], [353.6, 194.07], [444.3, 194.05], [377.74, 198.0], [246.17, 198.04]]}, {\"category\": \"car\", \"corners_3d\": [[393.51, 216.17], [429.12, 216.17], [402.35, 220.81], [361.4, 220.81], [393.51, 182.02], [429.12, 182.02], [402.35, 181.54], [361.4, 181.54]]}, {\"category\": \"car\", \"corners_3d\": [[366.64, 204.58], [395.43, 204.58], [374.56, 206.55], [342.83, 206.55], [366.64, 177.15], [395.43, 177.15], [374.56, 176.33], [342.83, 176.33]]}, {\"category\": \"car\", \"corners_3d\": [[539.9, 200.86], [563.96, 200.84], [562.46, 202.41], [535.99, 202.43], [539.9, 181.47], [563.96, 181.47], [562.46, 181.1], [535.99, 181.09]]}]\n```" - } - ] - }, - { - "id": 649, - "data_source": "Mono3DRefer", - "scene_name": "000345", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[505.23, 209.0], [553.68, 208.97], [522.23, 227.79], [435.23, 227.89], [505.23, 130.61], [553.68, 130.68], [522.23, 87.48], [435.23, 87.25]]}]\n```" - } - ] - }, - { - "id": 650, - "data_source": "Mono3DRefer", - "scene_name": "000345", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[573.43, 218.78], [618.54, 218.73], [622.52, 224.05], [570.26, 224.1], [573.43, 177.75], [618.54, 177.76], [622.52, 176.58], [570.26, 176.57]]}, {\"category\": \"car\", \"corners_3d\": [[353.6, 269.9], [444.3, 269.7], [377.74, 307.42], [246.17, 307.85], [353.6, 194.07], [444.3, 194.05], [377.74, 198.0], [246.17, 198.04]]}, {\"category\": \"car\", \"corners_3d\": [[393.51, 216.17], [429.12, 216.17], [402.35, 220.81], [361.4, 220.81], [393.51, 182.02], [429.12, 182.02], [402.35, 181.54], [361.4, 181.54]]}, {\"category\": \"truck\", \"corners_3d\": [[505.23, 209.0], [553.68, 208.97], [522.23, 227.79], [435.23, 227.89], [505.23, 130.61], [553.68, 130.68], [522.23, 87.48], [435.23, 87.25]]}, {\"category\": \"car\", \"corners_3d\": [[366.64, 204.58], [395.43, 204.58], [374.56, 206.55], [342.83, 206.55], [366.64, 177.15], [395.43, 177.15], [374.56, 176.33], [342.83, 176.33]]}, {\"category\": \"car\", \"corners_3d\": [[539.9, 200.86], [563.96, 200.84], [562.46, 202.41], [535.99, 202.43], [539.9, 181.47], [563.96, 181.47], [562.46, 181.1], [535.99, 181.09]]}]\n```" - } - ] - }, - { - "id": 651, - "data_source": "Mono3DRefer", - "scene_name": "000349", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[399.2, 287.74], [291.95, 287.41], [376.02, 258.35], [456.19, 258.54], [399.2, 190.97], [291.95, 190.92], [376.02, 186.34], [456.19, 186.37]]}, {\"category\": \"car\", \"corners_3d\": [[320.76, 259.02], [250.66, 258.24], [327.57, 242.73], [385.42, 243.25], [320.76, 191.13], [250.66, 190.97], [327.57, 187.68], [385.42, 187.79]]}, {\"category\": \"car\", \"corners_3d\": [[620.97, 219.75], [667.46, 219.94], [669.44, 226.61], [616.38, 226.37], [620.97, 174.8], [667.46, 174.81], [669.44, 175.08], [616.38, 175.07]]}]\n```" - } - ] - }, - { - "id": 652, - "data_source": "Mono3DRefer", - "scene_name": "000352", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[95.61, 310.38], [0.96, 334.36], [-286.14, 331.69], [-150.8, 308.44], [95.61, 204.57], [0.96, 210.1], [-286.14, 209.49], [-150.8, 204.12]]}, {\"category\": \"car\", \"corners_3d\": [[711.38, 245.02], [784.51, 245.03], [840.79, 268.32], [744.05, 268.31], [711.38, 172.37], [784.51, 172.37], [840.79, 172.22], [744.05, 172.22]]}, {\"category\": \"car\", \"corners_3d\": [[403.3, 284.35], [300.17, 284.99], [367.73, 257.93], [445.78, 257.56], [403.3, 197.15], [300.17, 197.29], [367.73, 191.4], [445.78, 191.31]]}, {\"category\": \"car\", \"corners_3d\": [[466.04, 248.8], [398.07, 248.88], [433.53, 235.65], [489.65, 235.59], [466.04, 177.77], [398.07, 177.77], [433.53, 176.92], [489.65, 176.91]]}, {\"category\": \"car\", \"corners_3d\": [[671.34, 221.03], [721.34, 221.03], [742.24, 230.09], [682.84, 230.08], [671.34, 174.65], [721.34, 174.65], [742.24, 174.99], [682.84, 174.99]]}, {\"category\": \"car\", \"corners_3d\": [[506.99, 234.06], [455.97, 234.11], [477.22, 225.18], [520.8, 225.14], [506.99, 181.59], [455.97, 181.59], [477.22, 180.32], [520.8, 180.32]]}, {\"category\": \"car\", \"corners_3d\": [[661.86, 212.87], [700.11, 212.68], [724.32, 218.9], [680.2, 219.15], [661.86, 177.82], [700.11, 177.8], [724.32, 178.57], [680.2, 178.6]]}, {\"category\": \"car\", \"corners_3d\": [[626.86, 197.39], [646.43, 197.35], [651.18, 198.81], [630.46, 198.86], [626.86, 179.69], [646.43, 179.68], [651.18, 180.08], [630.46, 180.09]]}, {\"category\": \"car\", \"corners_3d\": [[529.37, 199.28], [549.03, 199.12], [555.04, 200.7], [534.24, 200.88], [529.37, 183.61], [549.03, 183.55], [555.04, 184.19], [534.24, 184.26]]}, {\"category\": \"car\", \"corners_3d\": [[762.32, 290.5], [881.84, 290.32], [1107.07, 385.18], [891.4, 385.76], [762.32, 176.6], [881.84, 176.59], [1107.07, 179.61], [891.4, 179.63]]}]\n```" - } - ] - }, - { - "id": 653, - "data_source": "Mono3DRefer", - "scene_name": "000352", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[644.45, 204.54], [690.04, 204.54], [704.18, 210.15], [650.51, 210.15], [644.45, 151.72], [690.04, 151.72], [704.18, 147.98], [650.51, 147.98]]}]\n```" - } - ] - }, - { - "id": 654, - "data_source": "Mono3DRefer", - "scene_name": "000352", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[95.61, 310.38], [0.96, 334.36], [-286.14, 331.69], [-150.8, 308.44], [95.61, 204.57], [0.96, 210.1], [-286.14, 209.49], [-150.8, 204.12]]}, {\"category\": \"car\", \"corners_3d\": [[711.38, 245.02], [784.51, 245.03], [840.79, 268.32], [744.05, 268.31], [711.38, 172.37], [784.51, 172.37], [840.79, 172.22], [744.05, 172.22]]}, {\"category\": \"car\", \"corners_3d\": [[403.3, 284.35], [300.17, 284.99], [367.73, 257.93], [445.78, 257.56], [403.3, 197.15], [300.17, 197.29], [367.73, 191.4], [445.78, 191.31]]}, {\"category\": \"car\", \"corners_3d\": [[466.04, 248.8], [398.07, 248.88], [433.53, 235.65], [489.65, 235.59], [466.04, 177.77], [398.07, 177.77], [433.53, 176.92], [489.65, 176.91]]}, {\"category\": \"car\", \"corners_3d\": [[671.34, 221.03], [721.34, 221.03], [742.24, 230.09], [682.84, 230.08], [671.34, 174.65], [721.34, 174.65], [742.24, 174.99], [682.84, 174.99]]}, {\"category\": \"car\", \"corners_3d\": [[506.99, 234.06], [455.97, 234.11], [477.22, 225.18], [520.8, 225.14], [506.99, 181.59], [455.97, 181.59], [477.22, 180.32], [520.8, 180.32]]}, {\"category\": \"car\", \"corners_3d\": [[661.86, 212.87], [700.11, 212.68], [724.32, 218.9], [680.2, 219.15], [661.86, 177.82], [700.11, 177.8], [724.32, 178.57], [680.2, 178.6]]}, {\"category\": \"van\", \"corners_3d\": [[644.45, 204.54], [690.04, 204.54], [704.18, 210.15], [650.51, 210.15], [644.45, 151.72], [690.04, 151.72], [704.18, 147.98], [650.51, 147.98]]}, {\"category\": \"car\", \"corners_3d\": [[626.86, 197.39], [646.43, 197.35], [651.18, 198.81], [630.46, 198.86], [626.86, 179.69], [646.43, 179.68], [651.18, 180.08], [630.46, 180.09]]}, {\"category\": \"car\", \"corners_3d\": [[529.37, 199.28], [549.03, 199.12], [555.04, 200.7], [534.24, 200.88], [529.37, 183.61], [549.03, 183.55], [555.04, 184.19], [534.24, 184.26]]}, {\"category\": \"car\", \"corners_3d\": [[762.32, 290.5], [881.84, 290.32], [1107.07, 385.18], [891.4, 385.76], [762.32, 176.6], [881.84, 176.59], [1107.07, 179.61], [891.4, 179.63]]}]\n```" - } - ] - }, - { - "id": 655, - "data_source": "Mono3DRefer", - "scene_name": "000353", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[100.07, 356.63], [-139.02, 363.96], [146.88, 280.37], [277.41, 278.01], [100.07, 182.68], [-139.02, 183.07], [146.88, 178.6], [277.41, 178.48]]}, {\"category\": \"car\", \"corners_3d\": [[354.31, 238.92], [265.63, 239.56], [330.43, 224.42], [398.65, 224.04], [354.31, 159.62], [265.63, 159.49], [330.43, 162.52], [398.65, 162.6]]}]\n```" - } - ] - }, - { - "id": 656, - "data_source": "Mono3DRefer", - "scene_name": "000356", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[633.15, 244.05], [703.09, 243.36], [744.17, 260.87], [657.14, 261.95], [633.15, 179.05], [703.09, 178.99], [744.17, 180.51], [657.14, 180.61]]}, {\"category\": \"car\", \"corners_3d\": [[244.44, 451.92], [-136.44, 476.57], [205.36, 307.52], [369.36, 302.45], [244.44, 105.6], [-136.44, 99.65], [205.36, 140.4], [369.36, 141.62]]}, {\"category\": \"car\", \"corners_3d\": [[421.66, 224.42], [467.57, 223.94], [462.12, 232.24], [408.62, 232.88], [421.66, 179.86], [467.57, 179.8], [462.12, 180.93], [408.62, 181.02]]}, {\"category\": \"car\", \"corners_3d\": [[612.96, 226.14], [658.86, 225.77], [678.15, 233.72], [625.46, 234.21], [612.96, 179.31], [658.86, 179.27], [678.15, 180.23], [625.46, 180.29]]}, {\"category\": \"car\", \"corners_3d\": [[588.1, 219.02], [627.95, 218.69], [643.66, 225.12], [598.32, 225.55], [588.1, 179.99], [627.95, 179.94], [643.66, 180.93], [598.32, 181.0]]}, {\"category\": \"car\", \"corners_3d\": [[570.74, 204.99], [598.48, 204.89], [602.98, 208.01], [572.54, 208.13], [570.74, 179.62], [598.48, 179.6], [602.98, 180.26], [572.54, 180.29]]}]\n```" - } - ] - }, - { - "id": 657, - "data_source": "Mono3DRefer", - "scene_name": "000356", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[545.04, 193.97], [564.58, 193.87], [569.98, 195.32], [549.12, 195.43], [545.04, 172.74], [564.58, 172.74], [569.98, 172.74], [549.12, 172.73]]}]\n```" - } - ] - }, - { - "id": 658, - "data_source": "Mono3DRefer", - "scene_name": "000356", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[481.08, 211.41], [433.65, 211.69], [447.52, 206.06], [488.03, 205.85], [481.08, 152.77], [433.65, 152.63], [447.52, 155.56], [488.03, 155.67]]}]\n```" - } - ] - }, - { - "id": 659, - "data_source": "Mono3DRefer", - "scene_name": "000356", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[633.15, 244.05], [703.09, 243.36], [744.17, 260.87], [657.14, 261.95], [633.15, 179.05], [703.09, 178.99], [744.17, 180.51], [657.14, 180.61]]}, {\"category\": \"car\", \"corners_3d\": [[244.44, 451.92], [-136.44, 476.57], [205.36, 307.52], [369.36, 302.45], [244.44, 105.6], [-136.44, 99.65], [205.36, 140.4], [369.36, 141.62]]}, {\"category\": \"car\", \"corners_3d\": [[421.66, 224.42], [467.57, 223.94], [462.12, 232.24], [408.62, 232.88], [421.66, 179.86], [467.57, 179.8], [462.12, 180.93], [408.62, 181.02]]}, {\"category\": \"car\", \"corners_3d\": [[612.96, 226.14], [658.86, 225.77], [678.15, 233.72], [625.46, 234.21], [612.96, 179.31], [658.86, 179.27], [678.15, 180.23], [625.46, 180.29]]}, {\"category\": \"car\", \"corners_3d\": [[588.1, 219.02], [627.95, 218.69], [643.66, 225.12], [598.32, 225.55], [588.1, 179.99], [627.95, 179.94], [643.66, 180.93], [598.32, 181.0]]}, {\"category\": \"car\", \"corners_3d\": [[570.74, 204.99], [598.48, 204.89], [602.98, 208.01], [572.54, 208.13], [570.74, 179.62], [598.48, 179.6], [602.98, 180.26], [572.54, 180.29]]}, {\"category\": \"van\", \"corners_3d\": [[545.04, 193.97], [564.58, 193.87], [569.98, 195.32], [549.12, 195.43], [545.04, 172.74], [564.58, 172.74], [569.98, 172.74], [549.12, 172.73]]}]\n```" - } - ] - }, - { - "id": 660, - "data_source": "Mono3DRefer", - "scene_name": "000356", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[633.15, 244.05], [703.09, 243.36], [744.17, 260.87], [657.14, 261.95], [633.15, 179.05], [703.09, 178.99], [744.17, 180.51], [657.14, 180.61]]}, {\"category\": \"car\", \"corners_3d\": [[244.44, 451.92], [-136.44, 476.57], [205.36, 307.52], [369.36, 302.45], [244.44, 105.6], [-136.44, 99.65], [205.36, 140.4], [369.36, 141.62]]}, {\"category\": \"car\", \"corners_3d\": [[421.66, 224.42], [467.57, 223.94], [462.12, 232.24], [408.62, 232.88], [421.66, 179.86], [467.57, 179.8], [462.12, 180.93], [408.62, 181.02]]}, {\"category\": \"car\", \"corners_3d\": [[612.96, 226.14], [658.86, 225.77], [678.15, 233.72], [625.46, 234.21], [612.96, 179.31], [658.86, 179.27], [678.15, 180.23], [625.46, 180.29]]}, {\"category\": \"car\", \"corners_3d\": [[588.1, 219.02], [627.95, 218.69], [643.66, 225.12], [598.32, 225.55], [588.1, 179.99], [627.95, 179.94], [643.66, 180.93], [598.32, 181.0]]}, {\"category\": \"truck\", \"corners_3d\": [[481.08, 211.41], [433.65, 211.69], [447.52, 206.06], [488.03, 205.85], [481.08, 152.77], [433.65, 152.63], [447.52, 155.56], [488.03, 155.67]]}, {\"category\": \"car\", \"corners_3d\": [[570.74, 204.99], [598.48, 204.89], [602.98, 208.01], [572.54, 208.13], [570.74, 179.62], [598.48, 179.6], [602.98, 180.26], [572.54, 180.29]]}]\n```" - } - ] - }, - { - "id": 661, - "data_source": "Mono3DRefer", - "scene_name": "000356", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[481.08, 211.41], [433.65, 211.69], [447.52, 206.06], [488.03, 205.85], [481.08, 152.77], [433.65, 152.63], [447.52, 155.56], [488.03, 155.67]]}, {\"category\": \"van\", \"corners_3d\": [[545.04, 193.97], [564.58, 193.87], [569.98, 195.32], [549.12, 195.43], [545.04, 172.74], [564.58, 172.74], [569.98, 172.74], [549.12, 172.73]]}]\n```" - } - ] - }, - { - "id": 662, - "data_source": "Mono3DRefer", - "scene_name": "000356", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[633.15, 244.05], [703.09, 243.36], [744.17, 260.87], [657.14, 261.95], [633.15, 179.05], [703.09, 178.99], [744.17, 180.51], [657.14, 180.61]]}, {\"category\": \"car\", \"corners_3d\": [[244.44, 451.92], [-136.44, 476.57], [205.36, 307.52], [369.36, 302.45], [244.44, 105.6], [-136.44, 99.65], [205.36, 140.4], [369.36, 141.62]]}, {\"category\": \"car\", \"corners_3d\": [[421.66, 224.42], [467.57, 223.94], [462.12, 232.24], [408.62, 232.88], [421.66, 179.86], [467.57, 179.8], [462.12, 180.93], [408.62, 181.02]]}, {\"category\": \"car\", \"corners_3d\": [[612.96, 226.14], [658.86, 225.77], [678.15, 233.72], [625.46, 234.21], [612.96, 179.31], [658.86, 179.27], [678.15, 180.23], [625.46, 180.29]]}, {\"category\": \"car\", \"corners_3d\": [[588.1, 219.02], [627.95, 218.69], [643.66, 225.12], [598.32, 225.55], [588.1, 179.99], [627.95, 179.94], [643.66, 180.93], [598.32, 181.0]]}, {\"category\": \"truck\", \"corners_3d\": [[481.08, 211.41], [433.65, 211.69], [447.52, 206.06], [488.03, 205.85], [481.08, 152.77], [433.65, 152.63], [447.52, 155.56], [488.03, 155.67]]}, {\"category\": \"car\", \"corners_3d\": [[570.74, 204.99], [598.48, 204.89], [602.98, 208.01], [572.54, 208.13], [570.74, 179.62], [598.48, 179.6], [602.98, 180.26], [572.54, 180.29]]}, {\"category\": \"van\", \"corners_3d\": [[545.04, 193.97], [564.58, 193.87], [569.98, 195.32], [549.12, 195.43], [545.04, 172.74], [564.58, 172.74], [569.98, 172.74], [549.12, 172.73]]}]\n```" - } - ] - }, - { - "id": 663, - "data_source": "Mono3DRefer", - "scene_name": "000360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[368.41, 208.7], [337.45, 208.64], [363.65, 205.52], [391.95, 205.57], [368.41, 181.05], [337.45, 181.04], [363.65, 180.32], [391.95, 180.34]]}]\n```" - } - ] - }, - { - "id": 664, - "data_source": "Mono3DRefer", - "scene_name": "000360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[126.52, 228.77], [114.23, 228.79], [130.07, 226.94], [141.96, 226.92], [126.52, 197.1], [114.23, 197.11], [130.07, 196.31], [141.96, 196.3]]}]\n```" - } - ] - }, - { - "id": 665, - "data_source": "Mono3DRefer", - "scene_name": "000360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[593.41, 189.25], [616.59, 189.24], [619.02, 192.15], [591.73, 192.15], [593.41, 164.12], [616.59, 164.13], [619.02, 162.58], [591.73, 162.58]]}]\n```" - } - ] - }, - { - "id": 666, - "data_source": "Mono3DRefer", - "scene_name": "000360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[777.23, 236.53], [818.3, 233.95], [929.61, 244.41], [886.72, 247.98], [777.23, 183.95], [818.3, 183.5], [929.61, 185.32], [886.72, 185.94]]}, {\"category\": \"car\", \"corners_3d\": [[296.07, 217.28], [261.71, 217.26], [290.37, 213.67], [321.96, 213.68], [296.07, 188.1], [261.71, 188.1], [290.37, 186.86], [321.96, 186.87]]}]\n```" - } - ] - }, - { - "id": 667, - "data_source": "Mono3DRefer", - "scene_name": "000360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[368.41, 208.7], [337.45, 208.64], [363.65, 205.52], [391.95, 205.57], [368.41, 181.05], [337.45, 181.04], [363.65, 180.32], [391.95, 180.34]]}, {\"category\": \"cyclist\", \"corners_3d\": [[126.52, 228.77], [114.23, 228.79], [130.07, 226.94], [141.96, 226.92], [126.52, 197.1], [114.23, 197.11], [130.07, 196.31], [141.96, 196.3]]}]\n```" - } - ] - }, - { - "id": 668, - "data_source": "Mono3DRefer", - "scene_name": "000360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[593.41, 189.25], [616.59, 189.24], [619.02, 192.15], [591.73, 192.15], [593.41, 164.12], [616.59, 164.13], [619.02, 162.58], [591.73, 162.58]]}, {\"category\": \"van\", \"corners_3d\": [[368.41, 208.7], [337.45, 208.64], [363.65, 205.52], [391.95, 205.57], [368.41, 181.05], [337.45, 181.04], [363.65, 180.32], [391.95, 180.34]]}]\n```" - } - ] - }, - { - "id": 669, - "data_source": "Mono3DRefer", - "scene_name": "000360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[777.23, 236.53], [818.3, 233.95], [929.61, 244.41], [886.72, 247.98], [777.23, 183.95], [818.3, 183.5], [929.61, 185.32], [886.72, 185.94]]}, {\"category\": \"car\", \"corners_3d\": [[296.07, 217.28], [261.71, 217.26], [290.37, 213.67], [321.96, 213.68], [296.07, 188.1], [261.71, 188.1], [290.37, 186.86], [321.96, 186.87]]}, {\"category\": \"van\", \"corners_3d\": [[368.41, 208.7], [337.45, 208.64], [363.65, 205.52], [391.95, 205.57], [368.41, 181.05], [337.45, 181.04], [363.65, 180.32], [391.95, 180.34]]}]\n```" - } - ] - }, - { - "id": 670, - "data_source": "Mono3DRefer", - "scene_name": "000360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[593.41, 189.25], [616.59, 189.24], [619.02, 192.15], [591.73, 192.15], [593.41, 164.12], [616.59, 164.13], [619.02, 162.58], [591.73, 162.58]]}, {\"category\": \"cyclist\", \"corners_3d\": [[126.52, 228.77], [114.23, 228.79], [130.07, 226.94], [141.96, 226.92], [126.52, 197.1], [114.23, 197.11], [130.07, 196.31], [141.96, 196.3]]}]\n```" - } - ] - }, - { - "id": 671, - "data_source": "Mono3DRefer", - "scene_name": "000360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[777.23, 236.53], [818.3, 233.95], [929.61, 244.41], [886.72, 247.98], [777.23, 183.95], [818.3, 183.5], [929.61, 185.32], [886.72, 185.94]]}, {\"category\": \"car\", \"corners_3d\": [[296.07, 217.28], [261.71, 217.26], [290.37, 213.67], [321.96, 213.68], [296.07, 188.1], [261.71, 188.1], [290.37, 186.86], [321.96, 186.87]]}, {\"category\": \"cyclist\", \"corners_3d\": [[126.52, 228.77], [114.23, 228.79], [130.07, 226.94], [141.96, 226.92], [126.52, 197.1], [114.23, 197.11], [130.07, 196.31], [141.96, 196.3]]}]\n```" - } - ] - }, - { - "id": 672, - "data_source": "Mono3DRefer", - "scene_name": "000360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[593.41, 189.25], [616.59, 189.24], [619.02, 192.15], [591.73, 192.15], [593.41, 164.12], [616.59, 164.13], [619.02, 162.58], [591.73, 162.58]]}, {\"category\": \"car\", \"corners_3d\": [[777.23, 236.53], [818.3, 233.95], [929.61, 244.41], [886.72, 247.98], [777.23, 183.95], [818.3, 183.5], [929.61, 185.32], [886.72, 185.94]]}, {\"category\": \"car\", \"corners_3d\": [[296.07, 217.28], [261.71, 217.26], [290.37, 213.67], [321.96, 213.68], [296.07, 188.1], [261.71, 188.1], [290.37, 186.86], [321.96, 186.87]]}]\n```" - } - ] - }, - { - "id": 673, - "data_source": "Mono3DRefer", - "scene_name": "000360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[593.41, 189.25], [616.59, 189.24], [619.02, 192.15], [591.73, 192.15], [593.41, 164.12], [616.59, 164.13], [619.02, 162.58], [591.73, 162.58]]}, {\"category\": \"van\", \"corners_3d\": [[368.41, 208.7], [337.45, 208.64], [363.65, 205.52], [391.95, 205.57], [368.41, 181.05], [337.45, 181.04], [363.65, 180.32], [391.95, 180.34]]}, {\"category\": \"cyclist\", \"corners_3d\": [[126.52, 228.77], [114.23, 228.79], [130.07, 226.94], [141.96, 226.92], [126.52, 197.1], [114.23, 197.11], [130.07, 196.31], [141.96, 196.3]]}]\n```" - } - ] - }, - { - "id": 674, - "data_source": "Mono3DRefer", - "scene_name": "000360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[777.23, 236.53], [818.3, 233.95], [929.61, 244.41], [886.72, 247.98], [777.23, 183.95], [818.3, 183.5], [929.61, 185.32], [886.72, 185.94]]}, {\"category\": \"car\", \"corners_3d\": [[296.07, 217.28], [261.71, 217.26], [290.37, 213.67], [321.96, 213.68], [296.07, 188.1], [261.71, 188.1], [290.37, 186.86], [321.96, 186.87]]}, {\"category\": \"van\", \"corners_3d\": [[368.41, 208.7], [337.45, 208.64], [363.65, 205.52], [391.95, 205.57], [368.41, 181.05], [337.45, 181.04], [363.65, 180.32], [391.95, 180.34]]}, {\"category\": \"cyclist\", \"corners_3d\": [[126.52, 228.77], [114.23, 228.79], [130.07, 226.94], [141.96, 226.92], [126.52, 197.1], [114.23, 197.11], [130.07, 196.31], [141.96, 196.3]]}]\n```" - } - ] - }, - { - "id": 675, - "data_source": "Mono3DRefer", - "scene_name": "000360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[593.41, 189.25], [616.59, 189.24], [619.02, 192.15], [591.73, 192.15], [593.41, 164.12], [616.59, 164.13], [619.02, 162.58], [591.73, 162.58]]}, {\"category\": \"car\", \"corners_3d\": [[777.23, 236.53], [818.3, 233.95], [929.61, 244.41], [886.72, 247.98], [777.23, 183.95], [818.3, 183.5], [929.61, 185.32], [886.72, 185.94]]}, {\"category\": \"car\", \"corners_3d\": [[296.07, 217.28], [261.71, 217.26], [290.37, 213.67], [321.96, 213.68], [296.07, 188.1], [261.71, 188.1], [290.37, 186.86], [321.96, 186.87]]}, {\"category\": \"van\", \"corners_3d\": [[368.41, 208.7], [337.45, 208.64], [363.65, 205.52], [391.95, 205.57], [368.41, 181.05], [337.45, 181.04], [363.65, 180.32], [391.95, 180.34]]}]\n```" - } - ] - }, - { - "id": 676, - "data_source": "Mono3DRefer", - "scene_name": "000360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[593.41, 189.25], [616.59, 189.24], [619.02, 192.15], [591.73, 192.15], [593.41, 164.12], [616.59, 164.13], [619.02, 162.58], [591.73, 162.58]]}, {\"category\": \"car\", \"corners_3d\": [[777.23, 236.53], [818.3, 233.95], [929.61, 244.41], [886.72, 247.98], [777.23, 183.95], [818.3, 183.5], [929.61, 185.32], [886.72, 185.94]]}, {\"category\": \"car\", \"corners_3d\": [[296.07, 217.28], [261.71, 217.26], [290.37, 213.67], [321.96, 213.68], [296.07, 188.1], [261.71, 188.1], [290.37, 186.86], [321.96, 186.87]]}, {\"category\": \"cyclist\", \"corners_3d\": [[126.52, 228.77], [114.23, 228.79], [130.07, 226.94], [141.96, 226.92], [126.52, 197.1], [114.23, 197.11], [130.07, 196.31], [141.96, 196.3]]}]\n```" - } - ] - }, - { - "id": 677, - "data_source": "Mono3DRefer", - "scene_name": "000360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[593.41, 189.25], [616.59, 189.24], [619.02, 192.15], [591.73, 192.15], [593.41, 164.12], [616.59, 164.13], [619.02, 162.58], [591.73, 162.58]]}, {\"category\": \"car\", \"corners_3d\": [[777.23, 236.53], [818.3, 233.95], [929.61, 244.41], [886.72, 247.98], [777.23, 183.95], [818.3, 183.5], [929.61, 185.32], [886.72, 185.94]]}, {\"category\": \"car\", \"corners_3d\": [[296.07, 217.28], [261.71, 217.26], [290.37, 213.67], [321.96, 213.68], [296.07, 188.1], [261.71, 188.1], [290.37, 186.86], [321.96, 186.87]]}, {\"category\": \"van\", \"corners_3d\": [[368.41, 208.7], [337.45, 208.64], [363.65, 205.52], [391.95, 205.57], [368.41, 181.05], [337.45, 181.04], [363.65, 180.32], [391.95, 180.34]]}, {\"category\": \"cyclist\", \"corners_3d\": [[126.52, 228.77], [114.23, 228.79], [130.07, 226.94], [141.96, 226.92], [126.52, 197.1], [114.23, 197.11], [130.07, 196.31], [141.96, 196.3]]}]\n```" - } - ] - }, - { - "id": 678, - "data_source": "Mono3DRefer", - "scene_name": "000363", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[536.11, 216.0], [497.44, 215.95], [512.71, 210.73], [546.7, 210.76], [536.11, 177.69], [497.44, 177.69], [512.71, 177.1], [546.7, 177.11]]}, {\"category\": \"car\", \"corners_3d\": [[568.61, 201.32], [546.74, 201.33], [550.54, 199.37], [570.9, 199.36], [568.61, 179.9], [546.74, 179.9], [550.54, 179.41], [570.9, 179.41]]}, {\"category\": \"car\", \"corners_3d\": [[579.55, 192.58], [563.88, 192.58], [565.44, 191.77], [580.47, 191.76], [579.55, 178.27], [563.88, 178.27], [565.44, 178.04], [580.47, 178.04]]}, {\"category\": \"car\", \"corners_3d\": [[591.81, 213.63], [629.57, 213.61], [634.08, 220.49], [589.94, 220.52], [591.81, 176.09], [629.57, 176.09], [634.08, 176.63], [589.94, 176.64]]}]\n```" - } - ] - }, - { - "id": 679, - "data_source": "Mono3DRefer", - "scene_name": "000364", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[829.37, 260.65], [781.02, 258.12], [786.58, 255.56], [833.61, 257.94], [829.37, 170.78], [781.02, 170.84], [786.58, 170.9], [833.61, 170.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[762.25, 245.68], [731.44, 244.72], [734.6, 242.77], [764.61, 243.68], [762.25, 180.84], [731.44, 180.73], [734.6, 180.52], [764.61, 180.62]]}]\n```" - } - ] - }, - { - "id": 680, - "data_source": "Mono3DRefer", - "scene_name": "000364", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[1182.05, 340.69], [1263.88, 341.08], [1350.0, 363.75], [1256.95, 363.25], [1182.05, 214.01], [1263.88, 214.11], [1350.0, 219.67], [1256.95, 219.54]]}]\n```" - } - ] - }, - { - "id": 681, - "data_source": "Mono3DRefer", - "scene_name": "000364", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[829.37, 260.65], [781.02, 258.12], [786.58, 255.56], [833.61, 257.94], [829.37, 170.78], [781.02, 170.84], [786.58, 170.9], [833.61, 170.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[762.25, 245.68], [731.44, 244.72], [734.6, 242.77], [764.61, 243.68], [762.25, 180.84], [731.44, 180.73], [734.6, 180.52], [764.61, 180.62]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1182.05, 340.69], [1263.88, 341.08], [1350.0, 363.75], [1256.95, 363.25], [1182.05, 214.01], [1263.88, 214.11], [1350.0, 219.67], [1256.95, 219.54]]}]\n```" - } - ] - }, - { - "id": 682, - "data_source": "Mono3DRefer", - "scene_name": "000365", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[496.68, 316.12], [493.02, 325.91], [360.35, 328.07], [373.12, 318.0], [496.68, 159.74], [493.02, 158.24], [360.35, 157.91], [373.12, 159.45]]}]\n```" - } - ] - }, - { - "id": 683, - "data_source": "Mono3DRefer", - "scene_name": "000367", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[529.84, 203.78], [500.41, 203.82], [507.69, 201.22], [534.65, 201.19], [529.84, 179.33], [500.41, 179.33], [507.69, 178.79], [534.65, 178.78]]}]\n```" - } - ] - }, - { - "id": 684, - "data_source": "Mono3DRefer", - "scene_name": "000372", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[673.1, 217.75], [743.11, 217.79], [848.05, 255.29], [719.53, 255.13], [673.1, 142.02], [743.11, 141.99], [848.05, 116.24], [719.53, 116.35]]}]\n```" - } - ] - }, - { - "id": 685, - "data_source": "Mono3DRefer", - "scene_name": "000372", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[391.95, 208.56], [364.29, 208.52], [386.26, 205.57], [411.66, 205.61], [391.95, 185.73], [364.29, 185.71], [386.26, 184.65], [411.66, 184.66]]}, {\"category\": \"car\", \"corners_3d\": [[78.9, 264.13], [6.17, 264.04], [112.08, 248.21], [172.24, 248.27], [78.9, 200.7], [6.17, 200.67], [112.08, 195.84], [172.24, 195.86]]}]\n```" - } - ] - }, - { - "id": 686, - "data_source": "Mono3DRefer", - "scene_name": "000372", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[391.95, 208.56], [364.29, 208.52], [386.26, 205.57], [411.66, 205.61], [391.95, 185.73], [364.29, 185.71], [386.26, 184.65], [411.66, 184.66]]}, {\"category\": \"bus\", \"corners_3d\": [[673.1, 217.75], [743.11, 217.79], [848.05, 255.29], [719.53, 255.13], [673.1, 142.02], [743.11, 141.99], [848.05, 116.24], [719.53, 116.35]]}, {\"category\": \"car\", \"corners_3d\": [[78.9, 264.13], [6.17, 264.04], [112.08, 248.21], [172.24, 248.27], [78.9, 200.7], [6.17, 200.67], [112.08, 195.84], [172.24, 195.86]]}]\n```" - } - ] - }, - { - "id": 687, - "data_source": "Mono3DRefer", - "scene_name": "000375", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[284.93, 295.68], [154.45, 295.06], [281.6, 264.14], [376.91, 264.47], [284.93, 189.62], [154.45, 189.57], [281.6, 187.38], [376.91, 187.4]]}]\n```" - } - ] - }, - { - "id": 688, - "data_source": "Mono3DRefer", - "scene_name": "000379", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[27.74, 260.05], [-43.22, 260.06], [57.38, 246.61], [117.39, 246.6], [27.74, 198.25], [-43.22, 198.25], [57.38, 194.33], [117.39, 194.33]]}, {\"category\": \"car\", \"corners_3d\": [[367.71, 209.58], [341.13, 209.54], [362.2, 206.87], [386.87, 206.9], [367.71, 188.79], [341.13, 188.77], [362.2, 187.61], [386.87, 187.63]]}, {\"category\": \"car\", \"corners_3d\": [[489.37, 195.85], [472.32, 195.86], [478.66, 194.73], [494.88, 194.73], [489.37, 180.65], [472.32, 180.65], [478.66, 180.27], [494.88, 180.26]]}]\n```" - } - ] - }, - { - "id": 689, - "data_source": "Mono3DRefer", - "scene_name": "000379", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[698.66, 196.39], [706.35, 196.39], [709.51, 197.16], [701.57, 197.16], [698.66, 169.98], [706.35, 169.98], [709.51, 169.89], [701.57, 169.89]]}]\n```" - } - ] - }, - { - "id": 690, - "data_source": "Mono3DRefer", - "scene_name": "000379", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[27.74, 260.05], [-43.22, 260.06], [57.38, 246.61], [117.39, 246.6], [27.74, 198.25], [-43.22, 198.25], [57.38, 194.33], [117.39, 194.33]]}, {\"category\": \"car\", \"corners_3d\": [[367.71, 209.58], [341.13, 209.54], [362.2, 206.87], [386.87, 206.9], [367.71, 188.79], [341.13, 188.77], [362.2, 187.61], [386.87, 187.63]]}, {\"category\": \"car\", \"corners_3d\": [[489.37, 195.85], [472.32, 195.86], [478.66, 194.73], [494.88, 194.73], [489.37, 180.65], [472.32, 180.65], [478.66, 180.27], [494.88, 180.26]]}, {\"category\": \"cyclist\", \"corners_3d\": [[698.66, 196.39], [706.35, 196.39], [709.51, 197.16], [701.57, 197.16], [698.66, 169.98], [706.35, 169.98], [709.51, 169.89], [701.57, 169.89]]}]\n```" - } - ] - }, - { - "id": 691, - "data_source": "Mono3DRefer", - "scene_name": "000380", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.08, 232.01], [664.32, 232.01], [675.18, 243.87], [602.86, 243.87], [604.08, 183.53], [664.32, 183.53], [675.18, 185.68], [602.86, 185.67]]}, {\"category\": \"car\", \"corners_3d\": [[1111.62, 219.61], [1075.12, 216.53], [1191.03, 216.0], [1235.14, 219.01], [1111.62, 175.29], [1075.12, 175.13], [1191.03, 175.1], [1235.14, 175.26]]}, {\"category\": \"car\", \"corners_3d\": [[398.71, 217.22], [364.43, 217.21], [388.74, 212.92], [419.72, 212.94], [398.71, 185.44], [364.43, 185.44], [388.74, 184.22], [419.72, 184.22]]}, {\"category\": \"car\", \"corners_3d\": [[587.83, 217.51], [626.28, 217.51], [628.38, 223.33], [584.93, 223.33], [587.83, 182.52], [626.28, 182.52], [628.38, 183.77], [584.93, 183.77]]}]\n```" - } - ] - }, - { - "id": 692, - "data_source": "Mono3DRefer", - "scene_name": "000380", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[654.88, 207.41], [688.99, 207.46], [696.2, 211.82], [657.78, 211.76], [654.88, 168.64], [688.99, 168.64], [696.2, 168.1], [657.78, 168.11]]}, {\"category\": \"van\", \"corners_3d\": [[475.76, 202.0], [455.72, 202.0], [466.19, 200.01], [484.86, 200.01], [475.76, 180.04], [455.72, 180.04], [466.19, 179.55], [484.86, 179.55]]}]\n```" - } - ] - }, - { - "id": 693, - "data_source": "Mono3DRefer", - "scene_name": "000380", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[764.15, 217.01], [760.93, 215.72], [811.21, 215.9], [815.97, 217.21], [764.15, 174.37], [760.93, 174.32], [811.21, 174.33], [815.97, 174.37]]}]\n```" - } - ] - }, - { - "id": 694, - "data_source": "Mono3DRefer", - "scene_name": "000380", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1041.98, 250.53], [1014.28, 250.59], [990.27, 246.14], [1016.4, 246.09], [1041.98, 163.49], [1014.28, 163.49], [990.27, 164.02], [1016.4, 164.03]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[543.52, 217.5], [544.11, 218.43], [528.58, 218.56], [528.31, 217.62], [543.52, 174.01], [544.11, 174.03], [528.58, 174.04], [528.31, 174.01]]}]\n```" - } - ] - }, - { - "id": 695, - "data_source": "Mono3DRefer", - "scene_name": "000380", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.08, 232.01], [664.32, 232.01], [675.18, 243.87], [602.86, 243.87], [604.08, 183.53], [664.32, 183.53], [675.18, 185.68], [602.86, 185.67]]}, {\"category\": \"car\", \"corners_3d\": [[1111.62, 219.61], [1075.12, 216.53], [1191.03, 216.0], [1235.14, 219.01], [1111.62, 175.29], [1075.12, 175.13], [1191.03, 175.1], [1235.14, 175.26]]}, {\"category\": \"car\", \"corners_3d\": [[398.71, 217.22], [364.43, 217.21], [388.74, 212.92], [419.72, 212.94], [398.71, 185.44], [364.43, 185.44], [388.74, 184.22], [419.72, 184.22]]}, {\"category\": \"van\", \"corners_3d\": [[654.88, 207.41], [688.99, 207.46], [696.2, 211.82], [657.78, 211.76], [654.88, 168.64], [688.99, 168.64], [696.2, 168.1], [657.78, 168.11]]}, {\"category\": \"van\", \"corners_3d\": [[475.76, 202.0], [455.72, 202.0], [466.19, 200.01], [484.86, 200.01], [475.76, 180.04], [455.72, 180.04], [466.19, 179.55], [484.86, 179.55]]}, {\"category\": \"car\", \"corners_3d\": [[587.83, 217.51], [626.28, 217.51], [628.38, 223.33], [584.93, 223.33], [587.83, 182.52], [626.28, 182.52], [628.38, 183.77], [584.93, 183.77]]}]\n```" - } - ] - }, - { - "id": 696, - "data_source": "Mono3DRefer", - "scene_name": "000380", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.08, 232.01], [664.32, 232.01], [675.18, 243.87], [602.86, 243.87], [604.08, 183.53], [664.32, 183.53], [675.18, 185.68], [602.86, 185.67]]}, {\"category\": \"car\", \"corners_3d\": [[1111.62, 219.61], [1075.12, 216.53], [1191.03, 216.0], [1235.14, 219.01], [1111.62, 175.29], [1075.12, 175.13], [1191.03, 175.1], [1235.14, 175.26]]}, {\"category\": \"car\", \"corners_3d\": [[398.71, 217.22], [364.43, 217.21], [388.74, 212.92], [419.72, 212.94], [398.71, 185.44], [364.43, 185.44], [388.74, 184.22], [419.72, 184.22]]}, {\"category\": \"cyclist\", \"corners_3d\": [[764.15, 217.01], [760.93, 215.72], [811.21, 215.9], [815.97, 217.21], [764.15, 174.37], [760.93, 174.32], [811.21, 174.33], [815.97, 174.37]]}, {\"category\": \"car\", \"corners_3d\": [[587.83, 217.51], [626.28, 217.51], [628.38, 223.33], [584.93, 223.33], [587.83, 182.52], [626.28, 182.52], [628.38, 183.77], [584.93, 183.77]]}]\n```" - } - ] - }, - { - "id": 697, - "data_source": "Mono3DRefer", - "scene_name": "000380", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.08, 232.01], [664.32, 232.01], [675.18, 243.87], [602.86, 243.87], [604.08, 183.53], [664.32, 183.53], [675.18, 185.68], [602.86, 185.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1041.98, 250.53], [1014.28, 250.59], [990.27, 246.14], [1016.4, 246.09], [1041.98, 163.49], [1014.28, 163.49], [990.27, 164.02], [1016.4, 164.03]]}, {\"category\": \"car\", \"corners_3d\": [[1111.62, 219.61], [1075.12, 216.53], [1191.03, 216.0], [1235.14, 219.01], [1111.62, 175.29], [1075.12, 175.13], [1191.03, 175.1], [1235.14, 175.26]]}, {\"category\": \"car\", \"corners_3d\": [[398.71, 217.22], [364.43, 217.21], [388.74, 212.92], [419.72, 212.94], [398.71, 185.44], [364.43, 185.44], [388.74, 184.22], [419.72, 184.22]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[543.52, 217.5], [544.11, 218.43], [528.58, 218.56], [528.31, 217.62], [543.52, 174.01], [544.11, 174.03], [528.58, 174.04], [528.31, 174.01]]}, {\"category\": \"car\", \"corners_3d\": [[587.83, 217.51], [626.28, 217.51], [628.38, 223.33], [584.93, 223.33], [587.83, 182.52], [626.28, 182.52], [628.38, 183.77], [584.93, 183.77]]}]\n```" - } - ] - }, - { - "id": 698, - "data_source": "Mono3DRefer", - "scene_name": "000380", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[764.15, 217.01], [760.93, 215.72], [811.21, 215.9], [815.97, 217.21], [764.15, 174.37], [760.93, 174.32], [811.21, 174.33], [815.97, 174.37]]}, {\"category\": \"van\", \"corners_3d\": [[654.88, 207.41], [688.99, 207.46], [696.2, 211.82], [657.78, 211.76], [654.88, 168.64], [688.99, 168.64], [696.2, 168.1], [657.78, 168.11]]}, {\"category\": \"van\", \"corners_3d\": [[475.76, 202.0], [455.72, 202.0], [466.19, 200.01], [484.86, 200.01], [475.76, 180.04], [455.72, 180.04], [466.19, 179.55], [484.86, 179.55]]}]\n```" - } - ] - }, - { - "id": 699, - "data_source": "Mono3DRefer", - "scene_name": "000380", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1041.98, 250.53], [1014.28, 250.59], [990.27, 246.14], [1016.4, 246.09], [1041.98, 163.49], [1014.28, 163.49], [990.27, 164.02], [1016.4, 164.03]]}, {\"category\": \"van\", \"corners_3d\": [[654.88, 207.41], [688.99, 207.46], [696.2, 211.82], [657.78, 211.76], [654.88, 168.64], [688.99, 168.64], [696.2, 168.1], [657.78, 168.11]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[543.52, 217.5], [544.11, 218.43], [528.58, 218.56], [528.31, 217.62], [543.52, 174.01], [544.11, 174.03], [528.58, 174.04], [528.31, 174.01]]}, {\"category\": \"van\", \"corners_3d\": [[475.76, 202.0], [455.72, 202.0], [466.19, 200.01], [484.86, 200.01], [475.76, 180.04], [455.72, 180.04], [466.19, 179.55], [484.86, 179.55]]}]\n```" - } - ] - }, - { - "id": 700, - "data_source": "Mono3DRefer", - "scene_name": "000380", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1041.98, 250.53], [1014.28, 250.59], [990.27, 246.14], [1016.4, 246.09], [1041.98, 163.49], [1014.28, 163.49], [990.27, 164.02], [1016.4, 164.03]]}, {\"category\": \"cyclist\", \"corners_3d\": [[764.15, 217.01], [760.93, 215.72], [811.21, 215.9], [815.97, 217.21], [764.15, 174.37], [760.93, 174.32], [811.21, 174.33], [815.97, 174.37]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[543.52, 217.5], [544.11, 218.43], [528.58, 218.56], [528.31, 217.62], [543.52, 174.01], [544.11, 174.03], [528.58, 174.04], [528.31, 174.01]]}]\n```" - } - ] - }, - { - "id": 701, - "data_source": "Mono3DRefer", - "scene_name": "000380", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.08, 232.01], [664.32, 232.01], [675.18, 243.87], [602.86, 243.87], [604.08, 183.53], [664.32, 183.53], [675.18, 185.68], [602.86, 185.67]]}, {\"category\": \"car\", \"corners_3d\": [[1111.62, 219.61], [1075.12, 216.53], [1191.03, 216.0], [1235.14, 219.01], [1111.62, 175.29], [1075.12, 175.13], [1191.03, 175.1], [1235.14, 175.26]]}, {\"category\": \"car\", \"corners_3d\": [[398.71, 217.22], [364.43, 217.21], [388.74, 212.92], [419.72, 212.94], [398.71, 185.44], [364.43, 185.44], [388.74, 184.22], [419.72, 184.22]]}, {\"category\": \"cyclist\", \"corners_3d\": [[764.15, 217.01], [760.93, 215.72], [811.21, 215.9], [815.97, 217.21], [764.15, 174.37], [760.93, 174.32], [811.21, 174.33], [815.97, 174.37]]}, {\"category\": \"van\", \"corners_3d\": [[654.88, 207.41], [688.99, 207.46], [696.2, 211.82], [657.78, 211.76], [654.88, 168.64], [688.99, 168.64], [696.2, 168.1], [657.78, 168.11]]}, {\"category\": \"van\", \"corners_3d\": [[475.76, 202.0], [455.72, 202.0], [466.19, 200.01], [484.86, 200.01], [475.76, 180.04], [455.72, 180.04], [466.19, 179.55], [484.86, 179.55]]}, {\"category\": \"car\", \"corners_3d\": [[587.83, 217.51], [626.28, 217.51], [628.38, 223.33], [584.93, 223.33], [587.83, 182.52], [626.28, 182.52], [628.38, 183.77], [584.93, 183.77]]}]\n```" - } - ] - }, - { - "id": 702, - "data_source": "Mono3DRefer", - "scene_name": "000380", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.08, 232.01], [664.32, 232.01], [675.18, 243.87], [602.86, 243.87], [604.08, 183.53], [664.32, 183.53], [675.18, 185.68], [602.86, 185.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1041.98, 250.53], [1014.28, 250.59], [990.27, 246.14], [1016.4, 246.09], [1041.98, 163.49], [1014.28, 163.49], [990.27, 164.02], [1016.4, 164.03]]}, {\"category\": \"car\", \"corners_3d\": [[1111.62, 219.61], [1075.12, 216.53], [1191.03, 216.0], [1235.14, 219.01], [1111.62, 175.29], [1075.12, 175.13], [1191.03, 175.1], [1235.14, 175.26]]}, {\"category\": \"car\", \"corners_3d\": [[398.71, 217.22], [364.43, 217.21], [388.74, 212.92], [419.72, 212.94], [398.71, 185.44], [364.43, 185.44], [388.74, 184.22], [419.72, 184.22]]}, {\"category\": \"van\", \"corners_3d\": [[654.88, 207.41], [688.99, 207.46], [696.2, 211.82], [657.78, 211.76], [654.88, 168.64], [688.99, 168.64], [696.2, 168.1], [657.78, 168.11]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[543.52, 217.5], [544.11, 218.43], [528.58, 218.56], [528.31, 217.62], [543.52, 174.01], [544.11, 174.03], [528.58, 174.04], [528.31, 174.01]]}, {\"category\": \"van\", \"corners_3d\": [[475.76, 202.0], [455.72, 202.0], [466.19, 200.01], [484.86, 200.01], [475.76, 180.04], [455.72, 180.04], [466.19, 179.55], [484.86, 179.55]]}, {\"category\": \"car\", \"corners_3d\": [[587.83, 217.51], [626.28, 217.51], [628.38, 223.33], [584.93, 223.33], [587.83, 182.52], [626.28, 182.52], [628.38, 183.77], [584.93, 183.77]]}]\n```" - } - ] - }, - { - "id": 703, - "data_source": "Mono3DRefer", - "scene_name": "000380", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.08, 232.01], [664.32, 232.01], [675.18, 243.87], [602.86, 243.87], [604.08, 183.53], [664.32, 183.53], [675.18, 185.68], [602.86, 185.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1041.98, 250.53], [1014.28, 250.59], [990.27, 246.14], [1016.4, 246.09], [1041.98, 163.49], [1014.28, 163.49], [990.27, 164.02], [1016.4, 164.03]]}, {\"category\": \"car\", \"corners_3d\": [[1111.62, 219.61], [1075.12, 216.53], [1191.03, 216.0], [1235.14, 219.01], [1111.62, 175.29], [1075.12, 175.13], [1191.03, 175.1], [1235.14, 175.26]]}, {\"category\": \"car\", \"corners_3d\": [[398.71, 217.22], [364.43, 217.21], [388.74, 212.92], [419.72, 212.94], [398.71, 185.44], [364.43, 185.44], [388.74, 184.22], [419.72, 184.22]]}, {\"category\": \"cyclist\", \"corners_3d\": [[764.15, 217.01], [760.93, 215.72], [811.21, 215.9], [815.97, 217.21], [764.15, 174.37], [760.93, 174.32], [811.21, 174.33], [815.97, 174.37]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[543.52, 217.5], [544.11, 218.43], [528.58, 218.56], [528.31, 217.62], [543.52, 174.01], [544.11, 174.03], [528.58, 174.04], [528.31, 174.01]]}, {\"category\": \"car\", \"corners_3d\": [[587.83, 217.51], [626.28, 217.51], [628.38, 223.33], [584.93, 223.33], [587.83, 182.52], [626.28, 182.52], [628.38, 183.77], [584.93, 183.77]]}]\n```" - } - ] - }, - { - "id": 704, - "data_source": "Mono3DRefer", - "scene_name": "000380", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1041.98, 250.53], [1014.28, 250.59], [990.27, 246.14], [1016.4, 246.09], [1041.98, 163.49], [1014.28, 163.49], [990.27, 164.02], [1016.4, 164.03]]}, {\"category\": \"cyclist\", \"corners_3d\": [[764.15, 217.01], [760.93, 215.72], [811.21, 215.9], [815.97, 217.21], [764.15, 174.37], [760.93, 174.32], [811.21, 174.33], [815.97, 174.37]]}, {\"category\": \"van\", \"corners_3d\": [[654.88, 207.41], [688.99, 207.46], [696.2, 211.82], [657.78, 211.76], [654.88, 168.64], [688.99, 168.64], [696.2, 168.1], [657.78, 168.11]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[543.52, 217.5], [544.11, 218.43], [528.58, 218.56], [528.31, 217.62], [543.52, 174.01], [544.11, 174.03], [528.58, 174.04], [528.31, 174.01]]}, {\"category\": \"van\", \"corners_3d\": [[475.76, 202.0], [455.72, 202.0], [466.19, 200.01], [484.86, 200.01], [475.76, 180.04], [455.72, 180.04], [466.19, 179.55], [484.86, 179.55]]}]\n```" - } - ] - }, - { - "id": 705, - "data_source": "Mono3DRefer", - "scene_name": "000380", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.08, 232.01], [664.32, 232.01], [675.18, 243.87], [602.86, 243.87], [604.08, 183.53], [664.32, 183.53], [675.18, 185.68], [602.86, 185.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1041.98, 250.53], [1014.28, 250.59], [990.27, 246.14], [1016.4, 246.09], [1041.98, 163.49], [1014.28, 163.49], [990.27, 164.02], [1016.4, 164.03]]}, {\"category\": \"car\", \"corners_3d\": [[1111.62, 219.61], [1075.12, 216.53], [1191.03, 216.0], [1235.14, 219.01], [1111.62, 175.29], [1075.12, 175.13], [1191.03, 175.1], [1235.14, 175.26]]}, {\"category\": \"car\", \"corners_3d\": [[398.71, 217.22], [364.43, 217.21], [388.74, 212.92], [419.72, 212.94], [398.71, 185.44], [364.43, 185.44], [388.74, 184.22], [419.72, 184.22]]}, {\"category\": \"cyclist\", \"corners_3d\": [[764.15, 217.01], [760.93, 215.72], [811.21, 215.9], [815.97, 217.21], [764.15, 174.37], [760.93, 174.32], [811.21, 174.33], [815.97, 174.37]]}, {\"category\": \"van\", \"corners_3d\": [[654.88, 207.41], [688.99, 207.46], [696.2, 211.82], [657.78, 211.76], [654.88, 168.64], [688.99, 168.64], [696.2, 168.1], [657.78, 168.11]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[543.52, 217.5], [544.11, 218.43], [528.58, 218.56], [528.31, 217.62], [543.52, 174.01], [544.11, 174.03], [528.58, 174.04], [528.31, 174.01]]}, {\"category\": \"van\", \"corners_3d\": [[475.76, 202.0], [455.72, 202.0], [466.19, 200.01], [484.86, 200.01], [475.76, 180.04], [455.72, 180.04], [466.19, 179.55], [484.86, 179.55]]}, {\"category\": \"car\", \"corners_3d\": [[587.83, 217.51], [626.28, 217.51], [628.38, 223.33], [584.93, 223.33], [587.83, 182.52], [626.28, 182.52], [628.38, 183.77], [584.93, 183.77]]}]\n```" - } - ] - }, - { - "id": 706, - "data_source": "Mono3DRefer", - "scene_name": "000383", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1219.24, 187.39], [1236.39, 187.56], [1188.04, 187.61], [1171.95, 187.44], [1219.24, 137.71], [1236.39, 136.64], [1188.04, 136.31], [1171.95, 137.4]]}, {\"category\": \"cyclist\", \"corners_3d\": [[970.14, 208.77], [1000.66, 209.3], [1000.69, 211.58], [967.8, 210.96], [970.14, 150.08], [1000.66, 149.51], [1000.69, 147.05], [967.8, 147.72]]}, {\"category\": \"cyclist\", \"corners_3d\": [[532.6, 264.09], [540.51, 268.57], [429.4, 272.74], [427.05, 267.83], [532.6, 163.9], [540.51, 163.01], [429.4, 162.18], [427.05, 163.16]]}]\n```" - } - ] - }, - { - "id": 707, - "data_source": "Mono3DRefer", - "scene_name": "000383", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[801.49, 266.36], [808.8, 271.96], [738.37, 270.9], [735.4, 265.42], [801.49, 150.69], [808.8, 148.74], [738.37, 149.11], [735.4, 151.01]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[504.81, 279.03], [498.14, 284.84], [420.4, 284.61], [431.38, 278.83], [504.81, 158.31], [498.14, 157.0], [420.4, 157.05], [431.38, 158.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[511.84, 286.53], [504.89, 293.95], [420.73, 293.81], [433.18, 286.41], [511.84, 150.04], [504.89, 147.91], [420.73, 147.95], [433.18, 150.08]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[152.3, 266.1], [128.65, 269.68], [84.6, 268.67], [109.75, 265.18], [152.3, 176.23], [128.65, 176.06], [84.6, 176.11], [109.75, 176.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[131.61, 269.56], [110.53, 272.93], [62.03, 272.1], [84.68, 268.79], [131.61, 182.03], [110.53, 182.09], [62.03, 182.08], [84.68, 182.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[-35.73, 283.27], [-3.41, 278.65], [38.97, 279.38], [8.4, 284.07], [-35.73, 166.65], [-3.41, 167.27], [38.97, 167.17], [8.4, 166.54]]}]\n```" - } - ] - }, - { - "id": 708, - "data_source": "Mono3DRefer", - "scene_name": "000383", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[158.71, 361.19], [354.98, 361.23], [16.19, 606.47], [-446.09, 606.25], [158.71, 195.75], [354.98, 195.76], [16.19, 216.45], [-446.09, 216.43]]}, {\"category\": \"car\", \"corners_3d\": [[1333.69, 194.45], [1396.0, 195.66], [1250.3, 195.6], [1199.85, 194.39], [1333.69, 151.79], [1396.0, 149.29], [1250.3, 149.41], [1199.85, 151.89]]}, {\"category\": \"car\", \"corners_3d\": [[481.35, 191.77], [452.62, 191.82], [459.02, 190.83], [485.21, 190.79], [481.35, 168.38], [452.62, 168.32], [459.02, 169.39], [485.21, 169.44]]}]\n```" - } - ] - }, - { - "id": 709, - "data_source": "Mono3DRefer", - "scene_name": "000383", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[801.49, 266.36], [808.8, 271.96], [738.37, 270.9], [735.4, 265.42], [801.49, 150.69], [808.8, 148.74], [738.37, 149.11], [735.4, 151.01]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1219.24, 187.39], [1236.39, 187.56], [1188.04, 187.61], [1171.95, 187.44], [1219.24, 137.71], [1236.39, 136.64], [1188.04, 136.31], [1171.95, 137.4]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[504.81, 279.03], [498.14, 284.84], [420.4, 284.61], [431.38, 278.83], [504.81, 158.31], [498.14, 157.0], [420.4, 157.05], [431.38, 158.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[511.84, 286.53], [504.89, 293.95], [420.73, 293.81], [433.18, 286.41], [511.84, 150.04], [504.89, 147.91], [420.73, 147.95], [433.18, 150.08]]}, {\"category\": \"cyclist\", \"corners_3d\": [[970.14, 208.77], [1000.66, 209.3], [1000.69, 211.58], [967.8, 210.96], [970.14, 150.08], [1000.66, 149.51], [1000.69, 147.05], [967.8, 147.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[152.3, 266.1], [128.65, 269.68], [84.6, 268.67], [109.75, 265.18], [152.3, 176.23], [128.65, 176.06], [84.6, 176.11], [109.75, 176.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[131.61, 269.56], [110.53, 272.93], [62.03, 272.1], [84.68, 268.79], [131.61, 182.03], [110.53, 182.09], [62.03, 182.08], [84.68, 182.02]]}, {\"category\": \"cyclist\", \"corners_3d\": [[532.6, 264.09], [540.51, 268.57], [429.4, 272.74], [427.05, 267.83], [532.6, 163.9], [540.51, 163.01], [429.4, 162.18], [427.05, 163.16]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[-35.73, 283.27], [-3.41, 278.65], [38.97, 279.38], [8.4, 284.07], [-35.73, 166.65], [-3.41, 167.27], [38.97, 167.17], [8.4, 166.54]]}]\n```" - } - ] - }, - { - "id": 710, - "data_source": "Mono3DRefer", - "scene_name": "000383", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[158.71, 361.19], [354.98, 361.23], [16.19, 606.47], [-446.09, 606.25], [158.71, 195.75], [354.98, 195.76], [16.19, 216.45], [-446.09, 216.43]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1219.24, 187.39], [1236.39, 187.56], [1188.04, 187.61], [1171.95, 187.44], [1219.24, 137.71], [1236.39, 136.64], [1188.04, 136.31], [1171.95, 137.4]]}, {\"category\": \"cyclist\", \"corners_3d\": [[970.14, 208.77], [1000.66, 209.3], [1000.69, 211.58], [967.8, 210.96], [970.14, 150.08], [1000.66, 149.51], [1000.69, 147.05], [967.8, 147.72]]}, {\"category\": \"cyclist\", \"corners_3d\": [[532.6, 264.09], [540.51, 268.57], [429.4, 272.74], [427.05, 267.83], [532.6, 163.9], [540.51, 163.01], [429.4, 162.18], [427.05, 163.16]]}, {\"category\": \"car\", \"corners_3d\": [[1333.69, 194.45], [1396.0, 195.66], [1250.3, 195.6], [1199.85, 194.39], [1333.69, 151.79], [1396.0, 149.29], [1250.3, 149.41], [1199.85, 151.89]]}, {\"category\": \"car\", \"corners_3d\": [[481.35, 191.77], [452.62, 191.82], [459.02, 190.83], [485.21, 190.79], [481.35, 168.38], [452.62, 168.32], [459.02, 169.39], [485.21, 169.44]]}]\n```" - } - ] - }, - { - "id": 711, - "data_source": "Mono3DRefer", - "scene_name": "000383", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[158.71, 361.19], [354.98, 361.23], [16.19, 606.47], [-446.09, 606.25], [158.71, 195.75], [354.98, 195.76], [16.19, 216.45], [-446.09, 216.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[801.49, 266.36], [808.8, 271.96], [738.37, 270.9], [735.4, 265.42], [801.49, 150.69], [808.8, 148.74], [738.37, 149.11], [735.4, 151.01]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[504.81, 279.03], [498.14, 284.84], [420.4, 284.61], [431.38, 278.83], [504.81, 158.31], [498.14, 157.0], [420.4, 157.05], [431.38, 158.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[511.84, 286.53], [504.89, 293.95], [420.73, 293.81], [433.18, 286.41], [511.84, 150.04], [504.89, 147.91], [420.73, 147.95], [433.18, 150.08]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[152.3, 266.1], [128.65, 269.68], [84.6, 268.67], [109.75, 265.18], [152.3, 176.23], [128.65, 176.06], [84.6, 176.11], [109.75, 176.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[131.61, 269.56], [110.53, 272.93], [62.03, 272.1], [84.68, 268.79], [131.61, 182.03], [110.53, 182.09], [62.03, 182.08], [84.68, 182.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[-35.73, 283.27], [-3.41, 278.65], [38.97, 279.38], [8.4, 284.07], [-35.73, 166.65], [-3.41, 167.27], [38.97, 167.17], [8.4, 166.54]]}, {\"category\": \"car\", \"corners_3d\": [[1333.69, 194.45], [1396.0, 195.66], [1250.3, 195.6], [1199.85, 194.39], [1333.69, 151.79], [1396.0, 149.29], [1250.3, 149.41], [1199.85, 151.89]]}, {\"category\": \"car\", \"corners_3d\": [[481.35, 191.77], [452.62, 191.82], [459.02, 190.83], [485.21, 190.79], [481.35, 168.38], [452.62, 168.32], [459.02, 169.39], [485.21, 169.44]]}]\n```" - } - ] - }, - { - "id": 712, - "data_source": "Mono3DRefer", - "scene_name": "000383", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[158.71, 361.19], [354.98, 361.23], [16.19, 606.47], [-446.09, 606.25], [158.71, 195.75], [354.98, 195.76], [16.19, 216.45], [-446.09, 216.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[801.49, 266.36], [808.8, 271.96], [738.37, 270.9], [735.4, 265.42], [801.49, 150.69], [808.8, 148.74], [738.37, 149.11], [735.4, 151.01]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1219.24, 187.39], [1236.39, 187.56], [1188.04, 187.61], [1171.95, 187.44], [1219.24, 137.71], [1236.39, 136.64], [1188.04, 136.31], [1171.95, 137.4]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[504.81, 279.03], [498.14, 284.84], [420.4, 284.61], [431.38, 278.83], [504.81, 158.31], [498.14, 157.0], [420.4, 157.05], [431.38, 158.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[511.84, 286.53], [504.89, 293.95], [420.73, 293.81], [433.18, 286.41], [511.84, 150.04], [504.89, 147.91], [420.73, 147.95], [433.18, 150.08]]}, {\"category\": \"cyclist\", \"corners_3d\": [[970.14, 208.77], [1000.66, 209.3], [1000.69, 211.58], [967.8, 210.96], [970.14, 150.08], [1000.66, 149.51], [1000.69, 147.05], [967.8, 147.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[152.3, 266.1], [128.65, 269.68], [84.6, 268.67], [109.75, 265.18], [152.3, 176.23], [128.65, 176.06], [84.6, 176.11], [109.75, 176.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[131.61, 269.56], [110.53, 272.93], [62.03, 272.1], [84.68, 268.79], [131.61, 182.03], [110.53, 182.09], [62.03, 182.08], [84.68, 182.02]]}, {\"category\": \"cyclist\", \"corners_3d\": [[532.6, 264.09], [540.51, 268.57], [429.4, 272.74], [427.05, 267.83], [532.6, 163.9], [540.51, 163.01], [429.4, 162.18], [427.05, 163.16]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[-35.73, 283.27], [-3.41, 278.65], [38.97, 279.38], [8.4, 284.07], [-35.73, 166.65], [-3.41, 167.27], [38.97, 167.17], [8.4, 166.54]]}, {\"category\": \"car\", \"corners_3d\": [[1333.69, 194.45], [1396.0, 195.66], [1250.3, 195.6], [1199.85, 194.39], [1333.69, 151.79], [1396.0, 149.29], [1250.3, 149.41], [1199.85, 151.89]]}, {\"category\": \"car\", \"corners_3d\": [[481.35, 191.77], [452.62, 191.82], [459.02, 190.83], [485.21, 190.79], [481.35, 168.38], [452.62, 168.32], [459.02, 169.39], [485.21, 169.44]]}]\n```" - } - ] - }, - { - "id": 713, - "data_source": "Mono3DRefer", - "scene_name": "000385", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[755.84, 298.24], [870.15, 298.25], [1024.41, 372.64], [842.26, 372.6], [755.84, 189.89], [870.15, 189.9], [1024.41, 200.01], [842.26, 200.0]]}, {\"category\": \"car\", \"corners_3d\": [[556.29, 229.16], [506.61, 229.17], [523.23, 220.02], [564.85, 220.02], [556.29, 181.07], [506.61, 181.07], [523.23, 179.74], [564.85, 179.74]]}, {\"category\": \"car\", \"corners_3d\": [[743.87, 239.23], [684.07, 239.34], [668.76, 227.97], [718.36, 227.89], [743.87, 181.95], [684.07, 181.96], [668.76, 180.4], [718.36, 180.39]]}, {\"category\": \"car\", \"corners_3d\": [[582.44, 209.99], [552.51, 210.05], [554.7, 207.1], [582.26, 207.04], [582.44, 182.74], [552.51, 182.76], [554.7, 181.97], [582.26, 181.96]]}, {\"category\": \"car\", \"corners_3d\": [[648.85, 206.2], [677.81, 206.22], [683.32, 209.26], [651.72, 209.24], [648.85, 178.29], [677.81, 178.3], [683.32, 178.79], [651.72, 178.79]]}]\n```" - } - ] - }, - { - "id": 714, - "data_source": "Mono3DRefer", - "scene_name": "000386", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[513.08, 212.72], [546.63, 212.74], [537.12, 216.57], [498.91, 216.55], [513.08, 184.15], [546.63, 184.15], [537.12, 184.0], [498.91, 184.0]]}, {\"category\": \"car\", \"corners_3d\": [[150.58, 239.25], [205.29, 239.22], [109.68, 252.28], [41.65, 252.33], [150.58, 186.58], [205.29, 186.58], [109.68, 186.91], [41.65, 186.91]]}, {\"category\": \"car\", \"corners_3d\": [[572.45, 214.46], [611.47, 214.43], [613.89, 218.35], [569.64, 218.39], [572.45, 178.98], [611.47, 178.99], [613.89, 178.15], [569.64, 178.14]]}, {\"category\": \"car\", \"corners_3d\": [[261.83, 323.79], [410.35, 323.53], [214.68, 465.81], [-87.95, 466.89], [261.83, 182.53], [410.35, 182.54], [214.68, 179.79], [-87.95, 179.77]]}, {\"category\": \"car\", \"corners_3d\": [[117.54, 273.7], [208.8, 273.59], [47.62, 309.95], [-81.52, 310.16], [117.54, 191.92], [208.8, 191.91], [47.62, 194.67], [-81.52, 194.68]]}, {\"category\": \"car\", \"corners_3d\": [[352.97, 225.33], [400.36, 225.31], [363.14, 232.76], [306.91, 232.79], [352.97, 178.35], [400.36, 178.35], [363.14, 177.08], [306.91, 177.07]]}, {\"category\": \"car\", \"corners_3d\": [[455.51, 232.26], [509.05, 232.3], [486.83, 242.17], [422.09, 242.12], [455.51, 183.87], [509.05, 183.87], [486.83, 183.59], [422.09, 183.59]]}, {\"category\": \"car\", \"corners_3d\": [[409.26, 215.17], [447.87, 215.15], [427.36, 219.17], [383.54, 219.19], [409.26, 182.71], [447.87, 182.71], [427.36, 182.38], [383.54, 182.37]]}]\n```" - } - ] - }, - { - "id": 715, - "data_source": "Mono3DRefer", - "scene_name": "000386", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[538.69, 195.54], [560.99, 195.52], [559.25, 196.52], [534.79, 196.53], [538.69, 166.35], [560.99, 166.37], [559.25, 164.55], [534.79, 164.53]]}]\n```" - } - ] - }, - { - "id": 716, - "data_source": "Mono3DRefer", - "scene_name": "000386", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[513.08, 212.72], [546.63, 212.74], [537.12, 216.57], [498.91, 216.55], [513.08, 184.15], [546.63, 184.15], [537.12, 184.0], [498.91, 184.0]]}, {\"category\": \"van\", \"corners_3d\": [[538.69, 195.54], [560.99, 195.52], [559.25, 196.52], [534.79, 196.53], [538.69, 166.35], [560.99, 166.37], [559.25, 164.55], [534.79, 164.53]]}, {\"category\": \"car\", \"corners_3d\": [[150.58, 239.25], [205.29, 239.22], [109.68, 252.28], [41.65, 252.33], [150.58, 186.58], [205.29, 186.58], [109.68, 186.91], [41.65, 186.91]]}, {\"category\": \"car\", \"corners_3d\": [[572.45, 214.46], [611.47, 214.43], [613.89, 218.35], [569.64, 218.39], [572.45, 178.98], [611.47, 178.99], [613.89, 178.15], [569.64, 178.14]]}, {\"category\": \"car\", \"corners_3d\": [[261.83, 323.79], [410.35, 323.53], [214.68, 465.81], [-87.95, 466.89], [261.83, 182.53], [410.35, 182.54], [214.68, 179.79], [-87.95, 179.77]]}, {\"category\": \"car\", \"corners_3d\": [[117.54, 273.7], [208.8, 273.59], [47.62, 309.95], [-81.52, 310.16], [117.54, 191.92], [208.8, 191.91], [47.62, 194.67], [-81.52, 194.68]]}, {\"category\": \"car\", \"corners_3d\": [[352.97, 225.33], [400.36, 225.31], [363.14, 232.76], [306.91, 232.79], [352.97, 178.35], [400.36, 178.35], [363.14, 177.08], [306.91, 177.07]]}, {\"category\": \"car\", \"corners_3d\": [[455.51, 232.26], [509.05, 232.3], [486.83, 242.17], [422.09, 242.12], [455.51, 183.87], [509.05, 183.87], [486.83, 183.59], [422.09, 183.59]]}, {\"category\": \"car\", \"corners_3d\": [[409.26, 215.17], [447.87, 215.15], [427.36, 219.17], [383.54, 219.19], [409.26, 182.71], [447.87, 182.71], [427.36, 182.38], [383.54, 182.37]]}]\n```" - } - ] - }, - { - "id": 717, - "data_source": "Mono3DRefer", - "scene_name": "000387", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[415.23, 213.52], [385.48, 213.5], [409.33, 209.3], [436.01, 209.31], [415.23, 189.0], [385.48, 189.0], [409.33, 187.33], [436.01, 187.33]]}, {\"category\": \"car\", \"corners_3d\": [[484.66, 203.18], [466.23, 203.18], [474.85, 201.43], [492.22, 201.44], [484.66, 186.4], [466.23, 186.39], [474.85, 185.62], [492.22, 185.62]]}, {\"category\": \"car\", \"corners_3d\": [[661.68, 195.55], [674.62, 195.96], [646.97, 196.86], [634.04, 196.42], [661.68, 178.69], [674.62, 178.8], [646.97, 179.03], [634.04, 178.91]]}, {\"category\": \"car\", \"corners_3d\": [[584.81, 230.53], [634.92, 230.57], [638.15, 241.19], [578.83, 241.13], [584.81, 183.17], [634.92, 183.17], [638.15, 185.07], [578.83, 185.06]]}]\n```" - } - ] - }, - { - "id": 718, - "data_source": "Mono3DRefer", - "scene_name": "000388", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[530.46, 195.15], [549.8, 195.13], [548.24, 196.23], [527.95, 196.25], [530.46, 176.64], [549.8, 176.64], [548.24, 176.82], [527.95, 176.83]]}]\n```" - } - ] - }, - { - "id": 719, - "data_source": "Mono3DRefer", - "scene_name": "000388", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[488.56, 192.26], [511.29, 192.23], [493.83, 197.07], [465.39, 197.11], [488.56, 165.22], [511.29, 165.23], [493.83, 163.33], [465.39, 163.32]]}]\n```" - } - ] - }, - { - "id": 720, - "data_source": "Mono3DRefer", - "scene_name": "000388", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[566.08, 191.86], [589.01, 191.83], [591.63, 195.14], [564.69, 195.18], [566.08, 167.01], [589.01, 167.02], [591.63, 166.0], [564.69, 165.99]]}]\n```" - } - ] - }, - { - "id": 721, - "data_source": "Mono3DRefer", - "scene_name": "000388", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[530.46, 195.15], [549.8, 195.13], [548.24, 196.23], [527.95, 196.25], [530.46, 176.64], [549.8, 176.64], [548.24, 176.82], [527.95, 176.83]]}, {\"category\": \"tram\", \"corners_3d\": [[488.56, 192.26], [511.29, 192.23], [493.83, 197.07], [465.39, 197.11], [488.56, 165.22], [511.29, 165.23], [493.83, 163.33], [465.39, 163.32]]}]\n```" - } - ] - }, - { - "id": 722, - "data_source": "Mono3DRefer", - "scene_name": "000388", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[530.46, 195.15], [549.8, 195.13], [548.24, 196.23], [527.95, 196.25], [530.46, 176.64], [549.8, 176.64], [548.24, 176.82], [527.95, 176.83]]}, {\"category\": \"bus\", \"corners_3d\": [[566.08, 191.86], [589.01, 191.83], [591.63, 195.14], [564.69, 195.18], [566.08, 167.01], [589.01, 167.02], [591.63, 166.0], [564.69, 165.99]]}]\n```" - } - ] - }, - { - "id": 723, - "data_source": "Mono3DRefer", - "scene_name": "000388", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[566.08, 191.86], [589.01, 191.83], [591.63, 195.14], [564.69, 195.18], [566.08, 167.01], [589.01, 167.02], [591.63, 166.0], [564.69, 165.99]]}, {\"category\": \"tram\", \"corners_3d\": [[488.56, 192.26], [511.29, 192.23], [493.83, 197.07], [465.39, 197.11], [488.56, 165.22], [511.29, 165.23], [493.83, 163.33], [465.39, 163.32]]}]\n```" - } - ] - }, - { - "id": 724, - "data_source": "Mono3DRefer", - "scene_name": "000388", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[530.46, 195.15], [549.8, 195.13], [548.24, 196.23], [527.95, 196.25], [530.46, 176.64], [549.8, 176.64], [548.24, 176.82], [527.95, 176.83]]}, {\"category\": \"bus\", \"corners_3d\": [[566.08, 191.86], [589.01, 191.83], [591.63, 195.14], [564.69, 195.18], [566.08, 167.01], [589.01, 167.02], [591.63, 166.0], [564.69, 165.99]]}, {\"category\": \"tram\", \"corners_3d\": [[488.56, 192.26], [511.29, 192.23], [493.83, 197.07], [465.39, 197.11], [488.56, 165.22], [511.29, 165.23], [493.83, 163.33], [465.39, 163.32]]}]\n```" - } - ] - }, - { - "id": 725, - "data_source": "Mono3DRefer", - "scene_name": "000390", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[520.08, 225.29], [478.58, 225.65], [483.76, 219.35], [520.3, 219.07], [520.08, 183.13], [478.58, 183.2], [483.76, 181.97], [520.3, 181.91]]}, {\"category\": \"car\", \"corners_3d\": [[499.82, 210.78], [470.95, 211.23], [463.29, 207.78], [489.68, 207.41], [499.82, 186.11], [470.95, 186.26], [463.29, 185.06], [489.68, 184.93]]}]\n```" - } - ] - }, - { - "id": 726, - "data_source": "Mono3DRefer", - "scene_name": "000391", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[287.81, 240.14], [348.51, 240.1], [287.11, 253.47], [211.55, 253.53], [287.81, 185.54], [348.51, 185.54], [287.11, 185.62], [211.55, 185.62]]}, {\"category\": \"car\", \"corners_3d\": [[580.15, 208.45], [614.23, 208.41], [617.46, 211.09], [579.45, 211.13], [580.15, 177.44], [614.23, 177.45], [617.46, 176.56], [579.45, 176.55]]}]\n```" - } - ] - }, - { - "id": 727, - "data_source": "Mono3DRefer", - "scene_name": "000391", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[155.6, 270.43], [259.62, 270.19], [-59.53, 351.45], [-265.27, 352.34], [155.6, 131.19], [259.62, 131.33], [-59.53, 79.81], [-265.27, 79.24]]}]\n```" - } - ] - }, - { - "id": 728, - "data_source": "Mono3DRefer", - "scene_name": "000391", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[543.41, 195.6], [571.08, 195.58], [565.97, 199.08], [528.94, 199.11], [543.41, 150.82], [571.08, 150.86], [565.97, 139.26], [528.94, 139.19]]}]\n```" - } - ] - }, - { - "id": 729, - "data_source": "Mono3DRefer", - "scene_name": "000391", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[155.6, 270.43], [259.62, 270.19], [-59.53, 351.45], [-265.27, 352.34], [155.6, 131.19], [259.62, 131.33], [-59.53, 79.81], [-265.27, 79.24]]}, {\"category\": \"car\", \"corners_3d\": [[287.81, 240.14], [348.51, 240.1], [287.11, 253.47], [211.55, 253.53], [287.81, 185.54], [348.51, 185.54], [287.11, 185.62], [211.55, 185.62]]}, {\"category\": \"car\", \"corners_3d\": [[580.15, 208.45], [614.23, 208.41], [617.46, 211.09], [579.45, 211.13], [580.15, 177.44], [614.23, 177.45], [617.46, 176.56], [579.45, 176.55]]}]\n```" - } - ] - }, - { - "id": 730, - "data_source": "Mono3DRefer", - "scene_name": "000391", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[287.81, 240.14], [348.51, 240.1], [287.11, 253.47], [211.55, 253.53], [287.81, 185.54], [348.51, 185.54], [287.11, 185.62], [211.55, 185.62]]}, {\"category\": \"truck\", \"corners_3d\": [[543.41, 195.6], [571.08, 195.58], [565.97, 199.08], [528.94, 199.11], [543.41, 150.82], [571.08, 150.86], [565.97, 139.26], [528.94, 139.19]]}, {\"category\": \"car\", \"corners_3d\": [[580.15, 208.45], [614.23, 208.41], [617.46, 211.09], [579.45, 211.13], [580.15, 177.44], [614.23, 177.45], [617.46, 176.56], [579.45, 176.55]]}]\n```" - } - ] - }, - { - "id": 731, - "data_source": "Mono3DRefer", - "scene_name": "000391", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[155.6, 270.43], [259.62, 270.19], [-59.53, 351.45], [-265.27, 352.34], [155.6, 131.19], [259.62, 131.33], [-59.53, 79.81], [-265.27, 79.24]]}, {\"category\": \"truck\", \"corners_3d\": [[543.41, 195.6], [571.08, 195.58], [565.97, 199.08], [528.94, 199.11], [543.41, 150.82], [571.08, 150.86], [565.97, 139.26], [528.94, 139.19]]}]\n```" - } - ] - }, - { - "id": 732, - "data_source": "Mono3DRefer", - "scene_name": "000391", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[155.6, 270.43], [259.62, 270.19], [-59.53, 351.45], [-265.27, 352.34], [155.6, 131.19], [259.62, 131.33], [-59.53, 79.81], [-265.27, 79.24]]}, {\"category\": \"car\", \"corners_3d\": [[287.81, 240.14], [348.51, 240.1], [287.11, 253.47], [211.55, 253.53], [287.81, 185.54], [348.51, 185.54], [287.11, 185.62], [211.55, 185.62]]}, {\"category\": \"truck\", \"corners_3d\": [[543.41, 195.6], [571.08, 195.58], [565.97, 199.08], [528.94, 199.11], [543.41, 150.82], [571.08, 150.86], [565.97, 139.26], [528.94, 139.19]]}, {\"category\": \"car\", \"corners_3d\": [[580.15, 208.45], [614.23, 208.41], [617.46, 211.09], [579.45, 211.13], [580.15, 177.44], [614.23, 177.45], [617.46, 176.56], [579.45, 176.55]]}]\n```" - } - ] - }, - { - "id": 733, - "data_source": "Mono3DRefer", - "scene_name": "000394", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[594.13, 188.78], [617.27, 188.77], [621.08, 191.59], [593.85, 191.6], [594.13, 163.7], [617.27, 163.7], [621.08, 162.08], [593.85, 162.08]]}]\n```" - } - ] - }, - { - "id": 734, - "data_source": "Mono3DRefer", - "scene_name": "000394", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[120.05, 272.84], [133.34, 266.19], [177.94, 264.36], [168.04, 270.74], [120.05, 158.99], [133.34, 159.92], [177.94, 160.17], [168.04, 159.28]]}]\n```" - } - ] - }, - { - "id": 735, - "data_source": "Mono3DRefer", - "scene_name": "000394", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[346.29, 216.65], [412.16, 216.61], [272.16, 248.72], [157.69, 248.83], [346.29, 141.18], [412.16, 141.21], [272.16, 117.99], [157.69, 117.91]]}]\n```" - } - ] - }, - { - "id": 736, - "data_source": "Mono3DRefer", - "scene_name": "000394", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[565.23, 189.59], [580.96, 189.59], [579.8, 190.26], [563.43, 190.26], [565.23, 174.49], [580.96, 174.49], [579.8, 174.55], [563.43, 174.55]]}]\n```" - } - ] - }, - { - "id": 737, - "data_source": "Mono3DRefer", - "scene_name": "000394", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[120.05, 272.84], [133.34, 266.19], [177.94, 264.36], [168.04, 270.74], [120.05, 158.99], [133.34, 159.92], [177.94, 160.17], [168.04, 159.28]]}, {\"category\": \"bus\", \"corners_3d\": [[594.13, 188.78], [617.27, 188.77], [621.08, 191.59], [593.85, 191.6], [594.13, 163.7], [617.27, 163.7], [621.08, 162.08], [593.85, 162.08]]}]\n```" - } - ] - }, - { - "id": 738, - "data_source": "Mono3DRefer", - "scene_name": "000394", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[346.29, 216.65], [412.16, 216.61], [272.16, 248.72], [157.69, 248.83], [346.29, 141.18], [412.16, 141.21], [272.16, 117.99], [157.69, 117.91]]}, {\"category\": \"bus\", \"corners_3d\": [[594.13, 188.78], [617.27, 188.77], [621.08, 191.59], [593.85, 191.6], [594.13, 163.7], [617.27, 163.7], [621.08, 162.08], [593.85, 162.08]]}]\n```" - } - ] - }, - { - "id": 739, - "data_source": "Mono3DRefer", - "scene_name": "000394", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[565.23, 189.59], [580.96, 189.59], [579.8, 190.26], [563.43, 190.26], [565.23, 174.49], [580.96, 174.49], [579.8, 174.55], [563.43, 174.55]]}, {\"category\": \"bus\", \"corners_3d\": [[594.13, 188.78], [617.27, 188.77], [621.08, 191.59], [593.85, 191.6], [594.13, 163.7], [617.27, 163.7], [621.08, 162.08], [593.85, 162.08]]}]\n```" - } - ] - }, - { - "id": 740, - "data_source": "Mono3DRefer", - "scene_name": "000394", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[346.29, 216.65], [412.16, 216.61], [272.16, 248.72], [157.69, 248.83], [346.29, 141.18], [412.16, 141.21], [272.16, 117.99], [157.69, 117.91]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[120.05, 272.84], [133.34, 266.19], [177.94, 264.36], [168.04, 270.74], [120.05, 158.99], [133.34, 159.92], [177.94, 160.17], [168.04, 159.28]]}]\n```" - } - ] - }, - { - "id": 741, - "data_source": "Mono3DRefer", - "scene_name": "000394", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[120.05, 272.84], [133.34, 266.19], [177.94, 264.36], [168.04, 270.74], [120.05, 158.99], [133.34, 159.92], [177.94, 160.17], [168.04, 159.28]]}, {\"category\": \"car\", \"corners_3d\": [[565.23, 189.59], [580.96, 189.59], [579.8, 190.26], [563.43, 190.26], [565.23, 174.49], [580.96, 174.49], [579.8, 174.55], [563.43, 174.55]]}]\n```" - } - ] - }, - { - "id": 742, - "data_source": "Mono3DRefer", - "scene_name": "000394", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[346.29, 216.65], [412.16, 216.61], [272.16, 248.72], [157.69, 248.83], [346.29, 141.18], [412.16, 141.21], [272.16, 117.99], [157.69, 117.91]]}, {\"category\": \"car\", \"corners_3d\": [[565.23, 189.59], [580.96, 189.59], [579.8, 190.26], [563.43, 190.26], [565.23, 174.49], [580.96, 174.49], [579.8, 174.55], [563.43, 174.55]]}]\n```" - } - ] - }, - { - "id": 743, - "data_source": "Mono3DRefer", - "scene_name": "000394", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, pedestrian, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[346.29, 216.65], [412.16, 216.61], [272.16, 248.72], [157.69, 248.83], [346.29, 141.18], [412.16, 141.21], [272.16, 117.99], [157.69, 117.91]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[120.05, 272.84], [133.34, 266.19], [177.94, 264.36], [168.04, 270.74], [120.05, 158.99], [133.34, 159.92], [177.94, 160.17], [168.04, 159.28]]}, {\"category\": \"bus\", \"corners_3d\": [[594.13, 188.78], [617.27, 188.77], [621.08, 191.59], [593.85, 191.6], [594.13, 163.7], [617.27, 163.7], [621.08, 162.08], [593.85, 162.08]]}]\n```" - } - ] - }, - { - "id": 744, - "data_source": "Mono3DRefer", - "scene_name": "000394", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[120.05, 272.84], [133.34, 266.19], [177.94, 264.36], [168.04, 270.74], [120.05, 158.99], [133.34, 159.92], [177.94, 160.17], [168.04, 159.28]]}, {\"category\": \"car\", \"corners_3d\": [[565.23, 189.59], [580.96, 189.59], [579.8, 190.26], [563.43, 190.26], [565.23, 174.49], [580.96, 174.49], [579.8, 174.55], [563.43, 174.55]]}, {\"category\": \"bus\", \"corners_3d\": [[594.13, 188.78], [617.27, 188.77], [621.08, 191.59], [593.85, 191.6], [594.13, 163.7], [617.27, 163.7], [621.08, 162.08], [593.85, 162.08]]}]\n```" - } - ] - }, - { - "id": 745, - "data_source": "Mono3DRefer", - "scene_name": "000394", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[346.29, 216.65], [412.16, 216.61], [272.16, 248.72], [157.69, 248.83], [346.29, 141.18], [412.16, 141.21], [272.16, 117.99], [157.69, 117.91]]}, {\"category\": \"car\", \"corners_3d\": [[565.23, 189.59], [580.96, 189.59], [579.8, 190.26], [563.43, 190.26], [565.23, 174.49], [580.96, 174.49], [579.8, 174.55], [563.43, 174.55]]}, {\"category\": \"bus\", \"corners_3d\": [[594.13, 188.78], [617.27, 188.77], [621.08, 191.59], [593.85, 191.6], [594.13, 163.7], [617.27, 163.7], [621.08, 162.08], [593.85, 162.08]]}]\n```" - } - ] - }, - { - "id": 746, - "data_source": "Mono3DRefer", - "scene_name": "000394", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[346.29, 216.65], [412.16, 216.61], [272.16, 248.72], [157.69, 248.83], [346.29, 141.18], [412.16, 141.21], [272.16, 117.99], [157.69, 117.91]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[120.05, 272.84], [133.34, 266.19], [177.94, 264.36], [168.04, 270.74], [120.05, 158.99], [133.34, 159.92], [177.94, 160.17], [168.04, 159.28]]}, {\"category\": \"car\", \"corners_3d\": [[565.23, 189.59], [580.96, 189.59], [579.8, 190.26], [563.43, 190.26], [565.23, 174.49], [580.96, 174.49], [579.8, 174.55], [563.43, 174.55]]}]\n```" - } - ] - }, - { - "id": 747, - "data_source": "Mono3DRefer", - "scene_name": "000394", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, pedestrian, tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[346.29, 216.65], [412.16, 216.61], [272.16, 248.72], [157.69, 248.83], [346.29, 141.18], [412.16, 141.21], [272.16, 117.99], [157.69, 117.91]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[120.05, 272.84], [133.34, 266.19], [177.94, 264.36], [168.04, 270.74], [120.05, 158.99], [133.34, 159.92], [177.94, 160.17], [168.04, 159.28]]}, {\"category\": \"car\", \"corners_3d\": [[565.23, 189.59], [580.96, 189.59], [579.8, 190.26], [563.43, 190.26], [565.23, 174.49], [580.96, 174.49], [579.8, 174.55], [563.43, 174.55]]}, {\"category\": \"bus\", \"corners_3d\": [[594.13, 188.78], [617.27, 188.77], [621.08, 191.59], [593.85, 191.6], [594.13, 163.7], [617.27, 163.7], [621.08, 162.08], [593.85, 162.08]]}]\n```" - } - ] - }, - { - "id": 748, - "data_source": "Mono3DRefer", - "scene_name": "000397", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[461.59, 237.03], [450.22, 242.81], [277.2, 243.31], [303.05, 237.45], [461.59, 184.0], [450.22, 185.01], [277.2, 185.1], [303.05, 184.08]]}, {\"category\": \"car\", \"corners_3d\": [[510.41, 221.28], [504.7, 224.07], [388.8, 224.07], [400.83, 221.28], [510.41, 174.52], [504.7, 174.62], [388.8, 174.62], [400.83, 174.52]]}, {\"category\": \"car\", \"corners_3d\": [[646.85, 213.86], [680.76, 213.87], [689.02, 218.66], [651.14, 218.66], [646.85, 182.59], [680.76, 182.59], [689.02, 183.73], [651.14, 183.73]]}, {\"category\": \"car\", \"corners_3d\": [[891.6, 401.0], [1097.9, 398.41], [2601.77, 1054.96], [1837.07, 1096.08], [891.6, 219.79], [1097.9, 219.25], [2601.77, 354.32], [1837.07, 362.78]]}, {\"category\": \"car\", \"corners_3d\": [[660.21, 226.42], [709.34, 226.24], [737.81, 237.48], [678.42, 237.74], [660.21, 175.45], [709.34, 175.44], [737.81, 175.99], [678.42, 176.0]]}, {\"category\": \"car\", \"corners_3d\": [[249.11, 251.21], [283.94, 244.47], [423.5, 245.35], [401.51, 252.27], [249.11, 185.76], [283.94, 184.65], [423.5, 184.79], [401.51, 185.93]]}]\n```" - } - ] - }, - { - "id": 749, - "data_source": "Mono3DRefer", - "scene_name": "000397", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[122.1, 245.62], [99.86, 245.56], [112.24, 243.86], [133.97, 243.93], [122.1, 169.47], [99.86, 169.48], [112.24, 169.56], [133.97, 169.55]]}]\n```" - } - ] - }, - { - "id": 750, - "data_source": "Mono3DRefer", - "scene_name": "000397", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[461.59, 237.03], [450.22, 242.81], [277.2, 243.31], [303.05, 237.45], [461.59, 184.0], [450.22, 185.01], [277.2, 185.1], [303.05, 184.08]]}, {\"category\": \"car\", \"corners_3d\": [[510.41, 221.28], [504.7, 224.07], [388.8, 224.07], [400.83, 221.28], [510.41, 174.52], [504.7, 174.62], [388.8, 174.62], [400.83, 174.52]]}, {\"category\": \"car\", \"corners_3d\": [[646.85, 213.86], [680.76, 213.87], [689.02, 218.66], [651.14, 218.66], [646.85, 182.59], [680.76, 182.59], [689.02, 183.73], [651.14, 183.73]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[122.1, 245.62], [99.86, 245.56], [112.24, 243.86], [133.97, 243.93], [122.1, 169.47], [99.86, 169.48], [112.24, 169.56], [133.97, 169.55]]}, {\"category\": \"car\", \"corners_3d\": [[891.6, 401.0], [1097.9, 398.41], [2601.77, 1054.96], [1837.07, 1096.08], [891.6, 219.79], [1097.9, 219.25], [2601.77, 354.32], [1837.07, 362.78]]}, {\"category\": \"car\", \"corners_3d\": [[660.21, 226.42], [709.34, 226.24], [737.81, 237.48], [678.42, 237.74], [660.21, 175.45], [709.34, 175.44], [737.81, 175.99], [678.42, 176.0]]}, {\"category\": \"car\", \"corners_3d\": [[249.11, 251.21], [283.94, 244.47], [423.5, 245.35], [401.51, 252.27], [249.11, 185.76], [283.94, 184.65], [423.5, 184.79], [401.51, 185.93]]}]\n```" - } - ] - }, - { - "id": 751, - "data_source": "Mono3DRefer", - "scene_name": "000399", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[587.51, 189.86], [610.56, 189.85], [614.45, 192.84], [587.34, 192.86], [587.51, 164.87], [610.56, 164.88], [614.45, 163.47], [587.34, 163.46]]}]\n```" - } - ] - }, - { - "id": 752, - "data_source": "Mono3DRefer", - "scene_name": "000399", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[294.9, 211.61], [266.87, 211.61], [300.68, 207.78], [325.94, 207.78], [294.9, 180.25], [266.87, 180.25], [300.68, 179.52], [325.94, 179.52]]}]\n```" - } - ] - }, - { - "id": 753, - "data_source": "Mono3DRefer", - "scene_name": "000399", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[478.39, 195.47], [509.19, 195.47], [471.38, 203.94], [429.04, 203.93], [478.39, 158.66], [509.19, 158.66], [471.38, 153.34], [429.04, 153.34]]}]\n```" - } - ] - }, - { - "id": 754, - "data_source": "Mono3DRefer", - "scene_name": "000399", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[554.7, 194.28], [572.53, 194.28], [570.82, 195.26], [552.18, 195.26], [554.7, 177.18], [572.53, 177.18], [570.82, 177.37], [552.18, 177.37]]}]\n```" - } - ] - }, - { - "id": 755, - "data_source": "Mono3DRefer", - "scene_name": "000399", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[587.51, 189.86], [610.56, 189.85], [614.45, 192.84], [587.34, 192.86], [587.51, 164.87], [610.56, 164.88], [614.45, 163.47], [587.34, 163.46]]}, {\"category\": \"van\", \"corners_3d\": [[294.9, 211.61], [266.87, 211.61], [300.68, 207.78], [325.94, 207.78], [294.9, 180.25], [266.87, 180.25], [300.68, 179.52], [325.94, 179.52]]}]\n```" - } - ] - }, - { - "id": 756, - "data_source": "Mono3DRefer", - "scene_name": "000399", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[587.51, 189.86], [610.56, 189.85], [614.45, 192.84], [587.34, 192.86], [587.51, 164.87], [610.56, 164.88], [614.45, 163.47], [587.34, 163.46]]}, {\"category\": \"tram\", \"corners_3d\": [[478.39, 195.47], [509.19, 195.47], [471.38, 203.94], [429.04, 203.93], [478.39, 158.66], [509.19, 158.66], [471.38, 153.34], [429.04, 153.34]]}]\n```" - } - ] - }, - { - "id": 757, - "data_source": "Mono3DRefer", - "scene_name": "000399", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[554.7, 194.28], [572.53, 194.28], [570.82, 195.26], [552.18, 195.26], [554.7, 177.18], [572.53, 177.18], [570.82, 177.37], [552.18, 177.37]]}, {\"category\": \"bus\", \"corners_3d\": [[587.51, 189.86], [610.56, 189.85], [614.45, 192.84], [587.34, 192.86], [587.51, 164.87], [610.56, 164.88], [614.45, 163.47], [587.34, 163.46]]}]\n```" - } - ] - }, - { - "id": 758, - "data_source": "Mono3DRefer", - "scene_name": "000399", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[294.9, 211.61], [266.87, 211.61], [300.68, 207.78], [325.94, 207.78], [294.9, 180.25], [266.87, 180.25], [300.68, 179.52], [325.94, 179.52]]}, {\"category\": \"tram\", \"corners_3d\": [[478.39, 195.47], [509.19, 195.47], [471.38, 203.94], [429.04, 203.93], [478.39, 158.66], [509.19, 158.66], [471.38, 153.34], [429.04, 153.34]]}]\n```" - } - ] - }, - { - "id": 759, - "data_source": "Mono3DRefer", - "scene_name": "000399", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[554.7, 194.28], [572.53, 194.28], [570.82, 195.26], [552.18, 195.26], [554.7, 177.18], [572.53, 177.18], [570.82, 177.37], [552.18, 177.37]]}, {\"category\": \"van\", \"corners_3d\": [[294.9, 211.61], [266.87, 211.61], [300.68, 207.78], [325.94, 207.78], [294.9, 180.25], [266.87, 180.25], [300.68, 179.52], [325.94, 179.52]]}]\n```" - } - ] - }, - { - "id": 760, - "data_source": "Mono3DRefer", - "scene_name": "000399", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[554.7, 194.28], [572.53, 194.28], [570.82, 195.26], [552.18, 195.26], [554.7, 177.18], [572.53, 177.18], [570.82, 177.37], [552.18, 177.37]]}, {\"category\": \"tram\", \"corners_3d\": [[478.39, 195.47], [509.19, 195.47], [471.38, 203.94], [429.04, 203.93], [478.39, 158.66], [509.19, 158.66], [471.38, 153.34], [429.04, 153.34]]}]\n```" - } - ] - }, - { - "id": 761, - "data_source": "Mono3DRefer", - "scene_name": "000399", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, van, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[587.51, 189.86], [610.56, 189.85], [614.45, 192.84], [587.34, 192.86], [587.51, 164.87], [610.56, 164.88], [614.45, 163.47], [587.34, 163.46]]}, {\"category\": \"van\", \"corners_3d\": [[294.9, 211.61], [266.87, 211.61], [300.68, 207.78], [325.94, 207.78], [294.9, 180.25], [266.87, 180.25], [300.68, 179.52], [325.94, 179.52]]}, {\"category\": \"tram\", \"corners_3d\": [[478.39, 195.47], [509.19, 195.47], [471.38, 203.94], [429.04, 203.93], [478.39, 158.66], [509.19, 158.66], [471.38, 153.34], [429.04, 153.34]]}]\n```" - } - ] - }, - { - "id": 762, - "data_source": "Mono3DRefer", - "scene_name": "000399", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[554.7, 194.28], [572.53, 194.28], [570.82, 195.26], [552.18, 195.26], [554.7, 177.18], [572.53, 177.18], [570.82, 177.37], [552.18, 177.37]]}, {\"category\": \"bus\", \"corners_3d\": [[587.51, 189.86], [610.56, 189.85], [614.45, 192.84], [587.34, 192.86], [587.51, 164.87], [610.56, 164.88], [614.45, 163.47], [587.34, 163.46]]}, {\"category\": \"van\", \"corners_3d\": [[294.9, 211.61], [266.87, 211.61], [300.68, 207.78], [325.94, 207.78], [294.9, 180.25], [266.87, 180.25], [300.68, 179.52], [325.94, 179.52]]}]\n```" - } - ] - }, - { - "id": 763, - "data_source": "Mono3DRefer", - "scene_name": "000399", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[554.7, 194.28], [572.53, 194.28], [570.82, 195.26], [552.18, 195.26], [554.7, 177.18], [572.53, 177.18], [570.82, 177.37], [552.18, 177.37]]}, {\"category\": \"bus\", \"corners_3d\": [[587.51, 189.86], [610.56, 189.85], [614.45, 192.84], [587.34, 192.86], [587.51, 164.87], [610.56, 164.88], [614.45, 163.47], [587.34, 163.46]]}, {\"category\": \"tram\", \"corners_3d\": [[478.39, 195.47], [509.19, 195.47], [471.38, 203.94], [429.04, 203.93], [478.39, 158.66], [509.19, 158.66], [471.38, 153.34], [429.04, 153.34]]}]\n```" - } - ] - }, - { - "id": 764, - "data_source": "Mono3DRefer", - "scene_name": "000399", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[554.7, 194.28], [572.53, 194.28], [570.82, 195.26], [552.18, 195.26], [554.7, 177.18], [572.53, 177.18], [570.82, 177.37], [552.18, 177.37]]}, {\"category\": \"van\", \"corners_3d\": [[294.9, 211.61], [266.87, 211.61], [300.68, 207.78], [325.94, 207.78], [294.9, 180.25], [266.87, 180.25], [300.68, 179.52], [325.94, 179.52]]}, {\"category\": \"tram\", \"corners_3d\": [[478.39, 195.47], [509.19, 195.47], [471.38, 203.94], [429.04, 203.93], [478.39, 158.66], [509.19, 158.66], [471.38, 153.34], [429.04, 153.34]]}]\n```" - } - ] - }, - { - "id": 765, - "data_source": "Mono3DRefer", - "scene_name": "000399", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, van, tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[554.7, 194.28], [572.53, 194.28], [570.82, 195.26], [552.18, 195.26], [554.7, 177.18], [572.53, 177.18], [570.82, 177.37], [552.18, 177.37]]}, {\"category\": \"bus\", \"corners_3d\": [[587.51, 189.86], [610.56, 189.85], [614.45, 192.84], [587.34, 192.86], [587.51, 164.87], [610.56, 164.88], [614.45, 163.47], [587.34, 163.46]]}, {\"category\": \"van\", \"corners_3d\": [[294.9, 211.61], [266.87, 211.61], [300.68, 207.78], [325.94, 207.78], [294.9, 180.25], [266.87, 180.25], [300.68, 179.52], [325.94, 179.52]]}, {\"category\": \"tram\", \"corners_3d\": [[478.39, 195.47], [509.19, 195.47], [471.38, 203.94], [429.04, 203.93], [478.39, 158.66], [509.19, 158.66], [471.38, 153.34], [429.04, 153.34]]}]\n```" - } - ] - }, - { - "id": 766, - "data_source": "Mono3DRefer", - "scene_name": "000400", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[686.49, 225.0], [736.81, 224.89], [766.75, 235.4], [706.33, 235.55], [686.49, 178.15], [736.81, 178.14], [766.75, 179.21], [706.33, 179.22]]}, {\"category\": \"car\", \"corners_3d\": [[888.39, 333.01], [1039.99, 332.35], [1539.42, 511.65], [1220.26, 514.6], [888.39, 168.94], [1039.99, 168.95], [1539.42, 164.57], [1220.26, 164.49]]}, {\"category\": \"car\", \"corners_3d\": [[929.13, 207.57], [942.07, 206.73], [1016.88, 209.43], [1004.88, 210.4], [929.13, 177.52], [942.07, 177.4], [1016.88, 177.77], [1004.88, 177.9]]}, {\"category\": \"car\", \"corners_3d\": [[642.6, 205.55], [671.87, 205.52], [680.17, 208.77], [647.99, 208.81], [642.6, 178.36], [671.87, 178.36], [680.17, 178.91], [647.99, 178.91]]}]\n```" - } - ] - }, - { - "id": 767, - "data_source": "Mono3DRefer", - "scene_name": "000400", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[756.87, 215.08], [822.39, 214.93], [918.49, 231.7], [827.21, 232.0], [756.87, 123.09], [822.39, 123.27], [918.49, 103.51], [827.21, 103.16]]}]\n```" - } - ] - }, - { - "id": 768, - "data_source": "Mono3DRefer", - "scene_name": "000400", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[686.49, 225.0], [736.81, 224.89], [766.75, 235.4], [706.33, 235.55], [686.49, 178.15], [736.81, 178.14], [766.75, 179.21], [706.33, 179.22]]}, {\"category\": \"car\", \"corners_3d\": [[888.39, 333.01], [1039.99, 332.35], [1539.42, 511.65], [1220.26, 514.6], [888.39, 168.94], [1039.99, 168.95], [1539.42, 164.57], [1220.26, 164.49]]}, {\"category\": \"truck\", \"corners_3d\": [[756.87, 215.08], [822.39, 214.93], [918.49, 231.7], [827.21, 232.0], [756.87, 123.09], [822.39, 123.27], [918.49, 103.51], [827.21, 103.16]]}, {\"category\": \"car\", \"corners_3d\": [[929.13, 207.57], [942.07, 206.73], [1016.88, 209.43], [1004.88, 210.4], [929.13, 177.52], [942.07, 177.4], [1016.88, 177.77], [1004.88, 177.9]]}, {\"category\": \"car\", \"corners_3d\": [[642.6, 205.55], [671.87, 205.52], [680.17, 208.77], [647.99, 208.81], [642.6, 178.36], [671.87, 178.36], [680.17, 178.91], [647.99, 178.91]]}]\n```" - } - ] - }, - { - "id": 769, - "data_source": "Mono3DRefer", - "scene_name": "000404", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[508.62, 199.31], [527.1, 199.25], [530.19, 200.11], [510.58, 200.18], [508.62, 182.93], [527.1, 182.94], [530.19, 182.8], [510.58, 182.79]]}, {\"category\": \"car\", \"corners_3d\": [[542.28, 232.26], [600.14, 231.73], [619.8, 241.49], [549.9, 242.26], [542.28, 179.21], [600.14, 179.28], [619.8, 178.03], [549.9, 177.93]]}, {\"category\": \"car\", \"corners_3d\": [[-68.7, 333.61], [93.09, 329.47], [-265.41, 454.58], [-594.9, 469.36], [-68.7, 208.17], [93.09, 207.53], [-265.41, 226.88], [-594.9, 229.17]]}, {\"category\": \"car\", \"corners_3d\": [[202.36, 265.03], [287.46, 263.81], [207.1, 292.65], [88.37, 294.95], [202.36, 195.33], [287.46, 195.18], [207.1, 198.83], [88.37, 199.13]]}, {\"category\": \"car\", \"corners_3d\": [[451.6, 224.3], [498.38, 223.86], [502.08, 230.53], [447.16, 231.14], [451.6, 180.78], [498.38, 180.83], [502.08, 180.07], [447.16, 180.0]]}, {\"category\": \"car\", \"corners_3d\": [[298.24, 230.37], [353.19, 229.85], [320.33, 240.44], [251.69, 241.23], [298.24, 186.12], [353.19, 186.11], [320.33, 186.32], [251.69, 186.33]]}, {\"category\": \"car\", \"corners_3d\": [[206.93, 229.23], [260.6, 228.83], [209.05, 237.71], [143.76, 238.29], [206.93, 186.68], [260.6, 186.67], [209.05, 186.96], [143.76, 186.98]]}, {\"category\": \"car\", \"corners_3d\": [[369.38, 213.88], [402.09, 213.72], [386.89, 217.6], [349.6, 217.81], [369.38, 184.36], [402.09, 184.37], [386.89, 184.25], [349.6, 184.24]]}, {\"category\": \"car\", \"corners_3d\": [[392.06, 207.48], [419.71, 207.37], [410.82, 209.64], [380.26, 209.78], [392.06, 183.7], [419.71, 183.71], [410.82, 183.55], [380.26, 183.54]]}]\n```" - } - ] - }, - { - "id": 770, - "data_source": "Mono3DRefer", - "scene_name": "000404", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[0.43, 256.63], [96.62, 255.28], [-93.66, 289.02], [-243.0, 292.01], [0.43, 168.24], [96.62, 168.56], [-93.66, 160.54], [-243.0, 159.82]]}, {\"category\": \"van\", \"corners_3d\": [[262.96, 215.12], [305.74, 214.87], [275.61, 219.5], [225.8, 219.83], [262.96, 172.27], [305.74, 172.38], [275.61, 170.38], [225.8, 170.23]]}, {\"category\": \"van\", \"corners_3d\": [[418.45, 200.06], [441.73, 199.97], [438.45, 201.71], [412.4, 201.82], [418.45, 169.4], [441.73, 169.49], [438.45, 167.64], [412.4, 167.52]]}]\n```" - } - ] - }, - { - "id": 771, - "data_source": "Mono3DRefer", - "scene_name": "000404", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[508.62, 199.31], [527.1, 199.25], [530.19, 200.11], [510.58, 200.18], [508.62, 182.93], [527.1, 182.94], [530.19, 182.8], [510.58, 182.79]]}, {\"category\": \"car\", \"corners_3d\": [[542.28, 232.26], [600.14, 231.73], [619.8, 241.49], [549.9, 242.26], [542.28, 179.21], [600.14, 179.28], [619.8, 178.03], [549.9, 177.93]]}, {\"category\": \"car\", \"corners_3d\": [[-68.7, 333.61], [93.09, 329.47], [-265.41, 454.58], [-594.9, 469.36], [-68.7, 208.17], [93.09, 207.53], [-265.41, 226.88], [-594.9, 229.17]]}, {\"category\": \"van\", \"corners_3d\": [[0.43, 256.63], [96.62, 255.28], [-93.66, 289.02], [-243.0, 292.01], [0.43, 168.24], [96.62, 168.56], [-93.66, 160.54], [-243.0, 159.82]]}, {\"category\": \"car\", \"corners_3d\": [[202.36, 265.03], [287.46, 263.81], [207.1, 292.65], [88.37, 294.95], [202.36, 195.33], [287.46, 195.18], [207.1, 198.83], [88.37, 199.13]]}, {\"category\": \"car\", \"corners_3d\": [[451.6, 224.3], [498.38, 223.86], [502.08, 230.53], [447.16, 231.14], [451.6, 180.78], [498.38, 180.83], [502.08, 180.07], [447.16, 180.0]]}, {\"category\": \"car\", \"corners_3d\": [[298.24, 230.37], [353.19, 229.85], [320.33, 240.44], [251.69, 241.23], [298.24, 186.12], [353.19, 186.11], [320.33, 186.32], [251.69, 186.33]]}, {\"category\": \"car\", \"corners_3d\": [[206.93, 229.23], [260.6, 228.83], [209.05, 237.71], [143.76, 238.29], [206.93, 186.68], [260.6, 186.67], [209.05, 186.96], [143.76, 186.98]]}, {\"category\": \"car\", \"corners_3d\": [[369.38, 213.88], [402.09, 213.72], [386.89, 217.6], [349.6, 217.81], [369.38, 184.36], [402.09, 184.37], [386.89, 184.25], [349.6, 184.24]]}, {\"category\": \"van\", \"corners_3d\": [[262.96, 215.12], [305.74, 214.87], [275.61, 219.5], [225.8, 219.83], [262.96, 172.27], [305.74, 172.38], [275.61, 170.38], [225.8, 170.23]]}, {\"category\": \"car\", \"corners_3d\": [[392.06, 207.48], [419.71, 207.37], [410.82, 209.64], [380.26, 209.78], [392.06, 183.7], [419.71, 183.71], [410.82, 183.55], [380.26, 183.54]]}, {\"category\": \"van\", \"corners_3d\": [[418.45, 200.06], [441.73, 199.97], [438.45, 201.71], [412.4, 201.82], [418.45, 169.4], [441.73, 169.49], [438.45, 167.64], [412.4, 167.52]]}]\n```" - } - ] - }, - { - "id": 772, - "data_source": "Mono3DRefer", - "scene_name": "000408", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[602.09, 224.71], [585.7, 224.7], [587.13, 222.27], [602.75, 222.28], [602.09, 170.76], [585.7, 170.76], [587.13, 170.86], [602.75, 170.85]]}]\n```" - } - ] - }, - { - "id": 773, - "data_source": "Mono3DRefer", - "scene_name": "000408", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[528.04, 234.28], [468.58, 234.08], [497.63, 223.57], [546.94, 223.71], [528.04, 171.74], [468.58, 171.74], [497.63, 171.93], [546.94, 171.93]]}, {\"category\": \"car\", \"corners_3d\": [[720.69, 284.08], [828.9, 284.26], [893.02, 318.03], [751.95, 317.72], [720.69, 176.22], [828.9, 176.23], [893.02, 177.25], [751.95, 177.24]]}, {\"category\": \"car\", \"corners_3d\": [[692.77, 247.83], [765.96, 248.37], [792.9, 267.69], [700.9, 266.85], [692.77, 185.56], [765.96, 185.65], [792.9, 188.93], [700.9, 188.79]]}, {\"category\": \"car\", \"corners_3d\": [[529.17, 210.52], [567.37, 210.62], [556.76, 215.7], [513.47, 215.57], [529.17, 178.36], [567.37, 178.37], [556.76, 179.12], [513.47, 179.1]]}, {\"category\": \"car\", \"corners_3d\": [[663.95, 200.16], [637.94, 200.16], [635.7, 198.05], [659.71, 198.05], [663.95, 176.08], [637.94, 176.08], [635.7, 175.83], [659.71, 175.83]]}, {\"category\": \"car\", \"corners_3d\": [[147.52, 244.93], [194.0, 237.7], [390.72, 237.73], [366.16, 244.97], [147.52, 166.42], [194.0, 167.06], [390.72, 167.06], [366.16, 166.41]]}]\n```" - } - ] - }, - { - "id": 774, - "data_source": "Mono3DRefer", - "scene_name": "000408", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[528.04, 234.28], [468.58, 234.08], [497.63, 223.57], [546.94, 223.71], [528.04, 171.74], [468.58, 171.74], [497.63, 171.93], [546.94, 171.93]]}, {\"category\": \"car\", \"corners_3d\": [[720.69, 284.08], [828.9, 284.26], [893.02, 318.03], [751.95, 317.72], [720.69, 176.22], [828.9, 176.23], [893.02, 177.25], [751.95, 177.24]]}, {\"category\": \"cyclist\", \"corners_3d\": [[602.09, 224.71], [585.7, 224.7], [587.13, 222.27], [602.75, 222.28], [602.09, 170.76], [585.7, 170.76], [587.13, 170.86], [602.75, 170.85]]}, {\"category\": \"car\", \"corners_3d\": [[692.77, 247.83], [765.96, 248.37], [792.9, 267.69], [700.9, 266.85], [692.77, 185.56], [765.96, 185.65], [792.9, 188.93], [700.9, 188.79]]}, {\"category\": \"car\", \"corners_3d\": [[529.17, 210.52], [567.37, 210.62], [556.76, 215.7], [513.47, 215.57], [529.17, 178.36], [567.37, 178.37], [556.76, 179.12], [513.47, 179.1]]}, {\"category\": \"car\", \"corners_3d\": [[663.95, 200.16], [637.94, 200.16], [635.7, 198.05], [659.71, 198.05], [663.95, 176.08], [637.94, 176.08], [635.7, 175.83], [659.71, 175.83]]}, {\"category\": \"car\", \"corners_3d\": [[147.52, 244.93], [194.0, 237.7], [390.72, 237.73], [366.16, 244.97], [147.52, 166.42], [194.0, 167.06], [390.72, 167.06], [366.16, 166.41]]}]\n```" - } - ] - }, - { - "id": 775, - "data_source": "Mono3DRefer", - "scene_name": "000423", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[685.87, 224.55], [772.75, 224.56], [867.71, 254.75], [730.1, 254.74], [685.87, 109.3], [772.75, 109.3], [867.71, 72.18], [730.1, 72.2]]}, {\"category\": \"truck\", \"corners_3d\": [[627.28, 191.4], [661.59, 191.41], [668.1, 194.68], [627.76, 194.65], [627.28, 145.3], [661.59, 145.27], [668.1, 140.42], [627.76, 140.46]]}]\n```" - } - ] - }, - { - "id": 776, - "data_source": "Mono3DRefer", - "scene_name": "000423", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[589.68, 211.68], [633.99, 211.64], [639.35, 217.06], [588.84, 217.12], [589.68, 175.99], [633.99, 175.99], [639.35, 176.43], [588.84, 176.43]]}, {\"category\": \"car\", \"corners_3d\": [[267.18, 317.51], [98.58, 345.24], [-8.14, 305.08], [139.9, 288.13], [267.18, 179.0], [98.58, 180.18], [-8.14, 178.47], [139.9, 177.75]]}, {\"category\": \"car\", \"corners_3d\": [[411.2, 247.45], [345.64, 253.52], [258.69, 241.69], [320.9, 237.22], [411.2, 184.28], [345.64, 185.21], [258.69, 183.4], [320.9, 182.71]]}, {\"category\": \"car\", \"corners_3d\": [[489.05, 240.28], [552.28, 240.4], [534.31, 257.2], [455.43, 257.01], [489.05, 187.14], [552.28, 187.17], [534.31, 190.73], [455.43, 190.69]]}, {\"category\": \"car\", \"corners_3d\": [[523.16, 225.75], [573.98, 225.94], [559.75, 236.39], [499.03, 236.12], [523.16, 180.72], [573.98, 180.74], [559.75, 182.3], [499.03, 182.26]]}, {\"category\": \"car\", \"corners_3d\": [[503.43, 215.65], [468.33, 217.94], [393.21, 213.75], [428.68, 211.86], [503.43, 172.29], [468.33, 172.26], [393.21, 172.32], [428.68, 172.34]]}, {\"category\": \"car\", \"corners_3d\": [[780.02, 305.37], [906.1, 305.39], [1108.19, 395.88], [895.99, 395.83], [780.02, 193.15], [906.1, 193.16], [1108.19, 207.02], [895.99, 207.01]]}, {\"category\": \"car\", \"corners_3d\": [[653.3, 205.54], [689.57, 205.56], [698.8, 209.74], [657.89, 209.72], [653.3, 173.51], [689.57, 173.51], [698.8, 173.59], [657.89, 173.59]]}, {\"category\": \"car\", \"corners_3d\": [[323.05, 206.56], [287.1, 206.58], [320.13, 203.04], [352.29, 203.02], [323.05, 176.98], [287.1, 176.99], [320.13, 176.55], [352.29, 176.55]]}, {\"category\": \"car\", \"corners_3d\": [[585.22, 201.75], [616.16, 201.76], [616.78, 204.71], [582.67, 204.71], [585.22, 173.59], [616.16, 173.59], [616.78, 173.66], [582.67, 173.66]]}, {\"category\": \"car\", \"corners_3d\": [[118.93, 230.19], [66.87, 230.11], [150.17, 221.54], [194.53, 221.6], [118.93, 180.14], [66.87, 180.13], [150.17, 179.04], [194.53, 179.05]]}]\n```" - } - ] - }, - { - "id": 777, - "data_source": "Mono3DRefer", - "scene_name": "000423", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[867.73, 268.45], [897.08, 265.04], [932.98, 269.33], [903.6, 273.08], [867.73, 158.64], [897.08, 159.15], [932.98, 158.51], [903.6, 157.95]]}]\n```" - } - ] - }, - { - "id": 778, - "data_source": "Mono3DRefer", - "scene_name": "000423", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[589.68, 211.68], [633.99, 211.64], [639.35, 217.06], [588.84, 217.12], [589.68, 175.99], [633.99, 175.99], [639.35, 176.43], [588.84, 176.43]]}, {\"category\": \"car\", \"corners_3d\": [[267.18, 317.51], [98.58, 345.24], [-8.14, 305.08], [139.9, 288.13], [267.18, 179.0], [98.58, 180.18], [-8.14, 178.47], [139.9, 177.75]]}, {\"category\": \"car\", \"corners_3d\": [[411.2, 247.45], [345.64, 253.52], [258.69, 241.69], [320.9, 237.22], [411.2, 184.28], [345.64, 185.21], [258.69, 183.4], [320.9, 182.71]]}, {\"category\": \"car\", \"corners_3d\": [[489.05, 240.28], [552.28, 240.4], [534.31, 257.2], [455.43, 257.01], [489.05, 187.14], [552.28, 187.17], [534.31, 190.73], [455.43, 190.69]]}, {\"category\": \"car\", \"corners_3d\": [[523.16, 225.75], [573.98, 225.94], [559.75, 236.39], [499.03, 236.12], [523.16, 180.72], [573.98, 180.74], [559.75, 182.3], [499.03, 182.26]]}, {\"category\": \"car\", \"corners_3d\": [[503.43, 215.65], [468.33, 217.94], [393.21, 213.75], [428.68, 211.86], [503.43, 172.29], [468.33, 172.26], [393.21, 172.32], [428.68, 172.34]]}, {\"category\": \"car\", \"corners_3d\": [[780.02, 305.37], [906.1, 305.39], [1108.19, 395.88], [895.99, 395.83], [780.02, 193.15], [906.1, 193.16], [1108.19, 207.02], [895.99, 207.01]]}, {\"category\": \"truck\", \"corners_3d\": [[685.87, 224.55], [772.75, 224.56], [867.71, 254.75], [730.1, 254.74], [685.87, 109.3], [772.75, 109.3], [867.71, 72.18], [730.1, 72.2]]}, {\"category\": \"car\", \"corners_3d\": [[653.3, 205.54], [689.57, 205.56], [698.8, 209.74], [657.89, 209.72], [653.3, 173.51], [689.57, 173.51], [698.8, 173.59], [657.89, 173.59]]}, {\"category\": \"truck\", \"corners_3d\": [[627.28, 191.4], [661.59, 191.41], [668.1, 194.68], [627.76, 194.65], [627.28, 145.3], [661.59, 145.27], [668.1, 140.42], [627.76, 140.46]]}, {\"category\": \"car\", \"corners_3d\": [[323.05, 206.56], [287.1, 206.58], [320.13, 203.04], [352.29, 203.02], [323.05, 176.98], [287.1, 176.99], [320.13, 176.55], [352.29, 176.55]]}, {\"category\": \"car\", \"corners_3d\": [[585.22, 201.75], [616.16, 201.76], [616.78, 204.71], [582.67, 204.71], [585.22, 173.59], [616.16, 173.59], [616.78, 173.66], [582.67, 173.66]]}, {\"category\": \"car\", \"corners_3d\": [[118.93, 230.19], [66.87, 230.11], [150.17, 221.54], [194.53, 221.6], [118.93, 180.14], [66.87, 180.13], [150.17, 179.04], [194.53, 179.05]]}]\n```" - } - ] - }, - { - "id": 779, - "data_source": "Mono3DRefer", - "scene_name": "000423", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[685.87, 224.55], [772.75, 224.56], [867.71, 254.75], [730.1, 254.74], [685.87, 109.3], [772.75, 109.3], [867.71, 72.18], [730.1, 72.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[867.73, 268.45], [897.08, 265.04], [932.98, 269.33], [903.6, 273.08], [867.73, 158.64], [897.08, 159.15], [932.98, 158.51], [903.6, 157.95]]}, {\"category\": \"truck\", \"corners_3d\": [[627.28, 191.4], [661.59, 191.41], [668.1, 194.68], [627.76, 194.65], [627.28, 145.3], [661.59, 145.27], [668.1, 140.42], [627.76, 140.46]]}]\n```" - } - ] - }, - { - "id": 780, - "data_source": "Mono3DRefer", - "scene_name": "000423", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[589.68, 211.68], [633.99, 211.64], [639.35, 217.06], [588.84, 217.12], [589.68, 175.99], [633.99, 175.99], [639.35, 176.43], [588.84, 176.43]]}, {\"category\": \"car\", \"corners_3d\": [[267.18, 317.51], [98.58, 345.24], [-8.14, 305.08], [139.9, 288.13], [267.18, 179.0], [98.58, 180.18], [-8.14, 178.47], [139.9, 177.75]]}, {\"category\": \"car\", \"corners_3d\": [[411.2, 247.45], [345.64, 253.52], [258.69, 241.69], [320.9, 237.22], [411.2, 184.28], [345.64, 185.21], [258.69, 183.4], [320.9, 182.71]]}, {\"category\": \"car\", \"corners_3d\": [[489.05, 240.28], [552.28, 240.4], [534.31, 257.2], [455.43, 257.01], [489.05, 187.14], [552.28, 187.17], [534.31, 190.73], [455.43, 190.69]]}, {\"category\": \"car\", \"corners_3d\": [[523.16, 225.75], [573.98, 225.94], [559.75, 236.39], [499.03, 236.12], [523.16, 180.72], [573.98, 180.74], [559.75, 182.3], [499.03, 182.26]]}, {\"category\": \"car\", \"corners_3d\": [[503.43, 215.65], [468.33, 217.94], [393.21, 213.75], [428.68, 211.86], [503.43, 172.29], [468.33, 172.26], [393.21, 172.32], [428.68, 172.34]]}, {\"category\": \"car\", \"corners_3d\": [[780.02, 305.37], [906.1, 305.39], [1108.19, 395.88], [895.99, 395.83], [780.02, 193.15], [906.1, 193.16], [1108.19, 207.02], [895.99, 207.01]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[867.73, 268.45], [897.08, 265.04], [932.98, 269.33], [903.6, 273.08], [867.73, 158.64], [897.08, 159.15], [932.98, 158.51], [903.6, 157.95]]}, {\"category\": \"car\", \"corners_3d\": [[653.3, 205.54], [689.57, 205.56], [698.8, 209.74], [657.89, 209.72], [653.3, 173.51], [689.57, 173.51], [698.8, 173.59], [657.89, 173.59]]}, {\"category\": \"car\", \"corners_3d\": [[323.05, 206.56], [287.1, 206.58], [320.13, 203.04], [352.29, 203.02], [323.05, 176.98], [287.1, 176.99], [320.13, 176.55], [352.29, 176.55]]}, {\"category\": \"car\", \"corners_3d\": [[585.22, 201.75], [616.16, 201.76], [616.78, 204.71], [582.67, 204.71], [585.22, 173.59], [616.16, 173.59], [616.78, 173.66], [582.67, 173.66]]}, {\"category\": \"car\", \"corners_3d\": [[118.93, 230.19], [66.87, 230.11], [150.17, 221.54], [194.53, 221.6], [118.93, 180.14], [66.87, 180.13], [150.17, 179.04], [194.53, 179.05]]}]\n```" - } - ] - }, - { - "id": 781, - "data_source": "Mono3DRefer", - "scene_name": "000423", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[589.68, 211.68], [633.99, 211.64], [639.35, 217.06], [588.84, 217.12], [589.68, 175.99], [633.99, 175.99], [639.35, 176.43], [588.84, 176.43]]}, {\"category\": \"car\", \"corners_3d\": [[267.18, 317.51], [98.58, 345.24], [-8.14, 305.08], [139.9, 288.13], [267.18, 179.0], [98.58, 180.18], [-8.14, 178.47], [139.9, 177.75]]}, {\"category\": \"car\", \"corners_3d\": [[411.2, 247.45], [345.64, 253.52], [258.69, 241.69], [320.9, 237.22], [411.2, 184.28], [345.64, 185.21], [258.69, 183.4], [320.9, 182.71]]}, {\"category\": \"car\", \"corners_3d\": [[489.05, 240.28], [552.28, 240.4], [534.31, 257.2], [455.43, 257.01], [489.05, 187.14], [552.28, 187.17], [534.31, 190.73], [455.43, 190.69]]}, {\"category\": \"car\", \"corners_3d\": [[523.16, 225.75], [573.98, 225.94], [559.75, 236.39], [499.03, 236.12], [523.16, 180.72], [573.98, 180.74], [559.75, 182.3], [499.03, 182.26]]}, {\"category\": \"car\", \"corners_3d\": [[503.43, 215.65], [468.33, 217.94], [393.21, 213.75], [428.68, 211.86], [503.43, 172.29], [468.33, 172.26], [393.21, 172.32], [428.68, 172.34]]}, {\"category\": \"car\", \"corners_3d\": [[780.02, 305.37], [906.1, 305.39], [1108.19, 395.88], [895.99, 395.83], [780.02, 193.15], [906.1, 193.16], [1108.19, 207.02], [895.99, 207.01]]}, {\"category\": \"truck\", \"corners_3d\": [[685.87, 224.55], [772.75, 224.56], [867.71, 254.75], [730.1, 254.74], [685.87, 109.3], [772.75, 109.3], [867.71, 72.18], [730.1, 72.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[867.73, 268.45], [897.08, 265.04], [932.98, 269.33], [903.6, 273.08], [867.73, 158.64], [897.08, 159.15], [932.98, 158.51], [903.6, 157.95]]}, {\"category\": \"car\", \"corners_3d\": [[653.3, 205.54], [689.57, 205.56], [698.8, 209.74], [657.89, 209.72], [653.3, 173.51], [689.57, 173.51], [698.8, 173.59], [657.89, 173.59]]}, {\"category\": \"truck\", \"corners_3d\": [[627.28, 191.4], [661.59, 191.41], [668.1, 194.68], [627.76, 194.65], [627.28, 145.3], [661.59, 145.27], [668.1, 140.42], [627.76, 140.46]]}, {\"category\": \"car\", \"corners_3d\": [[323.05, 206.56], [287.1, 206.58], [320.13, 203.04], [352.29, 203.02], [323.05, 176.98], [287.1, 176.99], [320.13, 176.55], [352.29, 176.55]]}, {\"category\": \"car\", \"corners_3d\": [[585.22, 201.75], [616.16, 201.76], [616.78, 204.71], [582.67, 204.71], [585.22, 173.59], [616.16, 173.59], [616.78, 173.66], [582.67, 173.66]]}, {\"category\": \"car\", \"corners_3d\": [[118.93, 230.19], [66.87, 230.11], [150.17, 221.54], [194.53, 221.6], [118.93, 180.14], [66.87, 180.13], [150.17, 179.04], [194.53, 179.05]]}]\n```" - } - ] - }, - { - "id": 782, - "data_source": "Mono3DRefer", - "scene_name": "000424", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[334.43, 222.4], [388.12, 221.22], [398.31, 233.17], [331.27, 235.02], [334.43, 156.12], [388.12, 156.52], [398.31, 152.49], [331.27, 151.87]]}]\n```" - } - ] - }, - { - "id": 783, - "data_source": "Mono3DRefer", - "scene_name": "000424", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-195.63, 448.96], [-523.01, 471.44], [-52.47, 326.69], [103.65, 320.5], [-195.63, 244.29], [-523.01, 250.11], [-52.47, 212.66], [103.65, 211.06]]}, {\"category\": \"car\", \"corners_3d\": [[1463.87, 302.49], [1733.86, 334.5], [1423.33, 382.17], [1168.25, 331.46], [1463.87, 117.51], [1733.86, 103.84], [1423.33, 83.49], [1168.25, 105.14]]}, {\"category\": \"car\", \"corners_3d\": [[1025.8, 252.89], [1112.7, 263.12], [878.52, 277.46], [809.81, 263.96], [1025.8, 170.13], [1112.7, 169.78], [878.52, 169.29], [809.81, 169.75]]}, {\"category\": \"car\", \"corners_3d\": [[838.04, 237.42], [886.18, 243.15], [755.84, 249.77], [715.11, 242.96], [838.04, 179.83], [886.18, 180.45], [755.84, 181.17], [715.11, 180.43]]}, {\"category\": \"car\", \"corners_3d\": [[443.54, 229.91], [492.98, 228.45], [525.42, 238.5], [467.76, 240.55], [443.54, 191.56], [492.98, 191.08], [525.42, 194.38], [467.76, 195.05]]}, {\"category\": \"car\", \"corners_3d\": [[348.87, 229.72], [394.1, 228.64], [399.59, 238.51], [346.3, 240.02], [348.87, 191.28], [394.1, 190.93], [399.59, 194.13], [346.3, 194.62]]}]\n```" - } - ] - }, - { - "id": 784, - "data_source": "Mono3DRefer", - "scene_name": "000424", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[21.13, 424.39], [-73.81, 432.11], [52.94, 362.97], [120.33, 358.78], [21.13, 179.5], [-73.81, 179.7], [52.94, 177.88], [120.33, 177.76]]}]\n```" - } - ] - }, - { - "id": 785, - "data_source": "Mono3DRefer", - "scene_name": "000424", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-195.63, 448.96], [-523.01, 471.44], [-52.47, 326.69], [103.65, 320.5], [-195.63, 244.29], [-523.01, 250.11], [-52.47, 212.66], [103.65, 211.06]]}, {\"category\": \"car\", \"corners_3d\": [[1463.87, 302.49], [1733.86, 334.5], [1423.33, 382.17], [1168.25, 331.46], [1463.87, 117.51], [1733.86, 103.84], [1423.33, 83.49], [1168.25, 105.14]]}, {\"category\": \"car\", \"corners_3d\": [[1025.8, 252.89], [1112.7, 263.12], [878.52, 277.46], [809.81, 263.96], [1025.8, 170.13], [1112.7, 169.78], [878.52, 169.29], [809.81, 169.75]]}, {\"category\": \"car\", \"corners_3d\": [[838.04, 237.42], [886.18, 243.15], [755.84, 249.77], [715.11, 242.96], [838.04, 179.83], [886.18, 180.45], [755.84, 181.17], [715.11, 180.43]]}, {\"category\": \"car\", \"corners_3d\": [[443.54, 229.91], [492.98, 228.45], [525.42, 238.5], [467.76, 240.55], [443.54, 191.56], [492.98, 191.08], [525.42, 194.38], [467.76, 195.05]]}, {\"category\": \"car\", \"corners_3d\": [[348.87, 229.72], [394.1, 228.64], [399.59, 238.51], [346.3, 240.02], [348.87, 191.28], [394.1, 190.93], [399.59, 194.13], [346.3, 194.62]]}, {\"category\": \"truck\", \"corners_3d\": [[334.43, 222.4], [388.12, 221.22], [398.31, 233.17], [331.27, 235.02], [334.43, 156.12], [388.12, 156.52], [398.31, 152.49], [331.27, 151.87]]}]\n```" - } - ] - }, - { - "id": 786, - "data_source": "Mono3DRefer", - "scene_name": "000424", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[21.13, 424.39], [-73.81, 432.11], [52.94, 362.97], [120.33, 358.78], [21.13, 179.5], [-73.81, 179.7], [52.94, 177.88], [120.33, 177.76]]}, {\"category\": \"truck\", \"corners_3d\": [[334.43, 222.4], [388.12, 221.22], [398.31, 233.17], [331.27, 235.02], [334.43, 156.12], [388.12, 156.52], [398.31, 152.49], [331.27, 151.87]]}]\n```" - } - ] - }, - { - "id": 787, - "data_source": "Mono3DRefer", - "scene_name": "000424", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-195.63, 448.96], [-523.01, 471.44], [-52.47, 326.69], [103.65, 320.5], [-195.63, 244.29], [-523.01, 250.11], [-52.47, 212.66], [103.65, 211.06]]}, {\"category\": \"car\", \"corners_3d\": [[1463.87, 302.49], [1733.86, 334.5], [1423.33, 382.17], [1168.25, 331.46], [1463.87, 117.51], [1733.86, 103.84], [1423.33, 83.49], [1168.25, 105.14]]}, {\"category\": \"car\", \"corners_3d\": [[1025.8, 252.89], [1112.7, 263.12], [878.52, 277.46], [809.81, 263.96], [1025.8, 170.13], [1112.7, 169.78], [878.52, 169.29], [809.81, 169.75]]}, {\"category\": \"cyclist\", \"corners_3d\": [[21.13, 424.39], [-73.81, 432.11], [52.94, 362.97], [120.33, 358.78], [21.13, 179.5], [-73.81, 179.7], [52.94, 177.88], [120.33, 177.76]]}, {\"category\": \"car\", \"corners_3d\": [[838.04, 237.42], [886.18, 243.15], [755.84, 249.77], [715.11, 242.96], [838.04, 179.83], [886.18, 180.45], [755.84, 181.17], [715.11, 180.43]]}, {\"category\": \"car\", \"corners_3d\": [[443.54, 229.91], [492.98, 228.45], [525.42, 238.5], [467.76, 240.55], [443.54, 191.56], [492.98, 191.08], [525.42, 194.38], [467.76, 195.05]]}, {\"category\": \"car\", \"corners_3d\": [[348.87, 229.72], [394.1, 228.64], [399.59, 238.51], [346.3, 240.02], [348.87, 191.28], [394.1, 190.93], [399.59, 194.13], [346.3, 194.62]]}]\n```" - } - ] - }, - { - "id": 788, - "data_source": "Mono3DRefer", - "scene_name": "000424", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-195.63, 448.96], [-523.01, 471.44], [-52.47, 326.69], [103.65, 320.5], [-195.63, 244.29], [-523.01, 250.11], [-52.47, 212.66], [103.65, 211.06]]}, {\"category\": \"car\", \"corners_3d\": [[1463.87, 302.49], [1733.86, 334.5], [1423.33, 382.17], [1168.25, 331.46], [1463.87, 117.51], [1733.86, 103.84], [1423.33, 83.49], [1168.25, 105.14]]}, {\"category\": \"car\", \"corners_3d\": [[1025.8, 252.89], [1112.7, 263.12], [878.52, 277.46], [809.81, 263.96], [1025.8, 170.13], [1112.7, 169.78], [878.52, 169.29], [809.81, 169.75]]}, {\"category\": \"cyclist\", \"corners_3d\": [[21.13, 424.39], [-73.81, 432.11], [52.94, 362.97], [120.33, 358.78], [21.13, 179.5], [-73.81, 179.7], [52.94, 177.88], [120.33, 177.76]]}, {\"category\": \"car\", \"corners_3d\": [[838.04, 237.42], [886.18, 243.15], [755.84, 249.77], [715.11, 242.96], [838.04, 179.83], [886.18, 180.45], [755.84, 181.17], [715.11, 180.43]]}, {\"category\": \"car\", \"corners_3d\": [[443.54, 229.91], [492.98, 228.45], [525.42, 238.5], [467.76, 240.55], [443.54, 191.56], [492.98, 191.08], [525.42, 194.38], [467.76, 195.05]]}, {\"category\": \"car\", \"corners_3d\": [[348.87, 229.72], [394.1, 228.64], [399.59, 238.51], [346.3, 240.02], [348.87, 191.28], [394.1, 190.93], [399.59, 194.13], [346.3, 194.62]]}, {\"category\": \"truck\", \"corners_3d\": [[334.43, 222.4], [388.12, 221.22], [398.31, 233.17], [331.27, 235.02], [334.43, 156.12], [388.12, 156.52], [398.31, 152.49], [331.27, 151.87]]}]\n```" - } - ] - }, - { - "id": 789, - "data_source": "Mono3DRefer", - "scene_name": "000426", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[1008.75, 338.44], [788.96, 322.76], [826.65, 261.03], [952.82, 266.23], [1008.75, 82.31], [788.96, 90.88], [826.65, 124.64], [952.82, 121.79]]}]\n```" - } - ] - }, - { - "id": 790, - "data_source": "Mono3DRefer", - "scene_name": "000426", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[562.18, 197.5], [587.99, 197.33], [598.25, 199.48], [570.23, 199.68], [562.18, 172.36], [587.99, 172.36], [598.25, 172.31], [570.23, 172.31]]}, {\"category\": \"car\", \"corners_3d\": [[252.57, 231.28], [192.86, 234.4], [163.14, 227.63], [217.93, 225.14], [252.57, 176.59], [192.86, 176.79], [163.14, 176.35], [217.93, 176.19]]}, {\"category\": \"car\", \"corners_3d\": [[507.05, 190.77], [485.22, 190.83], [488.08, 189.66], [508.49, 189.61], [507.05, 168.81], [485.22, 168.79], [488.08, 169.06], [508.49, 169.07]]}]\n```" - } - ] - }, - { - "id": 791, - "data_source": "Mono3DRefer", - "scene_name": "000426", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[1008.75, 338.44], [788.96, 322.76], [826.65, 261.03], [952.82, 266.23], [1008.75, 82.31], [788.96, 90.88], [826.65, 124.64], [952.82, 121.79]]}, {\"category\": \"car\", \"corners_3d\": [[562.18, 197.5], [587.99, 197.33], [598.25, 199.48], [570.23, 199.68], [562.18, 172.36], [587.99, 172.36], [598.25, 172.31], [570.23, 172.31]]}, {\"category\": \"car\", \"corners_3d\": [[252.57, 231.28], [192.86, 234.4], [163.14, 227.63], [217.93, 225.14], [252.57, 176.59], [192.86, 176.79], [163.14, 176.35], [217.93, 176.19]]}, {\"category\": \"car\", \"corners_3d\": [[507.05, 190.77], [485.22, 190.83], [488.08, 189.66], [508.49, 189.61], [507.05, 168.81], [485.22, 168.79], [488.08, 169.06], [508.49, 169.07]]}]\n```" - } - ] - }, - { - "id": 792, - "data_source": "Mono3DRefer", - "scene_name": "000431", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[388.11, 304.7], [252.99, 304.22], [365.78, 264.22], [459.94, 264.45], [388.11, 200.27], [252.99, 200.17], [365.78, 191.85], [459.94, 191.9]]}, {\"category\": \"car\", \"corners_3d\": [[478.26, 217.16], [468.42, 219.76], [358.79, 219.4], [374.6, 216.84], [478.26, 180.58], [468.42, 181.04], [358.79, 180.97], [374.6, 180.53]]}]\n```" - } - ] - }, - { - "id": 793, - "data_source": "Mono3DRefer", - "scene_name": "000434", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[558.01, 210.73], [523.67, 210.71], [532.74, 207.0], [563.72, 207.01], [558.01, 182.56], [523.67, 182.56], [532.74, 181.6], [563.72, 181.61]]}]\n```" - } - ] - }, - { - "id": 794, - "data_source": "Mono3DRefer", - "scene_name": "000437", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[578.98, 192.29], [601.2, 192.28], [602.13, 193.7], [578.28, 193.72], [578.98, 174.79], [601.2, 174.79], [602.13, 174.93], [578.28, 174.94]]}, {\"category\": \"car\", \"corners_3d\": [[552.91, 188.27], [571.83, 188.26], [570.81, 189.19], [550.75, 189.2], [552.91, 172.39], [571.83, 172.39], [570.81, 172.36], [550.75, 172.36]]}]\n```" - } - ] - }, - { - "id": 795, - "data_source": "Mono3DRefer", - "scene_name": "000439", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[653.46, 200.05], [678.89, 200.08], [682.59, 201.65], [655.11, 201.62], [653.46, 176.95], [678.89, 176.94], [682.59, 176.66], [655.11, 176.66]]}, {\"category\": \"car\", \"corners_3d\": [[437.6, 207.42], [410.21, 207.27], [436.2, 204.89], [461.29, 205.02], [437.6, 180.29], [410.21, 180.29], [436.2, 180.31], [461.29, 180.31]]}, {\"category\": \"car\", \"corners_3d\": [[722.58, 239.61], [790.57, 240.29], [817.53, 256.3], [731.2, 255.22], [722.58, 174.51], [790.57, 174.45], [817.53, 172.82], [731.2, 172.93]]}, {\"category\": \"car\", \"corners_3d\": [[695.49, 225.58], [748.91, 225.62], [776.07, 234.55], [712.05, 234.5], [695.49, 176.52], [748.91, 176.52], [776.07, 175.73], [712.05, 175.74]]}, {\"category\": \"car\", \"corners_3d\": [[678.57, 215.54], [719.56, 215.64], [731.66, 220.99], [684.41, 220.85], [678.57, 177.9], [719.56, 177.89], [731.66, 177.5], [684.41, 177.51]]}, {\"category\": \"car\", \"corners_3d\": [[666.84, 209.55], [704.23, 209.61], [714.31, 213.73], [671.62, 213.65], [666.84, 177.69], [704.23, 177.68], [714.31, 177.28], [671.62, 177.29]]}]\n```" - } - ] - }, - { - "id": 796, - "data_source": "Mono3DRefer", - "scene_name": "000442", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-32.33, 363.29], [30.05, 325.97], [461.37, 295.79], [494.13, 318.75], [-32.33, 159.13], [30.05, 161.82], [461.37, 163.99], [494.13, 162.34]]}, {\"category\": \"car\", \"corners_3d\": [[1076.16, 283.05], [1220.1, 290.68], [1351.82, 344.64], [1135.94, 328.89], [1076.16, 174.22], [1220.1, 174.31], [1351.82, 174.98], [1135.94, 174.78]]}, {\"category\": \"car\", \"corners_3d\": [[709.24, 251.6], [633.35, 247.68], [696.94, 233.77], [760.81, 236.34], [709.24, 181.21], [633.35, 180.79], [696.94, 179.31], [760.81, 179.59]]}, {\"category\": \"car\", \"corners_3d\": [[1007.57, 226.98], [1073.01, 228.82], [1094.0, 238.37], [1017.02, 235.87], [1007.57, 174.58], [1073.01, 174.64], [1094.0, 174.94], [1017.02, 174.86]]}, {\"category\": \"car\", \"corners_3d\": [[820.97, 215.8], [773.86, 214.53], [797.73, 209.81], [840.0, 210.8], [820.97, 172.28], [773.86, 172.3], [797.73, 172.36], [840.0, 172.35]]}, {\"category\": \"car\", \"corners_3d\": [[844.27, 205.72], [807.89, 204.97], [823.96, 202.19], [857.47, 202.82], [844.27, 170.48], [807.89, 170.53], [823.96, 170.74], [857.47, 170.69]]}]\n```" - } - ] - }, - { - "id": 797, - "data_source": "Mono3DRefer", - "scene_name": "000442", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1064.51, 214.16], [1040.87, 213.72], [1035.74, 212.17], [1058.42, 212.58], [1064.51, 158.31], [1040.87, 158.47], [1035.74, 159.01], [1058.42, 158.87]]}]\n```" - } - ] - }, - { - "id": 798, - "data_source": "Mono3DRefer", - "scene_name": "000442", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-32.33, 363.29], [30.05, 325.97], [461.37, 295.79], [494.13, 318.75], [-32.33, 159.13], [30.05, 161.82], [461.37, 163.99], [494.13, 162.34]]}, {\"category\": \"car\", \"corners_3d\": [[1076.16, 283.05], [1220.1, 290.68], [1351.82, 344.64], [1135.94, 328.89], [1076.16, 174.22], [1220.1, 174.31], [1351.82, 174.98], [1135.94, 174.78]]}, {\"category\": \"car\", \"corners_3d\": [[709.24, 251.6], [633.35, 247.68], [696.94, 233.77], [760.81, 236.34], [709.24, 181.21], [633.35, 180.79], [696.94, 179.31], [760.81, 179.59]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1064.51, 214.16], [1040.87, 213.72], [1035.74, 212.17], [1058.42, 212.58], [1064.51, 158.31], [1040.87, 158.47], [1035.74, 159.01], [1058.42, 158.87]]}, {\"category\": \"car\", \"corners_3d\": [[1007.57, 226.98], [1073.01, 228.82], [1094.0, 238.37], [1017.02, 235.87], [1007.57, 174.58], [1073.01, 174.64], [1094.0, 174.94], [1017.02, 174.86]]}, {\"category\": \"car\", \"corners_3d\": [[820.97, 215.8], [773.86, 214.53], [797.73, 209.81], [840.0, 210.8], [820.97, 172.28], [773.86, 172.3], [797.73, 172.36], [840.0, 172.35]]}, {\"category\": \"car\", \"corners_3d\": [[844.27, 205.72], [807.89, 204.97], [823.96, 202.19], [857.47, 202.82], [844.27, 170.48], [807.89, 170.53], [823.96, 170.74], [857.47, 170.69]]}]\n```" - } - ] - }, - { - "id": 799, - "data_source": "Mono3DRefer", - "scene_name": "000444", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[535.65, 233.52], [479.73, 234.08], [487.21, 223.24], [533.25, 222.86], [535.65, 184.63], [479.73, 184.74], [487.21, 182.63], [533.25, 182.56]]}]\n```" - } - ] - }, - { - "id": 800, - "data_source": "Mono3DRefer", - "scene_name": "000445", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[861.5, 325.42], [935.43, 325.43], [965.18, 339.39], [884.49, 339.37], [861.5, 177.62], [935.43, 177.62], [965.18, 178.05], [884.49, 178.05]]}]\n```" - } - ] - }, - { - "id": 801, - "data_source": "Mono3DRefer", - "scene_name": "000445", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[42.28, 296.07], [175.33, 281.7], [309.09, 312.44], [155.37, 337.0], [42.28, 190.65], [175.33, 188.58], [309.09, 193.02], [155.37, 196.57]]}, {\"category\": \"car\", \"corners_3d\": [[415.66, 241.96], [348.54, 247.32], [277.7, 237.04], [340.58, 233.02], [415.66, 175.1], [348.54, 175.27], [277.7, 174.94], [340.58, 174.81]]}, {\"category\": \"car\", \"corners_3d\": [[469.38, 241.77], [547.1, 241.92], [524.29, 262.26], [423.81, 262.0], [469.38, 185.82], [547.1, 185.85], [524.29, 189.67], [423.81, 189.63]]}, {\"category\": \"car\", \"corners_3d\": [[499.62, 227.04], [553.26, 226.88], [547.54, 237.92], [482.88, 238.15], [499.62, 177.89], [553.26, 177.88], [547.54, 178.9], [482.88, 178.92]]}, {\"category\": \"car\", \"corners_3d\": [[495.51, 217.21], [461.48, 219.26], [406.44, 215.57], [440.13, 213.82], [495.51, 179.22], [461.48, 179.52], [406.44, 178.99], [440.13, 178.74]]}, {\"category\": \"car\", \"corners_3d\": [[737.47, 266.3], [841.86, 266.58], [949.39, 312.96], [793.09, 312.34], [737.47, 178.26], [841.86, 178.28], [949.39, 180.96], [793.09, 180.92]]}, {\"category\": \"car\", \"corners_3d\": [[697.7, 237.72], [770.76, 237.72], [821.64, 258.27], [725.44, 258.26], [697.7, 181.0], [770.76, 181.0], [821.64, 183.58], [725.44, 183.58]]}, {\"category\": \"car\", \"corners_3d\": [[681.28, 219.84], [728.48, 219.69], [755.58, 227.9], [700.17, 228.12], [681.28, 169.62], [728.48, 169.63], [755.58, 169.07], [700.17, 169.05]]}, {\"category\": \"car\", \"corners_3d\": [[643.2, 197.56], [670.68, 197.53], [678.66, 199.93], [648.53, 199.96], [643.2, 173.03], [670.68, 173.03], [678.66, 173.04], [648.53, 173.04]]}, {\"category\": \"car\", \"corners_3d\": [[594.78, 203.39], [628.8, 203.4], [630.1, 206.87], [592.23, 206.85], [594.78, 172.44], [628.8, 172.44], [630.1, 172.39], [592.23, 172.39]]}, {\"category\": \"car\", \"corners_3d\": [[230.94, 211.66], [190.96, 211.58], [242.78, 207.09], [278.22, 207.15], [230.94, 175.5], [190.96, 175.5], [242.78, 175.19], [278.22, 175.2]]}, {\"category\": \"car\", \"corners_3d\": [[589.36, 216.02], [640.67, 215.94], [649.31, 223.07], [589.51, 223.19], [589.36, 174.67], [640.67, 174.66], [649.31, 174.96], [589.51, 174.97]]}, {\"category\": \"car\", \"corners_3d\": [[-278.63, 352.9], [-37.1, 321.7], [125.8, 385.14], [-199.0, 454.81], [-278.63, 186.58], [-37.1, 184.2], [125.8, 189.04], [-199.0, 194.35]]}]\n```" - } - ] - }, - { - "id": 802, - "data_source": "Mono3DRefer", - "scene_name": "000445", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-123.74, 318.79], [-301.62, 318.82], [33.34, 265.13], [145.75, 265.11], [-123.74, 128.4], [-301.62, 128.39], [33.34, 144.74], [145.75, 144.75]]}]\n```" - } - ] - }, - { - "id": 803, - "data_source": "Mono3DRefer", - "scene_name": "000445", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[634.75, 189.8], [669.6, 189.8], [679.92, 192.74], [639.02, 192.74], [634.75, 143.56], [669.6, 143.56], [679.92, 138.48], [639.02, 138.48]]}]\n```" - } - ] - }, - { - "id": 804, - "data_source": "Mono3DRefer", - "scene_name": "000445", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[42.28, 296.07], [175.33, 281.7], [309.09, 312.44], [155.37, 337.0], [42.28, 190.65], [175.33, 188.58], [309.09, 193.02], [155.37, 196.57]]}, {\"category\": \"car\", \"corners_3d\": [[415.66, 241.96], [348.54, 247.32], [277.7, 237.04], [340.58, 233.02], [415.66, 175.1], [348.54, 175.27], [277.7, 174.94], [340.58, 174.81]]}, {\"category\": \"car\", \"corners_3d\": [[469.38, 241.77], [547.1, 241.92], [524.29, 262.26], [423.81, 262.0], [469.38, 185.82], [547.1, 185.85], [524.29, 189.67], [423.81, 189.63]]}, {\"category\": \"car\", \"corners_3d\": [[499.62, 227.04], [553.26, 226.88], [547.54, 237.92], [482.88, 238.15], [499.62, 177.89], [553.26, 177.88], [547.54, 178.9], [482.88, 178.92]]}, {\"category\": \"car\", \"corners_3d\": [[495.51, 217.21], [461.48, 219.26], [406.44, 215.57], [440.13, 213.82], [495.51, 179.22], [461.48, 179.52], [406.44, 178.99], [440.13, 178.74]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[861.5, 325.42], [935.43, 325.43], [965.18, 339.39], [884.49, 339.37], [861.5, 177.62], [935.43, 177.62], [965.18, 178.05], [884.49, 178.05]]}, {\"category\": \"car\", \"corners_3d\": [[737.47, 266.3], [841.86, 266.58], [949.39, 312.96], [793.09, 312.34], [737.47, 178.26], [841.86, 178.28], [949.39, 180.96], [793.09, 180.92]]}, {\"category\": \"car\", \"corners_3d\": [[697.7, 237.72], [770.76, 237.72], [821.64, 258.27], [725.44, 258.26], [697.7, 181.0], [770.76, 181.0], [821.64, 183.58], [725.44, 183.58]]}, {\"category\": \"car\", \"corners_3d\": [[681.28, 219.84], [728.48, 219.69], [755.58, 227.9], [700.17, 228.12], [681.28, 169.62], [728.48, 169.63], [755.58, 169.07], [700.17, 169.05]]}, {\"category\": \"car\", \"corners_3d\": [[643.2, 197.56], [670.68, 197.53], [678.66, 199.93], [648.53, 199.96], [643.2, 173.03], [670.68, 173.03], [678.66, 173.04], [648.53, 173.04]]}, {\"category\": \"car\", \"corners_3d\": [[594.78, 203.39], [628.8, 203.4], [630.1, 206.87], [592.23, 206.85], [594.78, 172.44], [628.8, 172.44], [630.1, 172.39], [592.23, 172.39]]}, {\"category\": \"car\", \"corners_3d\": [[230.94, 211.66], [190.96, 211.58], [242.78, 207.09], [278.22, 207.15], [230.94, 175.5], [190.96, 175.5], [242.78, 175.19], [278.22, 175.2]]}, {\"category\": \"car\", \"corners_3d\": [[589.36, 216.02], [640.67, 215.94], [649.31, 223.07], [589.51, 223.19], [589.36, 174.67], [640.67, 174.66], [649.31, 174.96], [589.51, 174.97]]}, {\"category\": \"car\", \"corners_3d\": [[-278.63, 352.9], [-37.1, 321.7], [125.8, 385.14], [-199.0, 454.81], [-278.63, 186.58], [-37.1, 184.2], [125.8, 189.04], [-199.0, 194.35]]}]\n```" - } - ] - }, - { - "id": 805, - "data_source": "Mono3DRefer", - "scene_name": "000445", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[861.5, 325.42], [935.43, 325.43], [965.18, 339.39], [884.49, 339.37], [861.5, 177.62], [935.43, 177.62], [965.18, 178.05], [884.49, 178.05]]}, {\"category\": \"van\", \"corners_3d\": [[-123.74, 318.79], [-301.62, 318.82], [33.34, 265.13], [145.75, 265.11], [-123.74, 128.4], [-301.62, 128.39], [33.34, 144.74], [145.75, 144.75]]}]\n```" - } - ] - }, - { - "id": 806, - "data_source": "Mono3DRefer", - "scene_name": "000445", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[861.5, 325.42], [935.43, 325.43], [965.18, 339.39], [884.49, 339.37], [861.5, 177.62], [935.43, 177.62], [965.18, 178.05], [884.49, 178.05]]}, {\"category\": \"truck\", \"corners_3d\": [[634.75, 189.8], [669.6, 189.8], [679.92, 192.74], [639.02, 192.74], [634.75, 143.56], [669.6, 143.56], [679.92, 138.48], [639.02, 138.48]]}]\n```" - } - ] - }, - { - "id": 807, - "data_source": "Mono3DRefer", - "scene_name": "000445", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[42.28, 296.07], [175.33, 281.7], [309.09, 312.44], [155.37, 337.0], [42.28, 190.65], [175.33, 188.58], [309.09, 193.02], [155.37, 196.57]]}, {\"category\": \"car\", \"corners_3d\": [[415.66, 241.96], [348.54, 247.32], [277.7, 237.04], [340.58, 233.02], [415.66, 175.1], [348.54, 175.27], [277.7, 174.94], [340.58, 174.81]]}, {\"category\": \"car\", \"corners_3d\": [[469.38, 241.77], [547.1, 241.92], [524.29, 262.26], [423.81, 262.0], [469.38, 185.82], [547.1, 185.85], [524.29, 189.67], [423.81, 189.63]]}, {\"category\": \"car\", \"corners_3d\": [[499.62, 227.04], [553.26, 226.88], [547.54, 237.92], [482.88, 238.15], [499.62, 177.89], [553.26, 177.88], [547.54, 178.9], [482.88, 178.92]]}, {\"category\": \"car\", \"corners_3d\": [[495.51, 217.21], [461.48, 219.26], [406.44, 215.57], [440.13, 213.82], [495.51, 179.22], [461.48, 179.52], [406.44, 178.99], [440.13, 178.74]]}, {\"category\": \"car\", \"corners_3d\": [[737.47, 266.3], [841.86, 266.58], [949.39, 312.96], [793.09, 312.34], [737.47, 178.26], [841.86, 178.28], [949.39, 180.96], [793.09, 180.92]]}, {\"category\": \"car\", \"corners_3d\": [[697.7, 237.72], [770.76, 237.72], [821.64, 258.27], [725.44, 258.26], [697.7, 181.0], [770.76, 181.0], [821.64, 183.58], [725.44, 183.58]]}, {\"category\": \"car\", \"corners_3d\": [[681.28, 219.84], [728.48, 219.69], [755.58, 227.9], [700.17, 228.12], [681.28, 169.62], [728.48, 169.63], [755.58, 169.07], [700.17, 169.05]]}, {\"category\": \"car\", \"corners_3d\": [[643.2, 197.56], [670.68, 197.53], [678.66, 199.93], [648.53, 199.96], [643.2, 173.03], [670.68, 173.03], [678.66, 173.04], [648.53, 173.04]]}, {\"category\": \"van\", \"corners_3d\": [[-123.74, 318.79], [-301.62, 318.82], [33.34, 265.13], [145.75, 265.11], [-123.74, 128.4], [-301.62, 128.39], [33.34, 144.74], [145.75, 144.75]]}, {\"category\": \"car\", \"corners_3d\": [[594.78, 203.39], [628.8, 203.4], [630.1, 206.87], [592.23, 206.85], [594.78, 172.44], [628.8, 172.44], [630.1, 172.39], [592.23, 172.39]]}, {\"category\": \"car\", \"corners_3d\": [[230.94, 211.66], [190.96, 211.58], [242.78, 207.09], [278.22, 207.15], [230.94, 175.5], [190.96, 175.5], [242.78, 175.19], [278.22, 175.2]]}, {\"category\": \"car\", \"corners_3d\": [[589.36, 216.02], [640.67, 215.94], [649.31, 223.07], [589.51, 223.19], [589.36, 174.67], [640.67, 174.66], [649.31, 174.96], [589.51, 174.97]]}, {\"category\": \"car\", \"corners_3d\": [[-278.63, 352.9], [-37.1, 321.7], [125.8, 385.14], [-199.0, 454.81], [-278.63, 186.58], [-37.1, 184.2], [125.8, 189.04], [-199.0, 194.35]]}]\n```" - } - ] - }, - { - "id": 808, - "data_source": "Mono3DRefer", - "scene_name": "000445", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[42.28, 296.07], [175.33, 281.7], [309.09, 312.44], [155.37, 337.0], [42.28, 190.65], [175.33, 188.58], [309.09, 193.02], [155.37, 196.57]]}, {\"category\": \"car\", \"corners_3d\": [[415.66, 241.96], [348.54, 247.32], [277.7, 237.04], [340.58, 233.02], [415.66, 175.1], [348.54, 175.27], [277.7, 174.94], [340.58, 174.81]]}, {\"category\": \"car\", \"corners_3d\": [[469.38, 241.77], [547.1, 241.92], [524.29, 262.26], [423.81, 262.0], [469.38, 185.82], [547.1, 185.85], [524.29, 189.67], [423.81, 189.63]]}, {\"category\": \"car\", \"corners_3d\": [[499.62, 227.04], [553.26, 226.88], [547.54, 237.92], [482.88, 238.15], [499.62, 177.89], [553.26, 177.88], [547.54, 178.9], [482.88, 178.92]]}, {\"category\": \"car\", \"corners_3d\": [[495.51, 217.21], [461.48, 219.26], [406.44, 215.57], [440.13, 213.82], [495.51, 179.22], [461.48, 179.52], [406.44, 178.99], [440.13, 178.74]]}, {\"category\": \"car\", \"corners_3d\": [[737.47, 266.3], [841.86, 266.58], [949.39, 312.96], [793.09, 312.34], [737.47, 178.26], [841.86, 178.28], [949.39, 180.96], [793.09, 180.92]]}, {\"category\": \"car\", \"corners_3d\": [[697.7, 237.72], [770.76, 237.72], [821.64, 258.27], [725.44, 258.26], [697.7, 181.0], [770.76, 181.0], [821.64, 183.58], [725.44, 183.58]]}, {\"category\": \"car\", \"corners_3d\": [[681.28, 219.84], [728.48, 219.69], [755.58, 227.9], [700.17, 228.12], [681.28, 169.62], [728.48, 169.63], [755.58, 169.07], [700.17, 169.05]]}, {\"category\": \"car\", \"corners_3d\": [[643.2, 197.56], [670.68, 197.53], [678.66, 199.93], [648.53, 199.96], [643.2, 173.03], [670.68, 173.03], [678.66, 173.04], [648.53, 173.04]]}, {\"category\": \"truck\", \"corners_3d\": [[634.75, 189.8], [669.6, 189.8], [679.92, 192.74], [639.02, 192.74], [634.75, 143.56], [669.6, 143.56], [679.92, 138.48], [639.02, 138.48]]}, {\"category\": \"car\", \"corners_3d\": [[594.78, 203.39], [628.8, 203.4], [630.1, 206.87], [592.23, 206.85], [594.78, 172.44], [628.8, 172.44], [630.1, 172.39], [592.23, 172.39]]}, {\"category\": \"car\", \"corners_3d\": [[230.94, 211.66], [190.96, 211.58], [242.78, 207.09], [278.22, 207.15], [230.94, 175.5], [190.96, 175.5], [242.78, 175.19], [278.22, 175.2]]}, {\"category\": \"car\", \"corners_3d\": [[589.36, 216.02], [640.67, 215.94], [649.31, 223.07], [589.51, 223.19], [589.36, 174.67], [640.67, 174.66], [649.31, 174.96], [589.51, 174.97]]}, {\"category\": \"car\", \"corners_3d\": [[-278.63, 352.9], [-37.1, 321.7], [125.8, 385.14], [-199.0, 454.81], [-278.63, 186.58], [-37.1, 184.2], [125.8, 189.04], [-199.0, 194.35]]}]\n```" - } - ] - }, - { - "id": 809, - "data_source": "Mono3DRefer", - "scene_name": "000445", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[634.75, 189.8], [669.6, 189.8], [679.92, 192.74], [639.02, 192.74], [634.75, 143.56], [669.6, 143.56], [679.92, 138.48], [639.02, 138.48]]}, {\"category\": \"van\", \"corners_3d\": [[-123.74, 318.79], [-301.62, 318.82], [33.34, 265.13], [145.75, 265.11], [-123.74, 128.4], [-301.62, 128.39], [33.34, 144.74], [145.75, 144.75]]}]\n```" - } - ] - }, - { - "id": 810, - "data_source": "Mono3DRefer", - "scene_name": "000445", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[42.28, 296.07], [175.33, 281.7], [309.09, 312.44], [155.37, 337.0], [42.28, 190.65], [175.33, 188.58], [309.09, 193.02], [155.37, 196.57]]}, {\"category\": \"car\", \"corners_3d\": [[415.66, 241.96], [348.54, 247.32], [277.7, 237.04], [340.58, 233.02], [415.66, 175.1], [348.54, 175.27], [277.7, 174.94], [340.58, 174.81]]}, {\"category\": \"car\", \"corners_3d\": [[469.38, 241.77], [547.1, 241.92], [524.29, 262.26], [423.81, 262.0], [469.38, 185.82], [547.1, 185.85], [524.29, 189.67], [423.81, 189.63]]}, {\"category\": \"car\", \"corners_3d\": [[499.62, 227.04], [553.26, 226.88], [547.54, 237.92], [482.88, 238.15], [499.62, 177.89], [553.26, 177.88], [547.54, 178.9], [482.88, 178.92]]}, {\"category\": \"car\", \"corners_3d\": [[495.51, 217.21], [461.48, 219.26], [406.44, 215.57], [440.13, 213.82], [495.51, 179.22], [461.48, 179.52], [406.44, 178.99], [440.13, 178.74]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[861.5, 325.42], [935.43, 325.43], [965.18, 339.39], [884.49, 339.37], [861.5, 177.62], [935.43, 177.62], [965.18, 178.05], [884.49, 178.05]]}, {\"category\": \"car\", \"corners_3d\": [[737.47, 266.3], [841.86, 266.58], [949.39, 312.96], [793.09, 312.34], [737.47, 178.26], [841.86, 178.28], [949.39, 180.96], [793.09, 180.92]]}, {\"category\": \"car\", \"corners_3d\": [[697.7, 237.72], [770.76, 237.72], [821.64, 258.27], [725.44, 258.26], [697.7, 181.0], [770.76, 181.0], [821.64, 183.58], [725.44, 183.58]]}, {\"category\": \"car\", \"corners_3d\": [[681.28, 219.84], [728.48, 219.69], [755.58, 227.9], [700.17, 228.12], [681.28, 169.62], [728.48, 169.63], [755.58, 169.07], [700.17, 169.05]]}, {\"category\": \"car\", \"corners_3d\": [[643.2, 197.56], [670.68, 197.53], [678.66, 199.93], [648.53, 199.96], [643.2, 173.03], [670.68, 173.03], [678.66, 173.04], [648.53, 173.04]]}, {\"category\": \"van\", \"corners_3d\": [[-123.74, 318.79], [-301.62, 318.82], [33.34, 265.13], [145.75, 265.11], [-123.74, 128.4], [-301.62, 128.39], [33.34, 144.74], [145.75, 144.75]]}, {\"category\": \"car\", \"corners_3d\": [[594.78, 203.39], [628.8, 203.4], [630.1, 206.87], [592.23, 206.85], [594.78, 172.44], [628.8, 172.44], [630.1, 172.39], [592.23, 172.39]]}, {\"category\": \"car\", \"corners_3d\": [[230.94, 211.66], [190.96, 211.58], [242.78, 207.09], [278.22, 207.15], [230.94, 175.5], [190.96, 175.5], [242.78, 175.19], [278.22, 175.2]]}, {\"category\": \"car\", \"corners_3d\": [[589.36, 216.02], [640.67, 215.94], [649.31, 223.07], [589.51, 223.19], [589.36, 174.67], [640.67, 174.66], [649.31, 174.96], [589.51, 174.97]]}, {\"category\": \"car\", \"corners_3d\": [[-278.63, 352.9], [-37.1, 321.7], [125.8, 385.14], [-199.0, 454.81], [-278.63, 186.58], [-37.1, 184.2], [125.8, 189.04], [-199.0, 194.35]]}]\n```" - } - ] - }, - { - "id": 811, - "data_source": "Mono3DRefer", - "scene_name": "000445", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[42.28, 296.07], [175.33, 281.7], [309.09, 312.44], [155.37, 337.0], [42.28, 190.65], [175.33, 188.58], [309.09, 193.02], [155.37, 196.57]]}, {\"category\": \"car\", \"corners_3d\": [[415.66, 241.96], [348.54, 247.32], [277.7, 237.04], [340.58, 233.02], [415.66, 175.1], [348.54, 175.27], [277.7, 174.94], [340.58, 174.81]]}, {\"category\": \"car\", \"corners_3d\": [[469.38, 241.77], [547.1, 241.92], [524.29, 262.26], [423.81, 262.0], [469.38, 185.82], [547.1, 185.85], [524.29, 189.67], [423.81, 189.63]]}, {\"category\": \"car\", \"corners_3d\": [[499.62, 227.04], [553.26, 226.88], [547.54, 237.92], [482.88, 238.15], [499.62, 177.89], [553.26, 177.88], [547.54, 178.9], [482.88, 178.92]]}, {\"category\": \"car\", \"corners_3d\": [[495.51, 217.21], [461.48, 219.26], [406.44, 215.57], [440.13, 213.82], [495.51, 179.22], [461.48, 179.52], [406.44, 178.99], [440.13, 178.74]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[861.5, 325.42], [935.43, 325.43], [965.18, 339.39], [884.49, 339.37], [861.5, 177.62], [935.43, 177.62], [965.18, 178.05], [884.49, 178.05]]}, {\"category\": \"car\", \"corners_3d\": [[737.47, 266.3], [841.86, 266.58], [949.39, 312.96], [793.09, 312.34], [737.47, 178.26], [841.86, 178.28], [949.39, 180.96], [793.09, 180.92]]}, {\"category\": \"car\", \"corners_3d\": [[697.7, 237.72], [770.76, 237.72], [821.64, 258.27], [725.44, 258.26], [697.7, 181.0], [770.76, 181.0], [821.64, 183.58], [725.44, 183.58]]}, {\"category\": \"car\", \"corners_3d\": [[681.28, 219.84], [728.48, 219.69], [755.58, 227.9], [700.17, 228.12], [681.28, 169.62], [728.48, 169.63], [755.58, 169.07], [700.17, 169.05]]}, {\"category\": \"car\", \"corners_3d\": [[643.2, 197.56], [670.68, 197.53], [678.66, 199.93], [648.53, 199.96], [643.2, 173.03], [670.68, 173.03], [678.66, 173.04], [648.53, 173.04]]}, {\"category\": \"truck\", \"corners_3d\": [[634.75, 189.8], [669.6, 189.8], [679.92, 192.74], [639.02, 192.74], [634.75, 143.56], [669.6, 143.56], [679.92, 138.48], [639.02, 138.48]]}, {\"category\": \"car\", \"corners_3d\": [[594.78, 203.39], [628.8, 203.4], [630.1, 206.87], [592.23, 206.85], [594.78, 172.44], [628.8, 172.44], [630.1, 172.39], [592.23, 172.39]]}, {\"category\": \"car\", \"corners_3d\": [[230.94, 211.66], [190.96, 211.58], [242.78, 207.09], [278.22, 207.15], [230.94, 175.5], [190.96, 175.5], [242.78, 175.19], [278.22, 175.2]]}, {\"category\": \"car\", \"corners_3d\": [[589.36, 216.02], [640.67, 215.94], [649.31, 223.07], [589.51, 223.19], [589.36, 174.67], [640.67, 174.66], [649.31, 174.96], [589.51, 174.97]]}, {\"category\": \"car\", \"corners_3d\": [[-278.63, 352.9], [-37.1, 321.7], [125.8, 385.14], [-199.0, 454.81], [-278.63, 186.58], [-37.1, 184.2], [125.8, 189.04], [-199.0, 194.35]]}]\n```" - } - ] - }, - { - "id": 812, - "data_source": "Mono3DRefer", - "scene_name": "000445", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[861.5, 325.42], [935.43, 325.43], [965.18, 339.39], [884.49, 339.37], [861.5, 177.62], [935.43, 177.62], [965.18, 178.05], [884.49, 178.05]]}, {\"category\": \"truck\", \"corners_3d\": [[634.75, 189.8], [669.6, 189.8], [679.92, 192.74], [639.02, 192.74], [634.75, 143.56], [669.6, 143.56], [679.92, 138.48], [639.02, 138.48]]}, {\"category\": \"van\", \"corners_3d\": [[-123.74, 318.79], [-301.62, 318.82], [33.34, 265.13], [145.75, 265.11], [-123.74, 128.4], [-301.62, 128.39], [33.34, 144.74], [145.75, 144.75]]}]\n```" - } - ] - }, - { - "id": 813, - "data_source": "Mono3DRefer", - "scene_name": "000445", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[42.28, 296.07], [175.33, 281.7], [309.09, 312.44], [155.37, 337.0], [42.28, 190.65], [175.33, 188.58], [309.09, 193.02], [155.37, 196.57]]}, {\"category\": \"car\", \"corners_3d\": [[415.66, 241.96], [348.54, 247.32], [277.7, 237.04], [340.58, 233.02], [415.66, 175.1], [348.54, 175.27], [277.7, 174.94], [340.58, 174.81]]}, {\"category\": \"car\", \"corners_3d\": [[469.38, 241.77], [547.1, 241.92], [524.29, 262.26], [423.81, 262.0], [469.38, 185.82], [547.1, 185.85], [524.29, 189.67], [423.81, 189.63]]}, {\"category\": \"car\", \"corners_3d\": [[499.62, 227.04], [553.26, 226.88], [547.54, 237.92], [482.88, 238.15], [499.62, 177.89], [553.26, 177.88], [547.54, 178.9], [482.88, 178.92]]}, {\"category\": \"car\", \"corners_3d\": [[495.51, 217.21], [461.48, 219.26], [406.44, 215.57], [440.13, 213.82], [495.51, 179.22], [461.48, 179.52], [406.44, 178.99], [440.13, 178.74]]}, {\"category\": \"car\", \"corners_3d\": [[737.47, 266.3], [841.86, 266.58], [949.39, 312.96], [793.09, 312.34], [737.47, 178.26], [841.86, 178.28], [949.39, 180.96], [793.09, 180.92]]}, {\"category\": \"car\", \"corners_3d\": [[697.7, 237.72], [770.76, 237.72], [821.64, 258.27], [725.44, 258.26], [697.7, 181.0], [770.76, 181.0], [821.64, 183.58], [725.44, 183.58]]}, {\"category\": \"car\", \"corners_3d\": [[681.28, 219.84], [728.48, 219.69], [755.58, 227.9], [700.17, 228.12], [681.28, 169.62], [728.48, 169.63], [755.58, 169.07], [700.17, 169.05]]}, {\"category\": \"car\", \"corners_3d\": [[643.2, 197.56], [670.68, 197.53], [678.66, 199.93], [648.53, 199.96], [643.2, 173.03], [670.68, 173.03], [678.66, 173.04], [648.53, 173.04]]}, {\"category\": \"truck\", \"corners_3d\": [[634.75, 189.8], [669.6, 189.8], [679.92, 192.74], [639.02, 192.74], [634.75, 143.56], [669.6, 143.56], [679.92, 138.48], [639.02, 138.48]]}, {\"category\": \"van\", \"corners_3d\": [[-123.74, 318.79], [-301.62, 318.82], [33.34, 265.13], [145.75, 265.11], [-123.74, 128.4], [-301.62, 128.39], [33.34, 144.74], [145.75, 144.75]]}, {\"category\": \"car\", \"corners_3d\": [[594.78, 203.39], [628.8, 203.4], [630.1, 206.87], [592.23, 206.85], [594.78, 172.44], [628.8, 172.44], [630.1, 172.39], [592.23, 172.39]]}, {\"category\": \"car\", \"corners_3d\": [[230.94, 211.66], [190.96, 211.58], [242.78, 207.09], [278.22, 207.15], [230.94, 175.5], [190.96, 175.5], [242.78, 175.19], [278.22, 175.2]]}, {\"category\": \"car\", \"corners_3d\": [[589.36, 216.02], [640.67, 215.94], [649.31, 223.07], [589.51, 223.19], [589.36, 174.67], [640.67, 174.66], [649.31, 174.96], [589.51, 174.97]]}, {\"category\": \"car\", \"corners_3d\": [[-278.63, 352.9], [-37.1, 321.7], [125.8, 385.14], [-199.0, 454.81], [-278.63, 186.58], [-37.1, 184.2], [125.8, 189.04], [-199.0, 194.35]]}]\n```" - } - ] - }, - { - "id": 814, - "data_source": "Mono3DRefer", - "scene_name": "000445", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[42.28, 296.07], [175.33, 281.7], [309.09, 312.44], [155.37, 337.0], [42.28, 190.65], [175.33, 188.58], [309.09, 193.02], [155.37, 196.57]]}, {\"category\": \"car\", \"corners_3d\": [[415.66, 241.96], [348.54, 247.32], [277.7, 237.04], [340.58, 233.02], [415.66, 175.1], [348.54, 175.27], [277.7, 174.94], [340.58, 174.81]]}, {\"category\": \"car\", \"corners_3d\": [[469.38, 241.77], [547.1, 241.92], [524.29, 262.26], [423.81, 262.0], [469.38, 185.82], [547.1, 185.85], [524.29, 189.67], [423.81, 189.63]]}, {\"category\": \"car\", \"corners_3d\": [[499.62, 227.04], [553.26, 226.88], [547.54, 237.92], [482.88, 238.15], [499.62, 177.89], [553.26, 177.88], [547.54, 178.9], [482.88, 178.92]]}, {\"category\": \"car\", \"corners_3d\": [[495.51, 217.21], [461.48, 219.26], [406.44, 215.57], [440.13, 213.82], [495.51, 179.22], [461.48, 179.52], [406.44, 178.99], [440.13, 178.74]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[861.5, 325.42], [935.43, 325.43], [965.18, 339.39], [884.49, 339.37], [861.5, 177.62], [935.43, 177.62], [965.18, 178.05], [884.49, 178.05]]}, {\"category\": \"car\", \"corners_3d\": [[737.47, 266.3], [841.86, 266.58], [949.39, 312.96], [793.09, 312.34], [737.47, 178.26], [841.86, 178.28], [949.39, 180.96], [793.09, 180.92]]}, {\"category\": \"car\", \"corners_3d\": [[697.7, 237.72], [770.76, 237.72], [821.64, 258.27], [725.44, 258.26], [697.7, 181.0], [770.76, 181.0], [821.64, 183.58], [725.44, 183.58]]}, {\"category\": \"car\", \"corners_3d\": [[681.28, 219.84], [728.48, 219.69], [755.58, 227.9], [700.17, 228.12], [681.28, 169.62], [728.48, 169.63], [755.58, 169.07], [700.17, 169.05]]}, {\"category\": \"car\", \"corners_3d\": [[643.2, 197.56], [670.68, 197.53], [678.66, 199.93], [648.53, 199.96], [643.2, 173.03], [670.68, 173.03], [678.66, 173.04], [648.53, 173.04]]}, {\"category\": \"truck\", \"corners_3d\": [[634.75, 189.8], [669.6, 189.8], [679.92, 192.74], [639.02, 192.74], [634.75, 143.56], [669.6, 143.56], [679.92, 138.48], [639.02, 138.48]]}, {\"category\": \"van\", \"corners_3d\": [[-123.74, 318.79], [-301.62, 318.82], [33.34, 265.13], [145.75, 265.11], [-123.74, 128.4], [-301.62, 128.39], [33.34, 144.74], [145.75, 144.75]]}, {\"category\": \"car\", \"corners_3d\": [[594.78, 203.39], [628.8, 203.4], [630.1, 206.87], [592.23, 206.85], [594.78, 172.44], [628.8, 172.44], [630.1, 172.39], [592.23, 172.39]]}, {\"category\": \"car\", \"corners_3d\": [[230.94, 211.66], [190.96, 211.58], [242.78, 207.09], [278.22, 207.15], [230.94, 175.5], [190.96, 175.5], [242.78, 175.19], [278.22, 175.2]]}, {\"category\": \"car\", \"corners_3d\": [[589.36, 216.02], [640.67, 215.94], [649.31, 223.07], [589.51, 223.19], [589.36, 174.67], [640.67, 174.66], [649.31, 174.96], [589.51, 174.97]]}, {\"category\": \"car\", \"corners_3d\": [[-278.63, 352.9], [-37.1, 321.7], [125.8, 385.14], [-199.0, 454.81], [-278.63, 186.58], [-37.1, 184.2], [125.8, 189.04], [-199.0, 194.35]]}]\n```" - } - ] - }, - { - "id": 815, - "data_source": "Mono3DRefer", - "scene_name": "000446", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[482.39, 207.67], [469.58, 207.59], [478.69, 206.2], [491.0, 206.27], [482.39, 177.14], [469.58, 177.13], [478.69, 176.96], [491.0, 176.96]]}]\n```" - } - ] - }, - { - "id": 816, - "data_source": "Mono3DRefer", - "scene_name": "000446", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[403.63, 291.67], [280.36, 291.49], [365.14, 261.54], [457.35, 261.64], [403.63, 191.22], [280.36, 191.19], [365.14, 186.56], [457.35, 186.58]]}]\n```" - } - ] - }, - { - "id": 817, - "data_source": "Mono3DRefer", - "scene_name": "000446", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[482.39, 207.67], [469.58, 207.59], [478.69, 206.2], [491.0, 206.27], [482.39, 177.14], [469.58, 177.13], [478.69, 176.96], [491.0, 176.96]]}, {\"category\": \"car\", \"corners_3d\": [[403.63, 291.67], [280.36, 291.49], [365.14, 261.54], [457.35, 261.64], [403.63, 191.22], [280.36, 191.19], [365.14, 186.56], [457.35, 186.58]]}]\n```" - } - ] - }, - { - "id": 818, - "data_source": "Mono3DRefer", - "scene_name": "000456", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[508.49, 266.48], [431.49, 266.59], [466.43, 247.36], [527.61, 247.29], [508.49, 185.57], [431.49, 185.59], [466.43, 182.97], [527.61, 182.96]]}, {\"category\": \"car\", \"corners_3d\": [[555.74, 220.62], [529.51, 220.59], [537.33, 216.61], [561.38, 216.64], [555.74, 191.41], [529.51, 191.39], [537.33, 189.85], [561.38, 189.86]]}]\n```" - } - ] - }, - { - "id": 819, - "data_source": "Mono3DRefer", - "scene_name": "000458", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[572.69, 214.33], [601.65, 214.32], [601.51, 218.99], [569.28, 219.01], [572.69, 187.61], [601.65, 187.61], [601.51, 189.27], [569.28, 189.27]]}, {\"category\": \"car\", \"corners_3d\": [[888.75, 318.76], [834.48, 293.73], [1138.27, 291.06], [1252.32, 314.88], [888.75, 181.89], [834.48, 180.34], [1138.27, 180.17], [1252.32, 181.65]]}, {\"category\": \"car\", \"corners_3d\": [[1054.12, 271.95], [1110.82, 285.95], [821.12, 283.33], [800.71, 269.93], [1054.12, 183.04], [1110.82, 184.48], [821.12, 184.22], [800.71, 182.84]]}, {\"category\": \"car\", \"corners_3d\": [[891.33, 239.18], [909.59, 245.4], [729.16, 243.32], [726.44, 237.44], [891.33, 183.53], [909.59, 184.53], [729.16, 184.19], [726.44, 183.25]]}, {\"category\": \"car\", \"corners_3d\": [[795.83, 269.09], [778.97, 259.17], [970.4, 259.89], [1009.62, 269.98], [795.83, 182.79], [778.97, 181.77], [970.4, 181.84], [1009.62, 182.88]]}]\n```" - } - ] - }, - { - "id": 820, - "data_source": "Mono3DRefer", - "scene_name": "000460", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1089.24, 252.61], [1169.83, 264.77], [888.89, 264.44], [848.96, 252.37], [1089.24, 169.63], [1169.83, 167.79], [888.89, 167.84], [848.96, 169.66]]}]\n```" - } - ] - }, - { - "id": 821, - "data_source": "Mono3DRefer", - "scene_name": "000460", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[660.37, 243.98], [686.89, 243.98], [697.04, 251.82], [667.25, 251.81], [660.37, 161.22], [686.89, 161.22], [697.04, 158.84], [667.25, 158.84]]}, {\"category\": \"cyclist\", \"corners_3d\": [[569.83, 200.3], [558.25, 200.28], [561.78, 199.42], [572.85, 199.43], [569.83, 167.32], [558.25, 167.33], [561.78, 167.91], [572.85, 167.9]]}]\n```" - } - ] - }, - { - "id": 822, - "data_source": "Mono3DRefer", - "scene_name": "000460", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[321.2, 267.37], [284.08, 267.37], [306.03, 261.4], [340.6, 261.4], [321.2, 168.23], [284.08, 168.23], [306.03, 169.07], [340.6, 169.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[699.06, 216.4], [715.32, 216.39], [720.0, 217.82], [703.1, 217.83], [699.06, 160.48], [715.32, 160.49], [720.0, 159.69], [703.1, 159.69]]}]\n```" - } - ] - }, - { - "id": 823, - "data_source": "Mono3DRefer", - "scene_name": "000460", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[660.37, 243.98], [686.89, 243.98], [697.04, 251.82], [667.25, 251.81], [660.37, 161.22], [686.89, 161.22], [697.04, 158.84], [667.25, 158.84]]}, {\"category\": \"car\", \"corners_3d\": [[1089.24, 252.61], [1169.83, 264.77], [888.89, 264.44], [848.96, 252.37], [1089.24, 169.63], [1169.83, 167.79], [888.89, 167.84], [848.96, 169.66]]}, {\"category\": \"cyclist\", \"corners_3d\": [[569.83, 200.3], [558.25, 200.28], [561.78, 199.42], [572.85, 199.43], [569.83, 167.32], [558.25, 167.33], [561.78, 167.91], [572.85, 167.9]]}]\n```" - } - ] - }, - { - "id": 824, - "data_source": "Mono3DRefer", - "scene_name": "000460", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[321.2, 267.37], [284.08, 267.37], [306.03, 261.4], [340.6, 261.4], [321.2, 168.23], [284.08, 168.23], [306.03, 169.07], [340.6, 169.07]]}, {\"category\": \"car\", \"corners_3d\": [[1089.24, 252.61], [1169.83, 264.77], [888.89, 264.44], [848.96, 252.37], [1089.24, 169.63], [1169.83, 167.79], [888.89, 167.84], [848.96, 169.66]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[699.06, 216.4], [715.32, 216.39], [720.0, 217.82], [703.1, 217.83], [699.06, 160.48], [715.32, 160.49], [720.0, 159.69], [703.1, 159.69]]}]\n```" - } - ] - }, - { - "id": 825, - "data_source": "Mono3DRefer", - "scene_name": "000460", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[321.2, 267.37], [284.08, 267.37], [306.03, 261.4], [340.6, 261.4], [321.2, 168.23], [284.08, 168.23], [306.03, 169.07], [340.6, 169.07]]}, {\"category\": \"cyclist\", \"corners_3d\": [[660.37, 243.98], [686.89, 243.98], [697.04, 251.82], [667.25, 251.81], [660.37, 161.22], [686.89, 161.22], [697.04, 158.84], [667.25, 158.84]]}, {\"category\": \"cyclist\", \"corners_3d\": [[569.83, 200.3], [558.25, 200.28], [561.78, 199.42], [572.85, 199.43], [569.83, 167.32], [558.25, 167.33], [561.78, 167.91], [572.85, 167.9]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[699.06, 216.4], [715.32, 216.39], [720.0, 217.82], [703.1, 217.83], [699.06, 160.48], [715.32, 160.49], [720.0, 159.69], [703.1, 159.69]]}]\n```" - } - ] - }, - { - "id": 826, - "data_source": "Mono3DRefer", - "scene_name": "000460", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[321.2, 267.37], [284.08, 267.37], [306.03, 261.4], [340.6, 261.4], [321.2, 168.23], [284.08, 168.23], [306.03, 169.07], [340.6, 169.07]]}, {\"category\": \"cyclist\", \"corners_3d\": [[660.37, 243.98], [686.89, 243.98], [697.04, 251.82], [667.25, 251.81], [660.37, 161.22], [686.89, 161.22], [697.04, 158.84], [667.25, 158.84]]}, {\"category\": \"car\", \"corners_3d\": [[1089.24, 252.61], [1169.83, 264.77], [888.89, 264.44], [848.96, 252.37], [1089.24, 169.63], [1169.83, 167.79], [888.89, 167.84], [848.96, 169.66]]}, {\"category\": \"cyclist\", \"corners_3d\": [[569.83, 200.3], [558.25, 200.28], [561.78, 199.42], [572.85, 199.43], [569.83, 167.32], [558.25, 167.33], [561.78, 167.91], [572.85, 167.9]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[699.06, 216.4], [715.32, 216.39], [720.0, 217.82], [703.1, 217.83], [699.06, 160.48], [715.32, 160.49], [720.0, 159.69], [703.1, 159.69]]}]\n```" - } - ] - }, - { - "id": 827, - "data_source": "Mono3DRefer", - "scene_name": "000466", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-5.27, 258.66], [-65.48, 259.14], [27.51, 246.28], [78.35, 245.94], [-5.27, 206.81], [-65.48, 206.99], [27.51, 201.91], [78.35, 201.77]]}, {\"category\": \"car\", \"corners_3d\": [[214.01, 226.12], [249.0, 225.97], [207.96, 232.83], [168.3, 233.02], [214.01, 197.68], [249.0, 197.61], [207.96, 200.81], [168.3, 200.9]]}, {\"category\": \"car\", \"corners_3d\": [[156.02, 225.65], [191.78, 225.47], [144.39, 232.25], [103.83, 232.47], [156.02, 195.01], [191.78, 194.94], [144.39, 197.79], [103.83, 197.88]]}]\n```" - } - ] - }, - { - "id": 828, - "data_source": "Mono3DRefer", - "scene_name": "000468", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[-9.41, 235.78], [-33.03, 235.79], [10.66, 231.5], [32.67, 231.5], [-9.41, 182.76], [-33.03, 182.76], [10.66, 182.09], [32.67, 182.09]]}]\n```" - } - ] - }, - { - "id": 829, - "data_source": "Mono3DRefer", - "scene_name": "000468", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[725.9, 194.06], [697.09, 194.62], [591.44, 190.21], [616.72, 189.85], [725.9, 138.61], [697.09, 137.7], [591.44, 144.83], [616.72, 145.41]]}, {\"category\": \"tram\", \"corners_3d\": [[902.44, 201.89], [876.04, 202.96], [693.4, 195.93], [718.83, 195.3], [902.44, 137.57], [876.04, 136.27], [693.4, 144.81], [718.83, 145.58]]}]\n```" - } - ] - }, - { - "id": 830, - "data_source": "Mono3DRefer", - "scene_name": "000468", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[565.34, 293.82], [689.95, 293.84], [724.89, 346.8], [545.74, 346.77], [565.34, 174.27], [689.95, 174.27], [724.89, 174.89], [545.74, 174.89]]}, {\"category\": \"car\", \"corners_3d\": [[123.91, 259.01], [49.0, 259.01], [156.02, 242.55], [216.61, 242.54], [123.91, 191.2], [49.0, 191.2], [156.02, 187.7], [216.61, 187.7]]}, {\"category\": \"car\", \"corners_3d\": [[393.71, 212.88], [357.9, 212.86], [384.28, 208.78], [416.44, 208.79], [393.71, 180.03], [357.9, 180.03], [384.28, 179.3], [416.44, 179.3]]}, {\"category\": \"car\", \"corners_3d\": [[452.24, 201.41], [428.54, 201.4], [443.56, 199.12], [465.37, 199.13], [452.24, 178.39], [428.54, 178.39], [443.56, 177.94], [465.37, 177.94]]}]\n```" - } - ] - }, - { - "id": 831, - "data_source": "Mono3DRefer", - "scene_name": "000468", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[725.9, 194.06], [697.09, 194.62], [591.44, 190.21], [616.72, 189.85], [725.9, 138.61], [697.09, 137.7], [591.44, 144.83], [616.72, 145.41]]}, {\"category\": \"cyclist\", \"corners_3d\": [[-9.41, 235.78], [-33.03, 235.79], [10.66, 231.5], [32.67, 231.5], [-9.41, 182.76], [-33.03, 182.76], [10.66, 182.09], [32.67, 182.09]]}, {\"category\": \"tram\", \"corners_3d\": [[902.44, 201.89], [876.04, 202.96], [693.4, 195.93], [718.83, 195.3], [902.44, 137.57], [876.04, 136.27], [693.4, 144.81], [718.83, 145.58]]}]\n```" - } - ] - }, - { - "id": 832, - "data_source": "Mono3DRefer", - "scene_name": "000468", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[565.34, 293.82], [689.95, 293.84], [724.89, 346.8], [545.74, 346.77], [565.34, 174.27], [689.95, 174.27], [724.89, 174.89], [545.74, 174.89]]}, {\"category\": \"car\", \"corners_3d\": [[123.91, 259.01], [49.0, 259.01], [156.02, 242.55], [216.61, 242.54], [123.91, 191.2], [49.0, 191.2], [156.02, 187.7], [216.61, 187.7]]}, {\"category\": \"car\", \"corners_3d\": [[393.71, 212.88], [357.9, 212.86], [384.28, 208.78], [416.44, 208.79], [393.71, 180.03], [357.9, 180.03], [384.28, 179.3], [416.44, 179.3]]}, {\"category\": \"car\", \"corners_3d\": [[452.24, 201.41], [428.54, 201.4], [443.56, 199.12], [465.37, 199.13], [452.24, 178.39], [428.54, 178.39], [443.56, 177.94], [465.37, 177.94]]}, {\"category\": \"cyclist\", \"corners_3d\": [[-9.41, 235.78], [-33.03, 235.79], [10.66, 231.5], [32.67, 231.5], [-9.41, 182.76], [-33.03, 182.76], [10.66, 182.09], [32.67, 182.09]]}]\n```" - } - ] - }, - { - "id": 833, - "data_source": "Mono3DRefer", - "scene_name": "000468", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[565.34, 293.82], [689.95, 293.84], [724.89, 346.8], [545.74, 346.77], [565.34, 174.27], [689.95, 174.27], [724.89, 174.89], [545.74, 174.89]]}, {\"category\": \"tram\", \"corners_3d\": [[725.9, 194.06], [697.09, 194.62], [591.44, 190.21], [616.72, 189.85], [725.9, 138.61], [697.09, 137.7], [591.44, 144.83], [616.72, 145.41]]}, {\"category\": \"car\", \"corners_3d\": [[123.91, 259.01], [49.0, 259.01], [156.02, 242.55], [216.61, 242.54], [123.91, 191.2], [49.0, 191.2], [156.02, 187.7], [216.61, 187.7]]}, {\"category\": \"car\", \"corners_3d\": [[393.71, 212.88], [357.9, 212.86], [384.28, 208.78], [416.44, 208.79], [393.71, 180.03], [357.9, 180.03], [384.28, 179.3], [416.44, 179.3]]}, {\"category\": \"car\", \"corners_3d\": [[452.24, 201.41], [428.54, 201.4], [443.56, 199.12], [465.37, 199.13], [452.24, 178.39], [428.54, 178.39], [443.56, 177.94], [465.37, 177.94]]}, {\"category\": \"tram\", \"corners_3d\": [[902.44, 201.89], [876.04, 202.96], [693.4, 195.93], [718.83, 195.3], [902.44, 137.57], [876.04, 136.27], [693.4, 144.81], [718.83, 145.58]]}]\n```" - } - ] - }, - { - "id": 834, - "data_source": "Mono3DRefer", - "scene_name": "000468", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[565.34, 293.82], [689.95, 293.84], [724.89, 346.8], [545.74, 346.77], [565.34, 174.27], [689.95, 174.27], [724.89, 174.89], [545.74, 174.89]]}, {\"category\": \"tram\", \"corners_3d\": [[725.9, 194.06], [697.09, 194.62], [591.44, 190.21], [616.72, 189.85], [725.9, 138.61], [697.09, 137.7], [591.44, 144.83], [616.72, 145.41]]}, {\"category\": \"car\", \"corners_3d\": [[123.91, 259.01], [49.0, 259.01], [156.02, 242.55], [216.61, 242.54], [123.91, 191.2], [49.0, 191.2], [156.02, 187.7], [216.61, 187.7]]}, {\"category\": \"car\", \"corners_3d\": [[393.71, 212.88], [357.9, 212.86], [384.28, 208.78], [416.44, 208.79], [393.71, 180.03], [357.9, 180.03], [384.28, 179.3], [416.44, 179.3]]}, {\"category\": \"car\", \"corners_3d\": [[452.24, 201.41], [428.54, 201.4], [443.56, 199.12], [465.37, 199.13], [452.24, 178.39], [428.54, 178.39], [443.56, 177.94], [465.37, 177.94]]}, {\"category\": \"cyclist\", \"corners_3d\": [[-9.41, 235.78], [-33.03, 235.79], [10.66, 231.5], [32.67, 231.5], [-9.41, 182.76], [-33.03, 182.76], [10.66, 182.09], [32.67, 182.09]]}, {\"category\": \"tram\", \"corners_3d\": [[902.44, 201.89], [876.04, 202.96], [693.4, 195.93], [718.83, 195.3], [902.44, 137.57], [876.04, 136.27], [693.4, 144.81], [718.83, 145.58]]}]\n```" - } - ] - }, - { - "id": 835, - "data_source": "Mono3DRefer", - "scene_name": "000469", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-2.4, 384.77], [266.59, 362.7], [50.03, 910.7], [-1811.46, 1468.44], [-2.4, 123.97], [266.59, 130.08], [50.03, -21.61], [-1811.46, -175.99]]}]\n```" - } - ] - }, - { - "id": 836, - "data_source": "Mono3DRefer", - "scene_name": "000469", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[634.14, 201.58], [642.25, 201.16], [660.19, 201.47], [652.32, 201.9], [634.14, 145.24], [642.25, 145.94], [660.19, 145.42], [652.32, 144.69]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[298.4, 254.52], [270.05, 258.65], [236.04, 257.4], [265.74, 253.4], [298.4, 141.58], [270.05, 139.41], [236.04, 140.07], [265.74, 142.17]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[509.37, 206.15], [522.86, 205.67], [548.59, 206.56], [535.12, 207.08], [509.37, 140.5], [522.86, 141.25], [548.59, 139.86], [535.12, 139.05]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1433.42, 800.61], [1341.14, 891.24], [978.91, 652.7], [1075.01, 610.91], [1433.42, 155.93], [1341.14, 152.34], [978.91, 161.79], [1075.01, 163.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1424.0, 678.63], [1318.59, 725.84], [1017.79, 557.29], [1111.11, 534.14], [1424.0, 79.55], [1318.59, 69.98], [1017.79, 104.14], [1111.11, 108.84]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1424.26, 594.82], [1347.16, 628.59], [1122.72, 530.59], [1197.41, 509.63], [1424.26, 195.42], [1347.16, 196.63], [1122.72, 193.11], [1197.41, 192.35]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1524.45, 572.48], [1464.76, 612.93], [1186.21, 512.06], [1253.42, 487.75], [1524.45, 143.75], [1464.76, 139.95], [1186.21, 149.41], [1253.42, 151.69]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[961.09, 487.31], [949.45, 534.0], [765.44, 490.6], [797.49, 454.07], [961.09, 168.42], [949.45, 166.59], [765.44, 168.29], [797.49, 169.73]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[712.75, 348.69], [646.55, 352.89], [618.99, 331.64], [677.79, 328.41], [712.75, 146.17], [646.55, 145.32], [618.99, 149.65], [677.79, 150.32]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[846.07, 342.53], [786.41, 347.85], [736.51, 328.21], [790.77, 324.05], [846.07, 161.93], [786.41, 161.32], [736.51, 163.57], [790.77, 164.05]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[926.86, 333.78], [863.77, 337.74], [806.49, 317.32], [862.83, 314.3], [926.86, 138.47], [863.77, 137.38], [806.49, 142.98], [862.83, 143.81]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1009.58, 325.2], [958.24, 328.51], [890.78, 310.12], [937.19, 307.57], [1009.58, 140.14], [958.24, 139.21], [890.78, 144.34], [937.19, 145.05]]}]\n```" - } - ] - }, - { - "id": 837, - "data_source": "Mono3DRefer", - "scene_name": "000469", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[634.14, 201.58], [642.25, 201.16], [660.19, 201.47], [652.32, 201.9], [634.14, 145.24], [642.25, 145.94], [660.19, 145.42], [652.32, 144.69]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[298.4, 254.52], [270.05, 258.65], [236.04, 257.4], [265.74, 253.4], [298.4, 141.58], [270.05, 139.41], [236.04, 140.07], [265.74, 142.17]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[509.37, 206.15], [522.86, 205.67], [548.59, 206.56], [535.12, 207.08], [509.37, 140.5], [522.86, 141.25], [548.59, 139.86], [535.12, 139.05]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1433.42, 800.61], [1341.14, 891.24], [978.91, 652.7], [1075.01, 610.91], [1433.42, 155.93], [1341.14, 152.34], [978.91, 161.79], [1075.01, 163.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1424.0, 678.63], [1318.59, 725.84], [1017.79, 557.29], [1111.11, 534.14], [1424.0, 79.55], [1318.59, 69.98], [1017.79, 104.14], [1111.11, 108.84]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1424.26, 594.82], [1347.16, 628.59], [1122.72, 530.59], [1197.41, 509.63], [1424.26, 195.42], [1347.16, 196.63], [1122.72, 193.11], [1197.41, 192.35]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1524.45, 572.48], [1464.76, 612.93], [1186.21, 512.06], [1253.42, 487.75], [1524.45, 143.75], [1464.76, 139.95], [1186.21, 149.41], [1253.42, 151.69]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[961.09, 487.31], [949.45, 534.0], [765.44, 490.6], [797.49, 454.07], [961.09, 168.42], [949.45, 166.59], [765.44, 168.29], [797.49, 169.73]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[712.75, 348.69], [646.55, 352.89], [618.99, 331.64], [677.79, 328.41], [712.75, 146.17], [646.55, 145.32], [618.99, 149.65], [677.79, 150.32]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[846.07, 342.53], [786.41, 347.85], [736.51, 328.21], [790.77, 324.05], [846.07, 161.93], [786.41, 161.32], [736.51, 163.57], [790.77, 164.05]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[926.86, 333.78], [863.77, 337.74], [806.49, 317.32], [862.83, 314.3], [926.86, 138.47], [863.77, 137.38], [806.49, 142.98], [862.83, 143.81]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1009.58, 325.2], [958.24, 328.51], [890.78, 310.12], [937.19, 307.57], [1009.58, 140.14], [958.24, 139.21], [890.78, 144.34], [937.19, 145.05]]}, {\"category\": \"van\", \"corners_3d\": [[-2.4, 384.77], [266.59, 362.7], [50.03, 910.7], [-1811.46, 1468.44], [-2.4, 123.97], [266.59, 130.08], [50.03, -21.61], [-1811.46, -175.99]]}]\n```" - } - ] - }, - { - "id": 838, - "data_source": "Mono3DRefer", - "scene_name": "000472", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[726.04, 252.62], [825.9, 252.52], [1029.54, 325.28], [838.76, 325.65], [726.04, 114.46], [825.9, 114.54], [1029.54, 61.28], [838.76, 61.01]]}]\n```" - } - ] - }, - { - "id": 839, - "data_source": "Mono3DRefer", - "scene_name": "000472", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[466.67, 249.27], [541.7, 249.35], [516.59, 274.75], [416.72, 274.6], [466.67, 190.85], [541.7, 190.87], [516.59, 196.85], [416.72, 196.81]]}, {\"category\": \"car\", \"corners_3d\": [[669.22, 209.33], [706.07, 209.29], [724.39, 215.34], [681.43, 215.39], [669.22, 174.03], [706.07, 174.03], [724.39, 174.22], [681.43, 174.23]]}, {\"category\": \"car\", \"corners_3d\": [[522.97, 223.07], [566.28, 223.17], [553.6, 232.91], [501.93, 232.78], [522.97, 182.78], [566.28, 182.79], [553.6, 184.72], [501.93, 184.69]]}, {\"category\": \"car\", \"corners_3d\": [[705.73, 239.87], [769.7, 239.93], [811.82, 258.48], [730.13, 258.38], [705.73, 176.05], [769.7, 176.05], [811.82, 176.94], [730.13, 176.93]]}, {\"category\": \"car\", \"corners_3d\": [[652.75, 196.81], [678.05, 196.88], [678.27, 198.82], [650.94, 198.73], [652.75, 175.61], [678.05, 175.62], [678.27, 175.84], [650.94, 175.83]]}]\n```" - } - ] - }, - { - "id": 840, - "data_source": "Mono3DRefer", - "scene_name": "000472", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[995.03, 329.13], [960.83, 316.04], [1004.6, 315.76], [1042.72, 328.8], [995.03, 124.18], [960.83, 128.26], [1004.6, 128.34], [1042.72, 124.28]]}]\n```" - } - ] - }, - { - "id": 841, - "data_source": "Mono3DRefer", - "scene_name": "000472", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[466.67, 249.27], [541.7, 249.35], [516.59, 274.75], [416.72, 274.6], [466.67, 190.85], [541.7, 190.87], [516.59, 196.85], [416.72, 196.81]]}, {\"category\": \"van\", \"corners_3d\": [[726.04, 252.62], [825.9, 252.52], [1029.54, 325.28], [838.76, 325.65], [726.04, 114.46], [825.9, 114.54], [1029.54, 61.28], [838.76, 61.01]]}, {\"category\": \"car\", \"corners_3d\": [[669.22, 209.33], [706.07, 209.29], [724.39, 215.34], [681.43, 215.39], [669.22, 174.03], [706.07, 174.03], [724.39, 174.22], [681.43, 174.23]]}, {\"category\": \"car\", \"corners_3d\": [[522.97, 223.07], [566.28, 223.17], [553.6, 232.91], [501.93, 232.78], [522.97, 182.78], [566.28, 182.79], [553.6, 184.72], [501.93, 184.69]]}, {\"category\": \"car\", \"corners_3d\": [[705.73, 239.87], [769.7, 239.93], [811.82, 258.48], [730.13, 258.38], [705.73, 176.05], [769.7, 176.05], [811.82, 176.94], [730.13, 176.93]]}, {\"category\": \"car\", \"corners_3d\": [[652.75, 196.81], [678.05, 196.88], [678.27, 198.82], [650.94, 198.73], [652.75, 175.61], [678.05, 175.62], [678.27, 175.84], [650.94, 175.83]]}]\n```" - } - ] - }, - { - "id": 842, - "data_source": "Mono3DRefer", - "scene_name": "000472", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[726.04, 252.62], [825.9, 252.52], [1029.54, 325.28], [838.76, 325.65], [726.04, 114.46], [825.9, 114.54], [1029.54, 61.28], [838.76, 61.01]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[995.03, 329.13], [960.83, 316.04], [1004.6, 315.76], [1042.72, 328.8], [995.03, 124.18], [960.83, 128.26], [1004.6, 128.34], [1042.72, 124.28]]}]\n```" - } - ] - }, - { - "id": 843, - "data_source": "Mono3DRefer", - "scene_name": "000472", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[466.67, 249.27], [541.7, 249.35], [516.59, 274.75], [416.72, 274.6], [466.67, 190.85], [541.7, 190.87], [516.59, 196.85], [416.72, 196.81]]}, {\"category\": \"car\", \"corners_3d\": [[669.22, 209.33], [706.07, 209.29], [724.39, 215.34], [681.43, 215.39], [669.22, 174.03], [706.07, 174.03], [724.39, 174.22], [681.43, 174.23]]}, {\"category\": \"car\", \"corners_3d\": [[522.97, 223.07], [566.28, 223.17], [553.6, 232.91], [501.93, 232.78], [522.97, 182.78], [566.28, 182.79], [553.6, 184.72], [501.93, 184.69]]}, {\"category\": \"car\", \"corners_3d\": [[705.73, 239.87], [769.7, 239.93], [811.82, 258.48], [730.13, 258.38], [705.73, 176.05], [769.7, 176.05], [811.82, 176.94], [730.13, 176.93]]}, {\"category\": \"car\", \"corners_3d\": [[652.75, 196.81], [678.05, 196.88], [678.27, 198.82], [650.94, 198.73], [652.75, 175.61], [678.05, 175.62], [678.27, 175.84], [650.94, 175.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[995.03, 329.13], [960.83, 316.04], [1004.6, 315.76], [1042.72, 328.8], [995.03, 124.18], [960.83, 128.26], [1004.6, 128.34], [1042.72, 124.28]]}]\n```" - } - ] - }, - { - "id": 844, - "data_source": "Mono3DRefer", - "scene_name": "000472", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[466.67, 249.27], [541.7, 249.35], [516.59, 274.75], [416.72, 274.6], [466.67, 190.85], [541.7, 190.87], [516.59, 196.85], [416.72, 196.81]]}, {\"category\": \"van\", \"corners_3d\": [[726.04, 252.62], [825.9, 252.52], [1029.54, 325.28], [838.76, 325.65], [726.04, 114.46], [825.9, 114.54], [1029.54, 61.28], [838.76, 61.01]]}, {\"category\": \"car\", \"corners_3d\": [[669.22, 209.33], [706.07, 209.29], [724.39, 215.34], [681.43, 215.39], [669.22, 174.03], [706.07, 174.03], [724.39, 174.22], [681.43, 174.23]]}, {\"category\": \"car\", \"corners_3d\": [[522.97, 223.07], [566.28, 223.17], [553.6, 232.91], [501.93, 232.78], [522.97, 182.78], [566.28, 182.79], [553.6, 184.72], [501.93, 184.69]]}, {\"category\": \"car\", \"corners_3d\": [[705.73, 239.87], [769.7, 239.93], [811.82, 258.48], [730.13, 258.38], [705.73, 176.05], [769.7, 176.05], [811.82, 176.94], [730.13, 176.93]]}, {\"category\": \"car\", \"corners_3d\": [[652.75, 196.81], [678.05, 196.88], [678.27, 198.82], [650.94, 198.73], [652.75, 175.61], [678.05, 175.62], [678.27, 175.84], [650.94, 175.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[995.03, 329.13], [960.83, 316.04], [1004.6, 315.76], [1042.72, 328.8], [995.03, 124.18], [960.83, 128.26], [1004.6, 128.34], [1042.72, 124.28]]}]\n```" - } - ] - }, - { - "id": 845, - "data_source": "Mono3DRefer", - "scene_name": "000473", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[620.39, 204.78], [589.78, 204.55], [602.05, 201.83], [630.1, 202.03], [620.39, 177.25], [589.78, 177.22], [602.05, 176.85], [630.1, 176.87]]}]\n```" - } - ] - }, - { - "id": 846, - "data_source": "Mono3DRefer", - "scene_name": "000474", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-408.88, 341.87], [-221.96, 309.67], [126.2, 306.45], [24.25, 336.98], [-408.88, 198.23], [-221.96, 193.39], [126.2, 192.91], [24.25, 197.49]]}, {\"category\": \"car\", \"corners_3d\": [[540.57, 200.45], [536.36, 201.32], [476.74, 201.1], [482.73, 200.25], [540.57, 182.58], [536.36, 182.88], [476.74, 182.81], [482.73, 182.5]]}]\n```" - } - ] - }, - { - "id": 847, - "data_source": "Mono3DRefer", - "scene_name": "000475", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-352.9, 552.55], [-641.83, 551.13], [-23.18, 365.11], [124.58, 365.47], [-352.9, 262.83], [-641.83, 262.5], [-23.18, 218.41], [124.58, 218.5]]}, {\"category\": \"car\", \"corners_3d\": [[1039.6, 326.21], [883.55, 327.95], [779.98, 276.24], [885.17, 275.47], [1039.6, 181.73], [883.55, 181.83], [779.98, 178.83], [885.17, 178.79]]}, {\"category\": \"car\", \"corners_3d\": [[323.64, 280.03], [237.23, 278.98], [342.41, 252.99], [408.29, 253.59], [323.64, 194.27], [237.23, 194.06], [342.41, 188.87], [408.29, 188.99]]}, {\"category\": \"car\", \"corners_3d\": [[712.54, 241.6], [783.6, 241.19], [848.86, 261.76], [756.74, 262.45], [712.54, 178.46], [783.6, 178.43], [848.86, 180.1], [756.74, 180.16]]}, {\"category\": \"car\", \"corners_3d\": [[483.62, 219.07], [441.14, 219.05], [461.41, 213.7], [498.98, 213.72], [483.62, 178.33], [441.14, 178.33], [461.41, 177.69], [498.98, 177.7]]}, {\"category\": \"car\", \"corners_3d\": [[673.22, 218.37], [718.72, 218.17], [744.72, 225.57], [691.88, 225.84], [673.22, 174.86], [718.72, 174.85], [744.72, 175.18], [691.88, 175.19]]}, {\"category\": \"car\", \"corners_3d\": [[663.79, 210.7], [698.58, 210.49], [719.0, 215.06], [680.07, 215.32], [663.79, 174.5], [698.58, 174.49], [719.0, 174.69], [680.07, 174.7]]}, {\"category\": \"car\", \"corners_3d\": [[503.7, 210.73], [472.15, 210.75], [484.73, 207.07], [513.22, 207.05], [503.7, 178.57], [472.15, 178.58], [484.73, 178.02], [513.22, 178.02]]}]\n```" - } - ] - }, - { - "id": 848, - "data_source": "Mono3DRefer", - "scene_name": "000476", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[562.77, 215.86], [523.57, 215.6], [541.43, 210.98], [576.47, 211.18], [562.77, 180.6], [523.57, 180.56], [541.43, 179.72], [576.47, 179.76]]}]\n```" - } - ] - }, - { - "id": 849, - "data_source": "Mono3DRefer", - "scene_name": "000479", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[355.46, 206.88], [327.89, 206.92], [349.76, 204.05], [374.98, 204.02], [355.46, 176.8], [327.89, 176.81], [349.76, 176.47], [374.98, 176.47]]}]\n```" - } - ] - }, - { - "id": 850, - "data_source": "Mono3DRefer", - "scene_name": "000479", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[619.99, 188.8], [642.89, 188.79], [651.14, 191.58], [624.24, 191.59], [619.99, 163.96], [642.89, 163.97], [651.14, 162.41], [624.24, 162.4]]}]\n```" - } - ] - }, - { - "id": 851, - "data_source": "Mono3DRefer", - "scene_name": "000479", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[591.1, 195.44], [612.33, 195.44], [612.45, 196.67], [590.06, 196.67], [591.1, 175.06], [612.33, 175.06], [612.45, 175.18], [590.06, 175.18]]}, {\"category\": \"car\", \"corners_3d\": [[191.93, 241.51], [137.39, 241.41], [202.47, 232.23], [249.78, 232.31], [191.93, 191.57], [137.39, 191.54], [202.47, 189.04], [249.78, 189.06]]}, {\"category\": \"car\", \"corners_3d\": [[391.97, 205.14], [365.12, 205.13], [383.2, 202.81], [408.12, 202.82], [391.97, 181.88], [365.12, 181.88], [383.2, 181.23], [408.12, 181.23]]}]\n```" - } - ] - }, - { - "id": 852, - "data_source": "Mono3DRefer", - "scene_name": "000479", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[619.99, 188.8], [642.89, 188.79], [651.14, 191.58], [624.24, 191.59], [619.99, 163.96], [642.89, 163.97], [651.14, 162.41], [624.24, 162.4]]}, {\"category\": \"van\", \"corners_3d\": [[355.46, 206.88], [327.89, 206.92], [349.76, 204.05], [374.98, 204.02], [355.46, 176.8], [327.89, 176.81], [349.76, 176.47], [374.98, 176.47]]}]\n```" - } - ] - }, - { - "id": 853, - "data_source": "Mono3DRefer", - "scene_name": "000479", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[591.1, 195.44], [612.33, 195.44], [612.45, 196.67], [590.06, 196.67], [591.1, 175.06], [612.33, 175.06], [612.45, 175.18], [590.06, 175.18]]}, {\"category\": \"car\", \"corners_3d\": [[191.93, 241.51], [137.39, 241.41], [202.47, 232.23], [249.78, 232.31], [191.93, 191.57], [137.39, 191.54], [202.47, 189.04], [249.78, 189.06]]}, {\"category\": \"car\", \"corners_3d\": [[391.97, 205.14], [365.12, 205.13], [383.2, 202.81], [408.12, 202.82], [391.97, 181.88], [365.12, 181.88], [383.2, 181.23], [408.12, 181.23]]}, {\"category\": \"van\", \"corners_3d\": [[355.46, 206.88], [327.89, 206.92], [349.76, 204.05], [374.98, 204.02], [355.46, 176.8], [327.89, 176.81], [349.76, 176.47], [374.98, 176.47]]}]\n```" - } - ] - }, - { - "id": 854, - "data_source": "Mono3DRefer", - "scene_name": "000479", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[591.1, 195.44], [612.33, 195.44], [612.45, 196.67], [590.06, 196.67], [591.1, 175.06], [612.33, 175.06], [612.45, 175.18], [590.06, 175.18]]}, {\"category\": \"bus\", \"corners_3d\": [[619.99, 188.8], [642.89, 188.79], [651.14, 191.58], [624.24, 191.59], [619.99, 163.96], [642.89, 163.97], [651.14, 162.41], [624.24, 162.4]]}, {\"category\": \"car\", \"corners_3d\": [[191.93, 241.51], [137.39, 241.41], [202.47, 232.23], [249.78, 232.31], [191.93, 191.57], [137.39, 191.54], [202.47, 189.04], [249.78, 189.06]]}, {\"category\": \"car\", \"corners_3d\": [[391.97, 205.14], [365.12, 205.13], [383.2, 202.81], [408.12, 202.82], [391.97, 181.88], [365.12, 181.88], [383.2, 181.23], [408.12, 181.23]]}]\n```" - } - ] - }, - { - "id": 855, - "data_source": "Mono3DRefer", - "scene_name": "000479", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[591.1, 195.44], [612.33, 195.44], [612.45, 196.67], [590.06, 196.67], [591.1, 175.06], [612.33, 175.06], [612.45, 175.18], [590.06, 175.18]]}, {\"category\": \"bus\", \"corners_3d\": [[619.99, 188.8], [642.89, 188.79], [651.14, 191.58], [624.24, 191.59], [619.99, 163.96], [642.89, 163.97], [651.14, 162.41], [624.24, 162.4]]}, {\"category\": \"car\", \"corners_3d\": [[191.93, 241.51], [137.39, 241.41], [202.47, 232.23], [249.78, 232.31], [191.93, 191.57], [137.39, 191.54], [202.47, 189.04], [249.78, 189.06]]}, {\"category\": \"car\", \"corners_3d\": [[391.97, 205.14], [365.12, 205.13], [383.2, 202.81], [408.12, 202.82], [391.97, 181.88], [365.12, 181.88], [383.2, 181.23], [408.12, 181.23]]}, {\"category\": \"van\", \"corners_3d\": [[355.46, 206.88], [327.89, 206.92], [349.76, 204.05], [374.98, 204.02], [355.46, 176.8], [327.89, 176.81], [349.76, 176.47], [374.98, 176.47]]}]\n```" - } - ] - }, - { - "id": 856, - "data_source": "Mono3DRefer", - "scene_name": "000480", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[620.69, 190.73], [643.6, 190.72], [651.99, 193.85], [625.07, 193.86], [620.69, 165.87], [643.6, 165.88], [651.99, 164.66], [625.07, 164.65]]}]\n```" - } - ] - }, - { - "id": 857, - "data_source": "Mono3DRefer", - "scene_name": "000480", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[243.97, 236.19], [196.18, 236.11], [247.09, 228.57], [289.23, 228.63], [243.97, 192.5], [196.18, 192.48], [247.09, 190.14], [289.23, 190.16]]}, {\"category\": \"car\", \"corners_3d\": [[591.64, 198.11], [612.96, 198.11], [613.12, 199.5], [590.63, 199.5], [591.64, 177.66], [612.96, 177.66], [613.12, 177.92], [590.63, 177.92]]}]\n```" - } - ] - }, - { - "id": 858, - "data_source": "Mono3DRefer", - "scene_name": "000480", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[364.0, 208.16], [337.29, 208.2], [357.77, 205.3], [382.27, 205.27], [364.0, 179.01], [337.29, 179.01], [357.77, 178.51], [382.27, 178.5]]}]\n```" - } - ] - }, - { - "id": 859, - "data_source": "Mono3DRefer", - "scene_name": "000480", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[620.69, 190.73], [643.6, 190.72], [651.99, 193.85], [625.07, 193.86], [620.69, 165.87], [643.6, 165.88], [651.99, 164.66], [625.07, 164.65]]}, {\"category\": \"car\", \"corners_3d\": [[243.97, 236.19], [196.18, 236.11], [247.09, 228.57], [289.23, 228.63], [243.97, 192.5], [196.18, 192.48], [247.09, 190.14], [289.23, 190.16]]}, {\"category\": \"car\", \"corners_3d\": [[591.64, 198.11], [612.96, 198.11], [613.12, 199.5], [590.63, 199.5], [591.64, 177.66], [612.96, 177.66], [613.12, 177.92], [590.63, 177.92]]}]\n```" - } - ] - }, - { - "id": 860, - "data_source": "Mono3DRefer", - "scene_name": "000480", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[620.69, 190.73], [643.6, 190.72], [651.99, 193.85], [625.07, 193.86], [620.69, 165.87], [643.6, 165.88], [651.99, 164.66], [625.07, 164.65]]}, {\"category\": \"van\", \"corners_3d\": [[364.0, 208.16], [337.29, 208.2], [357.77, 205.3], [382.27, 205.27], [364.0, 179.01], [337.29, 179.01], [357.77, 178.51], [382.27, 178.5]]}]\n```" - } - ] - }, - { - "id": 861, - "data_source": "Mono3DRefer", - "scene_name": "000480", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[243.97, 236.19], [196.18, 236.11], [247.09, 228.57], [289.23, 228.63], [243.97, 192.5], [196.18, 192.48], [247.09, 190.14], [289.23, 190.16]]}, {\"category\": \"van\", \"corners_3d\": [[364.0, 208.16], [337.29, 208.2], [357.77, 205.3], [382.27, 205.27], [364.0, 179.01], [337.29, 179.01], [357.77, 178.51], [382.27, 178.5]]}, {\"category\": \"car\", \"corners_3d\": [[591.64, 198.11], [612.96, 198.11], [613.12, 199.5], [590.63, 199.5], [591.64, 177.66], [612.96, 177.66], [613.12, 177.92], [590.63, 177.92]]}]\n```" - } - ] - }, - { - "id": 862, - "data_source": "Mono3DRefer", - "scene_name": "000480", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[620.69, 190.73], [643.6, 190.72], [651.99, 193.85], [625.07, 193.86], [620.69, 165.87], [643.6, 165.88], [651.99, 164.66], [625.07, 164.65]]}, {\"category\": \"car\", \"corners_3d\": [[243.97, 236.19], [196.18, 236.11], [247.09, 228.57], [289.23, 228.63], [243.97, 192.5], [196.18, 192.48], [247.09, 190.14], [289.23, 190.16]]}, {\"category\": \"van\", \"corners_3d\": [[364.0, 208.16], [337.29, 208.2], [357.77, 205.3], [382.27, 205.27], [364.0, 179.01], [337.29, 179.01], [357.77, 178.51], [382.27, 178.5]]}, {\"category\": \"car\", \"corners_3d\": [[591.64, 198.11], [612.96, 198.11], [613.12, 199.5], [590.63, 199.5], [591.64, 177.66], [612.96, 177.66], [613.12, 177.92], [590.63, 177.92]]}]\n```" - } - ] - }, - { - "id": 863, - "data_source": "Mono3DRefer", - "scene_name": "000481", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[464.0, 206.19], [416.94, 206.22], [476.08, 195.54], [508.06, 195.53], [464.0, 150.33], [416.94, 150.32], [476.08, 157.53], [508.06, 157.53]]}]\n```" - } - ] - }, - { - "id": 864, - "data_source": "Mono3DRefer", - "scene_name": "000481", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[733.47, 205.21], [744.26, 205.23], [749.14, 206.73], [737.84, 206.71], [733.47, 168.43], [744.26, 168.43], [749.14, 168.23], [737.84, 168.23]]}]\n```" - } - ] - }, - { - "id": 865, - "data_source": "Mono3DRefer", - "scene_name": "000481", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[196.03, 236.66], [148.1, 236.62], [216.42, 227.31], [257.38, 227.34], [196.03, 192.41], [148.1, 192.4], [216.42, 189.55], [257.38, 189.56]]}, {\"category\": \"car\", \"corners_3d\": [[376.54, 212.88], [343.2, 212.88], [367.56, 209.21], [397.85, 209.21], [376.54, 183.09], [343.2, 183.09], [367.56, 182.15], [397.85, 182.15]]}]\n```" - } - ] - }, - { - "id": 866, - "data_source": "Mono3DRefer", - "scene_name": "000481", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[464.0, 206.19], [416.94, 206.22], [476.08, 195.54], [508.06, 195.53], [464.0, 150.33], [416.94, 150.32], [476.08, 157.53], [508.06, 157.53]]}, {\"category\": \"cyclist\", \"corners_3d\": [[733.47, 205.21], [744.26, 205.23], [749.14, 206.73], [737.84, 206.71], [733.47, 168.43], [744.26, 168.43], [749.14, 168.23], [737.84, 168.23]]}]\n```" - } - ] - }, - { - "id": 867, - "data_source": "Mono3DRefer", - "scene_name": "000481", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[464.0, 206.19], [416.94, 206.22], [476.08, 195.54], [508.06, 195.53], [464.0, 150.33], [416.94, 150.32], [476.08, 157.53], [508.06, 157.53]]}, {\"category\": \"car\", \"corners_3d\": [[196.03, 236.66], [148.1, 236.62], [216.42, 227.31], [257.38, 227.34], [196.03, 192.41], [148.1, 192.4], [216.42, 189.55], [257.38, 189.56]]}, {\"category\": \"car\", \"corners_3d\": [[376.54, 212.88], [343.2, 212.88], [367.56, 209.21], [397.85, 209.21], [376.54, 183.09], [343.2, 183.09], [367.56, 182.15], [397.85, 182.15]]}]\n```" - } - ] - }, - { - "id": 868, - "data_source": "Mono3DRefer", - "scene_name": "000481", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[733.47, 205.21], [744.26, 205.23], [749.14, 206.73], [737.84, 206.71], [733.47, 168.43], [744.26, 168.43], [749.14, 168.23], [737.84, 168.23]]}, {\"category\": \"car\", \"corners_3d\": [[196.03, 236.66], [148.1, 236.62], [216.42, 227.31], [257.38, 227.34], [196.03, 192.41], [148.1, 192.4], [216.42, 189.55], [257.38, 189.56]]}, {\"category\": \"car\", \"corners_3d\": [[376.54, 212.88], [343.2, 212.88], [367.56, 209.21], [397.85, 209.21], [376.54, 183.09], [343.2, 183.09], [367.56, 182.15], [397.85, 182.15]]}]\n```" - } - ] - }, - { - "id": 869, - "data_source": "Mono3DRefer", - "scene_name": "000481", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[464.0, 206.19], [416.94, 206.22], [476.08, 195.54], [508.06, 195.53], [464.0, 150.33], [416.94, 150.32], [476.08, 157.53], [508.06, 157.53]]}, {\"category\": \"cyclist\", \"corners_3d\": [[733.47, 205.21], [744.26, 205.23], [749.14, 206.73], [737.84, 206.71], [733.47, 168.43], [744.26, 168.43], [749.14, 168.23], [737.84, 168.23]]}, {\"category\": \"car\", \"corners_3d\": [[196.03, 236.66], [148.1, 236.62], [216.42, 227.31], [257.38, 227.34], [196.03, 192.41], [148.1, 192.4], [216.42, 189.55], [257.38, 189.56]]}, {\"category\": \"car\", \"corners_3d\": [[376.54, 212.88], [343.2, 212.88], [367.56, 209.21], [397.85, 209.21], [376.54, 183.09], [343.2, 183.09], [367.56, 182.15], [397.85, 182.15]]}]\n```" - } - ] - }, - { - "id": 870, - "data_source": "Mono3DRefer", - "scene_name": "000482", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[742.56, 247.17], [807.93, 247.04], [861.19, 265.67], [779.5, 265.87], [742.56, 182.81], [807.93, 182.79], [861.19, 185.29], [779.5, 185.31]]}, {\"category\": \"car\", \"corners_3d\": [[932.0, 248.05], [1000.37, 247.99], [1105.34, 267.83], [1019.01, 267.93], [932.0, 181.66], [1000.37, 181.65], [1105.34, 183.97], [1019.01, 183.99]]}]\n```" - } - ] - }, - { - "id": 871, - "data_source": "Mono3DRefer", - "scene_name": "000483", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[717.7, 187.02], [733.31, 187.13], [726.43, 187.79], [710.15, 187.67], [717.7, 171.24], [733.31, 171.23], [726.43, 171.15], [710.15, 171.17]]}, {\"category\": \"car\", \"corners_3d\": [[38.6, 255.29], [-18.95, 254.34], [133.91, 237.09], [180.57, 237.68], [38.6, 192.03], [-18.95, 191.81], [133.91, 187.8], [180.57, 187.94]]}, {\"category\": \"car\", \"corners_3d\": [[626.43, 228.47], [682.66, 228.78], [687.28, 240.56], [619.26, 240.11], [626.43, 175.56], [682.66, 175.57], [687.28, 176.14], [619.26, 176.12]]}, {\"category\": \"car\", \"corners_3d\": [[942.24, 260.57], [1055.87, 270.49], [952.89, 301.12], [821.54, 284.53], [942.24, 174.08], [1055.87, 174.21], [952.89, 174.64], [821.54, 174.41]]}, {\"category\": \"car\", \"corners_3d\": [[166.32, 319.44], [23.63, 317.35], [239.88, 268.21], [335.65, 269.12], [166.32, 200.01], [23.63, 199.62], [239.88, 190.52], [335.65, 190.69]]}, {\"category\": \"car\", \"corners_3d\": [[554.11, 209.03], [521.94, 208.73], [542.74, 205.31], [571.98, 205.56], [554.11, 171.79], [521.94, 171.8], [542.74, 171.9], [571.98, 171.89]]}, {\"category\": \"car\", \"corners_3d\": [[245.92, 249.97], [186.31, 248.93], [286.16, 234.65], [335.55, 235.33], [245.92, 186.31], [186.31, 186.13], [286.16, 183.63], [335.55, 183.75]]}, {\"category\": \"car\", \"corners_3d\": [[448.51, 217.5], [411.77, 217.0], [456.22, 211.41], [488.7, 211.79], [448.51, 183.59], [411.77, 183.47], [456.22, 182.13], [488.7, 182.22]]}, {\"category\": \"car\", \"corners_3d\": [[696.31, 194.6], [719.88, 194.97], [698.35, 196.64], [673.41, 196.22], [696.31, 171.18], [719.88, 171.15], [698.35, 171.02], [673.41, 171.05]]}]\n```" - } - ] - }, - { - "id": 872, - "data_source": "Mono3DRefer", - "scene_name": "000484", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[598.9, 200.48], [620.56, 200.48], [621.83, 202.48], [598.61, 202.49], [598.9, 180.02], [620.56, 180.02], [621.83, 180.54], [598.61, 180.54]]}, {\"category\": \"car\", \"corners_3d\": [[801.27, 333.23], [956.34, 332.56], [1318.97, 493.16], [1009.7, 495.84], [801.27, 168.95], [956.34, 168.97], [1318.97, 165.07], [1009.7, 165.0]]}, {\"category\": \"car\", \"corners_3d\": [[684.71, 240.15], [742.91, 240.16], [773.58, 255.7], [701.94, 255.7], [684.71, 186.91], [742.91, 186.91], [773.58, 190.16], [701.94, 190.16]]}, {\"category\": \"car\", \"corners_3d\": [[672.13, 225.44], [719.8, 225.38], [743.01, 235.2], [686.44, 235.3], [672.13, 180.91], [719.8, 180.9], [743.01, 182.41], [686.44, 182.42]]}, {\"category\": \"car\", \"corners_3d\": [[658.77, 211.55], [689.63, 211.52], [700.97, 216.19], [666.4, 216.23], [658.77, 180.42], [689.63, 180.41], [700.97, 181.33], [666.4, 181.33]]}, {\"category\": \"car\", \"corners_3d\": [[647.33, 209.27], [677.59, 209.24], [686.65, 213.27], [653.04, 213.31], [647.33, 178.95], [677.59, 178.94], [686.65, 179.62], [653.04, 179.62]]}]\n```" - } - ] - }, - { - "id": 873, - "data_source": "Mono3DRefer", - "scene_name": "000484", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[694.65, 204.96], [707.68, 204.96], [709.45, 205.54], [696.18, 205.54], [694.65, 170.58], [707.68, 170.58], [709.45, 170.54], [696.18, 170.54]]}]\n```" - } - ] - }, - { - "id": 874, - "data_source": "Mono3DRefer", - "scene_name": "000484", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[598.9, 200.48], [620.56, 200.48], [621.83, 202.48], [598.61, 202.49], [598.9, 180.02], [620.56, 180.02], [621.83, 180.54], [598.61, 180.54]]}, {\"category\": \"car\", \"corners_3d\": [[801.27, 333.23], [956.34, 332.56], [1318.97, 493.16], [1009.7, 495.84], [801.27, 168.95], [956.34, 168.97], [1318.97, 165.07], [1009.7, 165.0]]}, {\"category\": \"car\", \"corners_3d\": [[684.71, 240.15], [742.91, 240.16], [773.58, 255.7], [701.94, 255.7], [684.71, 186.91], [742.91, 186.91], [773.58, 190.16], [701.94, 190.16]]}, {\"category\": \"car\", \"corners_3d\": [[672.13, 225.44], [719.8, 225.38], [743.01, 235.2], [686.44, 235.3], [672.13, 180.91], [719.8, 180.9], [743.01, 182.41], [686.44, 182.42]]}, {\"category\": \"car\", \"corners_3d\": [[658.77, 211.55], [689.63, 211.52], [700.97, 216.19], [666.4, 216.23], [658.77, 180.42], [689.63, 180.41], [700.97, 181.33], [666.4, 181.33]]}, {\"category\": \"car\", \"corners_3d\": [[647.33, 209.27], [677.59, 209.24], [686.65, 213.27], [653.04, 213.31], [647.33, 178.95], [677.59, 178.94], [686.65, 179.62], [653.04, 179.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[694.65, 204.96], [707.68, 204.96], [709.45, 205.54], [696.18, 205.54], [694.65, 170.58], [707.68, 170.58], [709.45, 170.54], [696.18, 170.54]]}]\n```" - } - ] - }, - { - "id": 875, - "data_source": "Mono3DRefer", - "scene_name": "000489", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[599.43, 190.94], [623.46, 190.94], [624.59, 194.3], [596.12, 194.29], [599.43, 164.9], [623.46, 164.9], [624.59, 163.43], [596.12, 163.43]]}]\n```" - } - ] - }, - { - "id": 876, - "data_source": "Mono3DRefer", - "scene_name": "000489", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[240.59, 235.27], [191.74, 235.19], [247.2, 227.18], [289.83, 227.25], [240.59, 196.49], [191.74, 196.46], [247.2, 193.42], [289.83, 193.45]]}]\n```" - } - ] - }, - { - "id": 877, - "data_source": "Mono3DRefer", - "scene_name": "000489", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[599.43, 190.94], [623.46, 190.94], [624.59, 194.3], [596.12, 194.29], [599.43, 164.9], [623.46, 164.9], [624.59, 163.43], [596.12, 163.43]]}, {\"category\": \"car\", \"corners_3d\": [[240.59, 235.27], [191.74, 235.19], [247.2, 227.18], [289.83, 227.25], [240.59, 196.49], [191.74, 196.46], [247.2, 193.42], [289.83, 193.45]]}]\n```" - } - ] - }, - { - "id": 878, - "data_source": "Mono3DRefer", - "scene_name": "000491", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[609.34, 212.95], [646.18, 213.01], [647.53, 216.78], [607.23, 216.7], [609.34, 175.91], [646.18, 175.92], [647.53, 176.2], [607.23, 176.2]]}, {\"category\": \"car\", \"corners_3d\": [[447.57, 268.52], [360.38, 267.94], [426.88, 245.73], [493.92, 246.07], [447.57, 192.92], [360.38, 192.8], [426.88, 188.14], [493.92, 188.21]]}, {\"category\": \"car\", \"corners_3d\": [[509.67, 234.4], [453.4, 234.54], [476.51, 223.9], [523.05, 223.8], [509.67, 183.04], [453.4, 183.07], [476.51, 181.3], [523.05, 181.29]]}, {\"category\": \"car\", \"corners_3d\": [[539.26, 209.56], [503.38, 209.5], [517.52, 205.43], [549.42, 205.47], [539.26, 178.85], [503.38, 178.84], [517.52, 178.18], [549.42, 178.19]]}, {\"category\": \"car\", \"corners_3d\": [[559.97, 194.09], [537.05, 194.06], [545.13, 192.47], [566.34, 192.5], [559.97, 171.64], [537.05, 171.65], [545.13, 171.74], [566.34, 171.73]]}]\n```" - } - ] - }, - { - "id": 879, - "data_source": "Mono3DRefer", - "scene_name": "000494", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[572.02, 219.07], [610.99, 219.0], [613.99, 225.14], [569.83, 225.24], [572.02, 180.55], [610.99, 180.54], [613.99, 181.56], [569.83, 181.58]]}, {\"category\": \"car\", \"corners_3d\": [[-11.55, 324.21], [-153.66, 324.53], [103.76, 272.84], [197.15, 272.71], [-11.55, 207.12], [-153.66, 207.19], [103.76, 195.49], [197.15, 195.46]]}, {\"category\": \"car\", \"corners_3d\": [[339.03, 243.74], [278.29, 243.8], [331.1, 232.22], [381.9, 232.17], [339.03, 189.63], [278.29, 189.65], [331.1, 186.91], [381.9, 186.89]]}, {\"category\": \"car\", \"corners_3d\": [[410.13, 225.65], [372.82, 225.63], [404.27, 218.81], [436.78, 218.83], [410.13, 186.43], [372.82, 186.43], [404.27, 184.67], [436.78, 184.68]]}, {\"category\": \"car\", \"corners_3d\": [[458.19, 213.18], [428.62, 213.28], [439.63, 210.03], [466.81, 209.95], [458.19, 183.63], [428.62, 183.65], [439.63, 182.79], [466.81, 182.76]]}, {\"category\": \"car\", \"corners_3d\": [[481.52, 207.24], [456.76, 207.23], [469.45, 204.49], [492.24, 204.5], [481.52, 186.7], [456.76, 186.69], [469.45, 185.59], [492.24, 185.59]]}]\n```" - } - ] - }, - { - "id": 880, - "data_source": "Mono3DRefer", - "scene_name": "000496", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[645.75, 232.11], [676.95, 232.25], [678.47, 234.6], [645.82, 234.45], [645.75, 168.15], [676.95, 168.11], [678.47, 167.49], [645.82, 167.53]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[666.73, 231.32], [699.94, 231.46], [702.78, 234.0], [667.88, 233.85], [666.73, 164.96], [699.94, 164.91], [702.78, 164.07], [667.88, 164.12]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[464.75, 231.13], [500.07, 231.03], [495.99, 233.84], [458.66, 233.94], [464.75, 163.11], [500.07, 163.15], [495.99, 162.11], [458.66, 162.07]]}]\n```" - } - ] - }, - { - "id": 881, - "data_source": "Mono3DRefer", - "scene_name": "000497", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[983.9, 197.67], [1022.74, 197.74], [1071.94, 200.99], [1027.89, 200.91], [983.9, 152.85], [1022.74, 152.8], [1071.94, 150.18], [1027.89, 150.25]]}, {\"category\": \"van\", \"corners_3d\": [[918.27, 192.69], [956.8, 192.79], [994.29, 195.51], [950.29, 195.38], [918.27, 153.43], [956.8, 153.33], [994.29, 150.66], [950.29, 150.79]]}]\n```" - } - ] - }, - { - "id": 882, - "data_source": "Mono3DRefer", - "scene_name": "000497", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[970.63, 391.59], [1186.02, 390.33], [3522.3, 1251.01], [2498.32, 1282.87], [970.63, 188.72], [1186.02, 188.63], [3522.3, 251.05], [2498.32, 253.36]]}, {\"category\": \"car\", \"corners_3d\": [[323.95, 329.07], [168.56, 329.43], [321.05, 274.31], [421.6, 274.16], [323.95, 166.27], [168.56, 166.26], [321.05, 168.58], [421.6, 168.58]]}, {\"category\": \"car\", \"corners_3d\": [[498.6, 251.99], [433.21, 250.34], [495.17, 236.1], [549.41, 237.19], [498.6, 171.86], [433.21, 171.88], [495.17, 172.06], [549.41, 172.05]]}, {\"category\": \"car\", \"corners_3d\": [[783.01, 264.76], [880.32, 265.75], [979.38, 309.08], [835.85, 306.97], [783.01, 173.96], [880.32, 173.97], [979.38, 174.5], [835.85, 174.47]]}, {\"category\": \"car\", \"corners_3d\": [[758.6, 235.22], [823.21, 236.15], [850.5, 255.51], [765.98, 253.93], [758.6, 175.21], [823.21, 175.24], [850.5, 175.97], [765.98, 175.92]]}, {\"category\": \"car\", \"corners_3d\": [[667.4, 216.99], [624.16, 215.79], [664.91, 210.4], [703.57, 211.31], [667.4, 172.56], [624.16, 172.56], [664.91, 172.6], [703.57, 172.59]]}]\n```" - } - ] - }, - { - "id": 883, - "data_source": "Mono3DRefer", - "scene_name": "000497", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[822.99, 179.59], [852.25, 179.7], [841.13, 180.76], [807.6, 180.62], [822.99, 145.68], [852.25, 145.26], [841.13, 140.97], [807.6, 141.54]]}]\n```" - } - ] - }, - { - "id": 884, - "data_source": "Mono3DRefer", - "scene_name": "000497", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[970.63, 391.59], [1186.02, 390.33], [3522.3, 1251.01], [2498.32, 1282.87], [970.63, 188.72], [1186.02, 188.63], [3522.3, 251.05], [2498.32, 253.36]]}, {\"category\": \"car\", \"corners_3d\": [[323.95, 329.07], [168.56, 329.43], [321.05, 274.31], [421.6, 274.16], [323.95, 166.27], [168.56, 166.26], [321.05, 168.58], [421.6, 168.58]]}, {\"category\": \"car\", \"corners_3d\": [[498.6, 251.99], [433.21, 250.34], [495.17, 236.1], [549.41, 237.19], [498.6, 171.86], [433.21, 171.88], [495.17, 172.06], [549.41, 172.05]]}, {\"category\": \"car\", \"corners_3d\": [[783.01, 264.76], [880.32, 265.75], [979.38, 309.08], [835.85, 306.97], [783.01, 173.96], [880.32, 173.97], [979.38, 174.5], [835.85, 174.47]]}, {\"category\": \"car\", \"corners_3d\": [[758.6, 235.22], [823.21, 236.15], [850.5, 255.51], [765.98, 253.93], [758.6, 175.21], [823.21, 175.24], [850.5, 175.97], [765.98, 175.92]]}, {\"category\": \"car\", \"corners_3d\": [[667.4, 216.99], [624.16, 215.79], [664.91, 210.4], [703.57, 211.31], [667.4, 172.56], [624.16, 172.56], [664.91, 172.6], [703.57, 172.59]]}, {\"category\": \"van\", \"corners_3d\": [[983.9, 197.67], [1022.74, 197.74], [1071.94, 200.99], [1027.89, 200.91], [983.9, 152.85], [1022.74, 152.8], [1071.94, 150.18], [1027.89, 150.25]]}, {\"category\": \"van\", \"corners_3d\": [[918.27, 192.69], [956.8, 192.79], [994.29, 195.51], [950.29, 195.38], [918.27, 153.43], [956.8, 153.33], [994.29, 150.66], [950.29, 150.79]]}]\n```" - } - ] - }, - { - "id": 885, - "data_source": "Mono3DRefer", - "scene_name": "000497", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[983.9, 197.67], [1022.74, 197.74], [1071.94, 200.99], [1027.89, 200.91], [983.9, 152.85], [1022.74, 152.8], [1071.94, 150.18], [1027.89, 150.25]]}, {\"category\": \"van\", \"corners_3d\": [[918.27, 192.69], [956.8, 192.79], [994.29, 195.51], [950.29, 195.38], [918.27, 153.43], [956.8, 153.33], [994.29, 150.66], [950.29, 150.79]]}, {\"category\": \"truck\", \"corners_3d\": [[822.99, 179.59], [852.25, 179.7], [841.13, 180.76], [807.6, 180.62], [822.99, 145.68], [852.25, 145.26], [841.13, 140.97], [807.6, 141.54]]}]\n```" - } - ] - }, - { - "id": 886, - "data_source": "Mono3DRefer", - "scene_name": "000497", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[970.63, 391.59], [1186.02, 390.33], [3522.3, 1251.01], [2498.32, 1282.87], [970.63, 188.72], [1186.02, 188.63], [3522.3, 251.05], [2498.32, 253.36]]}, {\"category\": \"car\", \"corners_3d\": [[323.95, 329.07], [168.56, 329.43], [321.05, 274.31], [421.6, 274.16], [323.95, 166.27], [168.56, 166.26], [321.05, 168.58], [421.6, 168.58]]}, {\"category\": \"car\", \"corners_3d\": [[498.6, 251.99], [433.21, 250.34], [495.17, 236.1], [549.41, 237.19], [498.6, 171.86], [433.21, 171.88], [495.17, 172.06], [549.41, 172.05]]}, {\"category\": \"car\", \"corners_3d\": [[783.01, 264.76], [880.32, 265.75], [979.38, 309.08], [835.85, 306.97], [783.01, 173.96], [880.32, 173.97], [979.38, 174.5], [835.85, 174.47]]}, {\"category\": \"car\", \"corners_3d\": [[758.6, 235.22], [823.21, 236.15], [850.5, 255.51], [765.98, 253.93], [758.6, 175.21], [823.21, 175.24], [850.5, 175.97], [765.98, 175.92]]}, {\"category\": \"car\", \"corners_3d\": [[667.4, 216.99], [624.16, 215.79], [664.91, 210.4], [703.57, 211.31], [667.4, 172.56], [624.16, 172.56], [664.91, 172.6], [703.57, 172.59]]}, {\"category\": \"truck\", \"corners_3d\": [[822.99, 179.59], [852.25, 179.7], [841.13, 180.76], [807.6, 180.62], [822.99, 145.68], [852.25, 145.26], [841.13, 140.97], [807.6, 141.54]]}]\n```" - } - ] - }, - { - "id": 887, - "data_source": "Mono3DRefer", - "scene_name": "000497", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[970.63, 391.59], [1186.02, 390.33], [3522.3, 1251.01], [2498.32, 1282.87], [970.63, 188.72], [1186.02, 188.63], [3522.3, 251.05], [2498.32, 253.36]]}, {\"category\": \"car\", \"corners_3d\": [[323.95, 329.07], [168.56, 329.43], [321.05, 274.31], [421.6, 274.16], [323.95, 166.27], [168.56, 166.26], [321.05, 168.58], [421.6, 168.58]]}, {\"category\": \"car\", \"corners_3d\": [[498.6, 251.99], [433.21, 250.34], [495.17, 236.1], [549.41, 237.19], [498.6, 171.86], [433.21, 171.88], [495.17, 172.06], [549.41, 172.05]]}, {\"category\": \"car\", \"corners_3d\": [[783.01, 264.76], [880.32, 265.75], [979.38, 309.08], [835.85, 306.97], [783.01, 173.96], [880.32, 173.97], [979.38, 174.5], [835.85, 174.47]]}, {\"category\": \"car\", \"corners_3d\": [[758.6, 235.22], [823.21, 236.15], [850.5, 255.51], [765.98, 253.93], [758.6, 175.21], [823.21, 175.24], [850.5, 175.97], [765.98, 175.92]]}, {\"category\": \"car\", \"corners_3d\": [[667.4, 216.99], [624.16, 215.79], [664.91, 210.4], [703.57, 211.31], [667.4, 172.56], [624.16, 172.56], [664.91, 172.6], [703.57, 172.59]]}, {\"category\": \"van\", \"corners_3d\": [[983.9, 197.67], [1022.74, 197.74], [1071.94, 200.99], [1027.89, 200.91], [983.9, 152.85], [1022.74, 152.8], [1071.94, 150.18], [1027.89, 150.25]]}, {\"category\": \"van\", \"corners_3d\": [[918.27, 192.69], [956.8, 192.79], [994.29, 195.51], [950.29, 195.38], [918.27, 153.43], [956.8, 153.33], [994.29, 150.66], [950.29, 150.79]]}, {\"category\": \"truck\", \"corners_3d\": [[822.99, 179.59], [852.25, 179.7], [841.13, 180.76], [807.6, 180.62], [822.99, 145.68], [852.25, 145.26], [841.13, 140.97], [807.6, 141.54]]}]\n```" - } - ] - }, - { - "id": 888, - "data_source": "Mono3DRefer", - "scene_name": "000505", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[704.26, 223.16], [723.38, 223.14], [729.65, 225.15], [709.63, 225.18], [704.26, 157.27], [723.38, 157.28], [729.65, 156.19], [709.63, 156.18]]}]\n```" - } - ] - }, - { - "id": 889, - "data_source": "Mono3DRefer", - "scene_name": "000505", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[843.53, 209.88], [860.59, 211.97], [724.0, 211.97], [716.02, 209.88], [843.53, 169.21], [860.59, 168.4], [724.0, 168.4], [716.02, 169.21]]}, {\"category\": \"car\", \"corners_3d\": [[1505.45, 329.65], [1847.54, 385.38], [1238.39, 389.04], [1059.72, 331.58], [1505.45, 173.13], [1847.54, 170.37], [1238.39, 170.19], [1059.72, 173.03]]}]\n```" - } - ] - }, - { - "id": 890, - "data_source": "Mono3DRefer", - "scene_name": "000505", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[645.58, 240.49], [670.35, 240.4], [681.13, 247.27], [653.53, 247.38], [645.58, 162.84], [670.35, 162.87], [681.13, 160.84], [653.53, 160.81]]}, {\"category\": \"cyclist\", \"corners_3d\": [[536.67, 215.26], [519.49, 215.19], [528.48, 212.98], [544.58, 213.04], [536.67, 165.91], [519.49, 165.93], [528.48, 166.87], [544.58, 166.84]]}]\n```" - } - ] - }, - { - "id": 891, - "data_source": "Mono3DRefer", - "scene_name": "000505", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[704.26, 223.16], [723.38, 223.14], [729.65, 225.15], [709.63, 225.18], [704.26, 157.27], [723.38, 157.28], [729.65, 156.19], [709.63, 156.18]]}, {\"category\": \"car\", \"corners_3d\": [[843.53, 209.88], [860.59, 211.97], [724.0, 211.97], [716.02, 209.88], [843.53, 169.21], [860.59, 168.4], [724.0, 168.4], [716.02, 169.21]]}, {\"category\": \"car\", \"corners_3d\": [[1505.45, 329.65], [1847.54, 385.38], [1238.39, 389.04], [1059.72, 331.58], [1505.45, 173.13], [1847.54, 170.37], [1238.39, 170.19], [1059.72, 173.03]]}]\n```" - } - ] - }, - { - "id": 892, - "data_source": "Mono3DRefer", - "scene_name": "000505", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[704.26, 223.16], [723.38, 223.14], [729.65, 225.15], [709.63, 225.18], [704.26, 157.27], [723.38, 157.28], [729.65, 156.19], [709.63, 156.18]]}, {\"category\": \"cyclist\", \"corners_3d\": [[645.58, 240.49], [670.35, 240.4], [681.13, 247.27], [653.53, 247.38], [645.58, 162.84], [670.35, 162.87], [681.13, 160.84], [653.53, 160.81]]}, {\"category\": \"cyclist\", \"corners_3d\": [[536.67, 215.26], [519.49, 215.19], [528.48, 212.98], [544.58, 213.04], [536.67, 165.91], [519.49, 165.93], [528.48, 166.87], [544.58, 166.84]]}]\n```" - } - ] - }, - { - "id": 893, - "data_source": "Mono3DRefer", - "scene_name": "000505", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[843.53, 209.88], [860.59, 211.97], [724.0, 211.97], [716.02, 209.88], [843.53, 169.21], [860.59, 168.4], [724.0, 168.4], [716.02, 169.21]]}, {\"category\": \"cyclist\", \"corners_3d\": [[645.58, 240.49], [670.35, 240.4], [681.13, 247.27], [653.53, 247.38], [645.58, 162.84], [670.35, 162.87], [681.13, 160.84], [653.53, 160.81]]}, {\"category\": \"car\", \"corners_3d\": [[1505.45, 329.65], [1847.54, 385.38], [1238.39, 389.04], [1059.72, 331.58], [1505.45, 173.13], [1847.54, 170.37], [1238.39, 170.19], [1059.72, 173.03]]}, {\"category\": \"cyclist\", \"corners_3d\": [[536.67, 215.26], [519.49, 215.19], [528.48, 212.98], [544.58, 213.04], [536.67, 165.91], [519.49, 165.93], [528.48, 166.87], [544.58, 166.84]]}]\n```" - } - ] - }, - { - "id": 894, - "data_source": "Mono3DRefer", - "scene_name": "000505", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[704.26, 223.16], [723.38, 223.14], [729.65, 225.15], [709.63, 225.18], [704.26, 157.27], [723.38, 157.28], [729.65, 156.19], [709.63, 156.18]]}, {\"category\": \"car\", \"corners_3d\": [[843.53, 209.88], [860.59, 211.97], [724.0, 211.97], [716.02, 209.88], [843.53, 169.21], [860.59, 168.4], [724.0, 168.4], [716.02, 169.21]]}, {\"category\": \"cyclist\", \"corners_3d\": [[645.58, 240.49], [670.35, 240.4], [681.13, 247.27], [653.53, 247.38], [645.58, 162.84], [670.35, 162.87], [681.13, 160.84], [653.53, 160.81]]}, {\"category\": \"car\", \"corners_3d\": [[1505.45, 329.65], [1847.54, 385.38], [1238.39, 389.04], [1059.72, 331.58], [1505.45, 173.13], [1847.54, 170.37], [1238.39, 170.19], [1059.72, 173.03]]}, {\"category\": \"cyclist\", \"corners_3d\": [[536.67, 215.26], [519.49, 215.19], [528.48, 212.98], [544.58, 213.04], [536.67, 165.91], [519.49, 165.93], [528.48, 166.87], [544.58, 166.84]]}]\n```" - } - ] - }, - { - "id": 895, - "data_source": "Mono3DRefer", - "scene_name": "000507", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[170.17, 639.27], [-186.82, 617.44], [284.11, 374.25], [451.79, 378.61], [170.17, 241.59], [-186.82, 238.37], [284.11, 202.53], [451.79, 203.18]]}, {\"category\": \"car\", \"corners_3d\": [[371.44, 340.79], [252.5, 341.08], [346.37, 295.87], [433.27, 295.71], [371.44, 200.13], [252.5, 200.18], [346.37, 192.83], [433.27, 192.81]]}, {\"category\": \"car\", \"corners_3d\": [[522.49, 205.16], [500.78, 205.35], [498.28, 203.22], [518.59, 203.05], [522.49, 183.0], [500.78, 183.06], [498.28, 182.39], [518.59, 182.34]]}]\n```" - } - ] - }, - { - "id": 896, - "data_source": "Mono3DRefer", - "scene_name": "000508", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[709.11, 192.69], [732.4, 192.55], [781.95, 196.19], [754.72, 196.37], [709.11, 166.02], [732.4, 166.07], [781.95, 164.81], [754.72, 164.75]]}]\n```" - } - ] - }, - { - "id": 897, - "data_source": "Mono3DRefer", - "scene_name": "000508", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[593.7, 205.69], [625.09, 205.69], [626.33, 208.42], [592.34, 208.41], [593.7, 175.57], [625.09, 175.57], [626.33, 175.79], [592.34, 175.79]]}]\n```" - } - ] - }, - { - "id": 898, - "data_source": "Mono3DRefer", - "scene_name": "000508", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[487.11, 196.77], [458.26, 196.8], [473.01, 194.06], [498.55, 194.04], [487.11, 159.73], [458.26, 159.71], [473.01, 161.21], [498.55, 161.23]]}]\n```" - } - ] - }, - { - "id": 899, - "data_source": "Mono3DRefer", - "scene_name": "000508", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[593.7, 205.69], [625.09, 205.69], [626.33, 208.42], [592.34, 208.41], [593.7, 175.57], [625.09, 175.57], [626.33, 175.79], [592.34, 175.79]]}, {\"category\": \"bus\", \"corners_3d\": [[709.11, 192.69], [732.4, 192.55], [781.95, 196.19], [754.72, 196.37], [709.11, 166.02], [732.4, 166.07], [781.95, 164.81], [754.72, 164.75]]}]\n```" - } - ] - }, - { - "id": 900, - "data_source": "Mono3DRefer", - "scene_name": "000508", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[709.11, 192.69], [732.4, 192.55], [781.95, 196.19], [754.72, 196.37], [709.11, 166.02], [732.4, 166.07], [781.95, 164.81], [754.72, 164.75]]}, {\"category\": \"truck\", \"corners_3d\": [[487.11, 196.77], [458.26, 196.8], [473.01, 194.06], [498.55, 194.04], [487.11, 159.73], [458.26, 159.71], [473.01, 161.21], [498.55, 161.23]]}]\n```" - } - ] - }, - { - "id": 901, - "data_source": "Mono3DRefer", - "scene_name": "000508", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[593.7, 205.69], [625.09, 205.69], [626.33, 208.42], [592.34, 208.41], [593.7, 175.57], [625.09, 175.57], [626.33, 175.79], [592.34, 175.79]]}, {\"category\": \"truck\", \"corners_3d\": [[487.11, 196.77], [458.26, 196.8], [473.01, 194.06], [498.55, 194.04], [487.11, 159.73], [458.26, 159.71], [473.01, 161.21], [498.55, 161.23]]}]\n```" - } - ] - }, - { - "id": 902, - "data_source": "Mono3DRefer", - "scene_name": "000508", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[593.7, 205.69], [625.09, 205.69], [626.33, 208.42], [592.34, 208.41], [593.7, 175.57], [625.09, 175.57], [626.33, 175.79], [592.34, 175.79]]}, {\"category\": \"bus\", \"corners_3d\": [[709.11, 192.69], [732.4, 192.55], [781.95, 196.19], [754.72, 196.37], [709.11, 166.02], [732.4, 166.07], [781.95, 164.81], [754.72, 164.75]]}, {\"category\": \"truck\", \"corners_3d\": [[487.11, 196.77], [458.26, 196.8], [473.01, 194.06], [498.55, 194.04], [487.11, 159.73], [458.26, 159.71], [473.01, 161.21], [498.55, 161.23]]}]\n```" - } - ] - }, - { - "id": 903, - "data_source": "Mono3DRefer", - "scene_name": "000511", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[14.49, 228.51], [-27.04, 228.69], [44.21, 221.98], [80.55, 221.84], [14.49, 187.92], [-27.04, 187.98], [44.21, 186.16], [80.55, 186.12]]}, {\"category\": \"car\", \"corners_3d\": [[114.55, 219.55], [74.88, 219.68], [122.97, 215.16], [158.7, 215.06], [114.55, 187.2], [74.88, 187.24], [122.97, 185.85], [158.7, 185.82]]}, {\"category\": \"car\", \"corners_3d\": [[171.71, 213.25], [141.6, 213.25], [187.84, 209.25], [214.97, 209.25], [171.71, 185.92], [141.6, 185.92], [187.84, 184.63], [214.97, 184.63]]}, {\"category\": \"car\", \"corners_3d\": [[230.67, 206.96], [205.17, 207.01], [232.57, 204.51], [256.17, 204.47], [230.67, 180.31], [205.17, 180.33], [232.57, 179.78], [256.17, 179.77]]}, {\"category\": \"car\", \"corners_3d\": [[265.6, 204.73], [242.47, 204.78], [264.21, 202.68], [285.79, 202.64], [265.6, 182.92], [242.47, 182.93], [264.21, 182.27], [285.79, 182.26]]}]\n```" - } - ] - }, - { - "id": 904, - "data_source": "Mono3DRefer", - "scene_name": "000511", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[588.55, 197.47], [635.62, 197.43], [648.5, 205.36], [586.24, 205.42], [588.55, 140.08], [635.62, 140.12], [648.5, 129.56], [586.24, 129.49]]}]\n```" - } - ] - }, - { - "id": 905, - "data_source": "Mono3DRefer", - "scene_name": "000511", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[588.55, 197.47], [635.62, 197.43], [648.5, 205.36], [586.24, 205.42], [588.55, 140.08], [635.62, 140.12], [648.5, 129.56], [586.24, 129.49]]}, {\"category\": \"car\", \"corners_3d\": [[14.49, 228.51], [-27.04, 228.69], [44.21, 221.98], [80.55, 221.84], [14.49, 187.92], [-27.04, 187.98], [44.21, 186.16], [80.55, 186.12]]}, {\"category\": \"car\", \"corners_3d\": [[114.55, 219.55], [74.88, 219.68], [122.97, 215.16], [158.7, 215.06], [114.55, 187.2], [74.88, 187.24], [122.97, 185.85], [158.7, 185.82]]}, {\"category\": \"car\", \"corners_3d\": [[171.71, 213.25], [141.6, 213.25], [187.84, 209.25], [214.97, 209.25], [171.71, 185.92], [141.6, 185.92], [187.84, 184.63], [214.97, 184.63]]}, {\"category\": \"car\", \"corners_3d\": [[230.67, 206.96], [205.17, 207.01], [232.57, 204.51], [256.17, 204.47], [230.67, 180.31], [205.17, 180.33], [232.57, 179.78], [256.17, 179.77]]}, {\"category\": \"car\", \"corners_3d\": [[265.6, 204.73], [242.47, 204.78], [264.21, 202.68], [285.79, 202.64], [265.6, 182.92], [242.47, 182.93], [264.21, 182.27], [285.79, 182.26]]}]\n```" - } - ] - }, - { - "id": 906, - "data_source": "Mono3DRefer", - "scene_name": "000512", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[389.59, 217.16], [353.99, 217.16], [378.16, 212.96], [410.39, 212.96], [389.59, 185.06], [353.99, 185.06], [378.16, 183.9], [410.39, 183.9]]}]\n```" - } - ] - }, - { - "id": 907, - "data_source": "Mono3DRefer", - "scene_name": "000512", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[613.63, 198.36], [645.14, 198.47], [637.39, 205.06], [597.83, 204.9], [613.63, 164.23], [645.14, 164.19], [637.39, 161.96], [597.83, 162.01]]}]\n```" - } - ] - }, - { - "id": 908, - "data_source": "Mono3DRefer", - "scene_name": "000512", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[613.63, 198.36], [645.14, 198.47], [637.39, 205.06], [597.83, 204.9], [613.63, 164.23], [645.14, 164.19], [637.39, 161.96], [597.83, 162.01]]}, {\"category\": \"car\", \"corners_3d\": [[389.59, 217.16], [353.99, 217.16], [378.16, 212.96], [410.39, 212.96], [389.59, 185.06], [353.99, 185.06], [378.16, 183.9], [410.39, 183.9]]}]\n```" - } - ] - }, - { - "id": 909, - "data_source": "Mono3DRefer", - "scene_name": "000514", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[888.58, 221.48], [907.34, 221.59], [928.58, 226.06], [908.06, 225.92], [888.58, 159.05], [907.34, 159.02], [928.58, 157.75], [908.06, 157.79]]}]\n```" - } - ] - }, - { - "id": 910, - "data_source": "Mono3DRefer", - "scene_name": "000514", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[706.92, 200.82], [732.82, 200.98], [733.38, 203.48], [705.19, 203.29], [706.92, 175.61], [732.82, 175.63], [733.38, 175.87], [705.19, 175.86]]}]\n```" - } - ] - }, - { - "id": 911, - "data_source": "Mono3DRefer", - "scene_name": "000514", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[888.58, 221.48], [907.34, 221.59], [928.58, 226.06], [908.06, 225.92], [888.58, 159.05], [907.34, 159.02], [928.58, 157.75], [908.06, 157.79]]}, {\"category\": \"car\", \"corners_3d\": [[706.92, 200.82], [732.82, 200.98], [733.38, 203.48], [705.19, 203.29], [706.92, 175.61], [732.82, 175.63], [733.38, 175.87], [705.19, 175.86]]}]\n```" - } - ] - }, - { - "id": 912, - "data_source": "Mono3DRefer", - "scene_name": "000523", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[309.16, 227.53], [326.48, 223.93], [444.08, 223.51], [435.13, 227.04], [309.16, 184.54], [326.48, 183.77], [444.08, 183.68], [435.13, 184.44]]}, {\"category\": \"car\", \"corners_3d\": [[567.1, 194.49], [564.22, 195.01], [519.28, 194.85], [523.18, 194.35], [567.1, 180.93], [564.22, 181.12], [519.28, 181.06], [523.18, 180.87]]}]\n```" - } - ] - }, - { - "id": 913, - "data_source": "Mono3DRefer", - "scene_name": "000526", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[569.91, 192.73], [546.28, 192.71], [552.81, 192.0], [574.97, 192.02], [569.91, 173.73], [546.28, 173.75], [552.81, 174.24], [574.97, 174.22]]}]\n```" - } - ] - }, - { - "id": 914, - "data_source": "Mono3DRefer", - "scene_name": "000528", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[2704.18, 1052.58], [1845.89, 1064.25], [924.26, 403.92], [1152.08, 403.12], [2704.18, 226.64], [1845.89, 227.36], [924.26, 186.98], [1152.08, 186.93]]}, {\"category\": \"car\", \"corners_3d\": [[532.44, 234.3], [478.43, 233.7], [514.2, 224.02], [559.85, 224.45], [532.44, 178.83], [478.43, 178.77], [514.2, 177.83], [559.85, 177.87]]}, {\"category\": \"car\", \"corners_3d\": [[582.35, 221.76], [544.24, 221.29], [568.59, 215.27], [602.12, 215.63], [582.35, 186.45], [544.24, 186.31], [568.59, 184.64], [602.12, 184.74]]}, {\"category\": \"car\", \"corners_3d\": [[651.65, 206.86], [626.34, 206.64], [634.92, 204.25], [658.49, 204.44], [651.65, 184.91], [626.34, 184.83], [634.92, 183.98], [658.49, 184.05]]}]\n```" - } - ] - }, - { - "id": 915, - "data_source": "Mono3DRefer", - "scene_name": "000528", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[699.81, 222.78], [717.76, 222.93], [719.18, 226.37], [699.99, 226.2], [699.81, 176.96], [717.76, 176.98], [719.18, 177.26], [699.99, 177.24]]}]\n```" - } - ] - }, - { - "id": 916, - "data_source": "Mono3DRefer", - "scene_name": "000528", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[2704.18, 1052.58], [1845.89, 1064.25], [924.26, 403.92], [1152.08, 403.12], [2704.18, 226.64], [1845.89, 227.36], [924.26, 186.98], [1152.08, 186.93]]}, {\"category\": \"car\", \"corners_3d\": [[532.44, 234.3], [478.43, 233.7], [514.2, 224.02], [559.85, 224.45], [532.44, 178.83], [478.43, 178.77], [514.2, 177.83], [559.85, 177.87]]}, {\"category\": \"car\", \"corners_3d\": [[582.35, 221.76], [544.24, 221.29], [568.59, 215.27], [602.12, 215.63], [582.35, 186.45], [544.24, 186.31], [568.59, 184.64], [602.12, 184.74]]}, {\"category\": \"car\", \"corners_3d\": [[651.65, 206.86], [626.34, 206.64], [634.92, 204.25], [658.49, 204.44], [651.65, 184.91], [626.34, 184.83], [634.92, 183.98], [658.49, 184.05]]}, {\"category\": \"cyclist\", \"corners_3d\": [[699.81, 222.78], [717.76, 222.93], [719.18, 226.37], [699.99, 226.2], [699.81, 176.96], [717.76, 176.98], [719.18, 177.26], [699.99, 177.24]]}]\n```" - } - ] - }, - { - "id": 917, - "data_source": "Mono3DRefer", - "scene_name": "000529", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[833.71, 212.58], [847.31, 212.58], [864.72, 215.49], [850.11, 215.49], [833.71, 164.04], [847.31, 164.04], [864.72, 163.4], [850.11, 163.4]]}, {\"category\": \"cyclist\", \"corners_3d\": [[799.24, 205.75], [808.93, 205.75], [820.08, 207.6], [809.84, 207.6], [799.24, 169.24], [808.93, 169.24], [820.08, 169.04], [809.84, 169.04]]}, {\"category\": \"cyclist\", \"corners_3d\": [[779.38, 203.44], [787.92, 203.45], [796.53, 204.99], [787.56, 204.99], [779.38, 172.07], [787.92, 172.07], [796.53, 172.03], [787.56, 172.03]]}]\n```" - } - ] - }, - { - "id": 918, - "data_source": "Mono3DRefer", - "scene_name": "000529", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-19.51, 288.04], [56.73, 273.89], [257.8, 273.57], [209.89, 287.62], [-19.51, 196.24], [56.73, 193.37], [257.8, 193.31], [209.89, 196.16]]}, {\"category\": \"car\", \"corners_3d\": [[103.64, 269.08], [151.3, 259.73], [334.67, 259.26], [306.93, 268.5], [103.64, 192.89], [151.3, 190.94], [334.67, 190.84], [306.93, 192.77]]}, {\"category\": \"car\", \"corners_3d\": [[325.43, 227.26], [344.64, 223.77], [454.53, 223.94], [442.81, 227.45], [325.43, 174.51], [344.64, 174.4], [454.53, 174.41], [442.81, 174.52]]}, {\"category\": \"car\", \"corners_3d\": [[519.12, 205.45], [516.06, 206.64], [456.11, 206.67], [461.3, 205.47], [519.12, 181.74], [516.06, 182.07], [456.11, 182.07], [461.3, 181.75]]}]\n```" - } - ] - }, - { - "id": 919, - "data_source": "Mono3DRefer", - "scene_name": "000529", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-19.51, 288.04], [56.73, 273.89], [257.8, 273.57], [209.89, 287.62], [-19.51, 196.24], [56.73, 193.37], [257.8, 193.31], [209.89, 196.16]]}, {\"category\": \"car\", \"corners_3d\": [[103.64, 269.08], [151.3, 259.73], [334.67, 259.26], [306.93, 268.5], [103.64, 192.89], [151.3, 190.94], [334.67, 190.84], [306.93, 192.77]]}, {\"category\": \"car\", \"corners_3d\": [[325.43, 227.26], [344.64, 223.77], [454.53, 223.94], [442.81, 227.45], [325.43, 174.51], [344.64, 174.4], [454.53, 174.41], [442.81, 174.52]]}, {\"category\": \"car\", \"corners_3d\": [[519.12, 205.45], [516.06, 206.64], [456.11, 206.67], [461.3, 205.47], [519.12, 181.74], [516.06, 182.07], [456.11, 182.07], [461.3, 181.75]]}, {\"category\": \"cyclist\", \"corners_3d\": [[833.71, 212.58], [847.31, 212.58], [864.72, 215.49], [850.11, 215.49], [833.71, 164.04], [847.31, 164.04], [864.72, 163.4], [850.11, 163.4]]}, {\"category\": \"cyclist\", \"corners_3d\": [[799.24, 205.75], [808.93, 205.75], [820.08, 207.6], [809.84, 207.6], [799.24, 169.24], [808.93, 169.24], [820.08, 169.04], [809.84, 169.04]]}, {\"category\": \"cyclist\", \"corners_3d\": [[779.38, 203.44], [787.92, 203.45], [796.53, 204.99], [787.56, 204.99], [779.38, 172.07], [787.92, 172.07], [796.53, 172.03], [787.56, 172.03]]}]\n```" - } - ] - }, - { - "id": 920, - "data_source": "Mono3DRefer", - "scene_name": "000530", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[537.18, 203.17], [570.81, 203.02], [580.23, 206.24], [541.82, 206.43], [537.18, 164.11], [570.81, 164.22], [580.23, 161.89], [541.82, 161.75]]}]\n```" - } - ] - }, - { - "id": 921, - "data_source": "Mono3DRefer", - "scene_name": "000530", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[427.7, 250.62], [401.9, 250.05], [420.52, 246.37], [445.09, 246.89], [427.7, 162.99], [401.9, 163.14], [420.52, 164.05], [445.09, 163.92]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[409.69, 254.31], [378.98, 254.01], [395.62, 249.77], [424.61, 250.03], [409.69, 166.53], [378.98, 166.59], [395.62, 167.39], [424.61, 167.34]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[396.61, 214.34], [377.85, 214.39], [383.57, 213.34], [401.73, 213.29], [396.61, 169.35], [377.85, 169.33], [383.57, 169.68], [401.73, 169.69]]}]\n```" - } - ] - }, - { - "id": 922, - "data_source": "Mono3DRefer", - "scene_name": "000530", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[537.18, 203.17], [570.81, 203.02], [580.23, 206.24], [541.82, 206.43], [537.18, 164.11], [570.81, 164.22], [580.23, 161.89], [541.82, 161.75]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[427.7, 250.62], [401.9, 250.05], [420.52, 246.37], [445.09, 246.89], [427.7, 162.99], [401.9, 163.14], [420.52, 164.05], [445.09, 163.92]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[409.69, 254.31], [378.98, 254.01], [395.62, 249.77], [424.61, 250.03], [409.69, 166.53], [378.98, 166.59], [395.62, 167.39], [424.61, 167.34]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[396.61, 214.34], [377.85, 214.39], [383.57, 213.34], [401.73, 213.29], [396.61, 169.35], [377.85, 169.33], [383.57, 169.68], [401.73, 169.69]]}]\n```" - } - ] - }, - { - "id": 923, - "data_source": "Mono3DRefer", - "scene_name": "000531", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[771.79, 195.3], [736.17, 195.28], [721.83, 192.42], [752.9, 192.44], [771.79, 147.44], [736.17, 147.46], [721.83, 150.69], [752.9, 150.67]]}, {\"category\": \"van\", \"corners_3d\": [[610.94, 237.33], [680.35, 237.83], [684.32, 258.64], [592.86, 257.77], [610.94, 161.33], [680.35, 161.24], [684.32, 157.51], [592.86, 157.67]]}, {\"category\": \"van\", \"corners_3d\": [[1622.21, 266.2], [1855.02, 288.46], [1293.35, 284.31], [1172.3, 263.48], [1622.21, 92.52], [1855.02, 73.37], [1293.35, 76.93], [1172.3, 94.86]]}]\n```" - } - ] - }, - { - "id": 924, - "data_source": "Mono3DRefer", - "scene_name": "000531", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[712.92, 197.56], [670.03, 197.49], [665.71, 193.23], [701.16, 193.28], [712.92, 129.7], [670.03, 129.82], [665.71, 137.26], [701.16, 137.18]]}]\n```" - } - ] - }, - { - "id": 925, - "data_source": "Mono3DRefer", - "scene_name": "000531", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[939.55, 227.59], [1005.56, 227.88], [1109.49, 244.13], [1023.41, 243.63], [939.55, 166.92], [1005.56, 166.89], [1109.49, 165.13], [1023.41, 165.18]]}, {\"category\": \"car\", \"corners_3d\": [[1048.46, 227.51], [1120.04, 227.88], [1262.9, 244.99], [1168.0, 244.36], [1048.46, 175.12], [1120.04, 175.13], [1262.9, 175.84], [1168.0, 175.82]]}, {\"category\": \"car\", \"corners_3d\": [[57.8, 245.81], [111.91, 246.22], [-54.75, 268.51], [-123.96, 267.81], [57.8, 191.62], [111.91, 191.72], [-54.75, 197.45], [-123.96, 197.27]]}, {\"category\": \"car\", \"corners_3d\": [[346.73, 242.11], [409.98, 242.36], [336.87, 264.55], [253.87, 264.11], [346.73, 186.69], [409.98, 186.74], [336.87, 191.18], [253.87, 191.09]]}, {\"category\": \"car\", \"corners_3d\": [[244.55, 242.6], [302.27, 243.2], [185.67, 264.77], [111.73, 263.76], [244.55, 168.09], [302.27, 168.05], [185.67, 166.58], [111.73, 166.65]]}, {\"category\": \"car\", \"corners_3d\": [[523.96, 214.52], [570.63, 214.63], [557.94, 222.38], [502.68, 222.22], [523.96, 173.99], [570.63, 173.99], [557.94, 174.2], [502.68, 174.19]]}, {\"category\": \"car\", \"corners_3d\": [[851.83, 231.06], [917.26, 231.23], [1004.86, 249.14], [919.11, 248.86], [851.83, 174.34], [917.26, 174.35], [1004.86, 174.81], [919.11, 174.8]]}, {\"category\": \"car\", \"corners_3d\": [[676.36, 199.4], [712.35, 199.53], [716.56, 203.29], [675.49, 203.12], [676.36, 165.71], [712.35, 165.68], [716.56, 164.67], [675.49, 164.71]]}]\n```" - } - ] - }, - { - "id": 926, - "data_source": "Mono3DRefer", - "scene_name": "000531", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[771.79, 195.3], [736.17, 195.28], [721.83, 192.42], [752.9, 192.44], [771.79, 147.44], [736.17, 147.46], [721.83, 150.69], [752.9, 150.67]]}, {\"category\": \"truck\", \"corners_3d\": [[712.92, 197.56], [670.03, 197.49], [665.71, 193.23], [701.16, 193.28], [712.92, 129.7], [670.03, 129.82], [665.71, 137.26], [701.16, 137.18]]}, {\"category\": \"van\", \"corners_3d\": [[610.94, 237.33], [680.35, 237.83], [684.32, 258.64], [592.86, 257.77], [610.94, 161.33], [680.35, 161.24], [684.32, 157.51], [592.86, 157.67]]}, {\"category\": \"van\", \"corners_3d\": [[1622.21, 266.2], [1855.02, 288.46], [1293.35, 284.31], [1172.3, 263.48], [1622.21, 92.52], [1855.02, 73.37], [1293.35, 76.93], [1172.3, 94.86]]}]\n```" - } - ] - }, - { - "id": 927, - "data_source": "Mono3DRefer", - "scene_name": "000531", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[771.79, 195.3], [736.17, 195.28], [721.83, 192.42], [752.9, 192.44], [771.79, 147.44], [736.17, 147.46], [721.83, 150.69], [752.9, 150.67]]}, {\"category\": \"car\", \"corners_3d\": [[939.55, 227.59], [1005.56, 227.88], [1109.49, 244.13], [1023.41, 243.63], [939.55, 166.92], [1005.56, 166.89], [1109.49, 165.13], [1023.41, 165.18]]}, {\"category\": \"car\", \"corners_3d\": [[1048.46, 227.51], [1120.04, 227.88], [1262.9, 244.99], [1168.0, 244.36], [1048.46, 175.12], [1120.04, 175.13], [1262.9, 175.84], [1168.0, 175.82]]}, {\"category\": \"car\", \"corners_3d\": [[57.8, 245.81], [111.91, 246.22], [-54.75, 268.51], [-123.96, 267.81], [57.8, 191.62], [111.91, 191.72], [-54.75, 197.45], [-123.96, 197.27]]}, {\"category\": \"van\", \"corners_3d\": [[610.94, 237.33], [680.35, 237.83], [684.32, 258.64], [592.86, 257.77], [610.94, 161.33], [680.35, 161.24], [684.32, 157.51], [592.86, 157.67]]}, {\"category\": \"van\", \"corners_3d\": [[1622.21, 266.2], [1855.02, 288.46], [1293.35, 284.31], [1172.3, 263.48], [1622.21, 92.52], [1855.02, 73.37], [1293.35, 76.93], [1172.3, 94.86]]}, {\"category\": \"car\", \"corners_3d\": [[346.73, 242.11], [409.98, 242.36], [336.87, 264.55], [253.87, 264.11], [346.73, 186.69], [409.98, 186.74], [336.87, 191.18], [253.87, 191.09]]}, {\"category\": \"car\", \"corners_3d\": [[244.55, 242.6], [302.27, 243.2], [185.67, 264.77], [111.73, 263.76], [244.55, 168.09], [302.27, 168.05], [185.67, 166.58], [111.73, 166.65]]}, {\"category\": \"car\", \"corners_3d\": [[523.96, 214.52], [570.63, 214.63], [557.94, 222.38], [502.68, 222.22], [523.96, 173.99], [570.63, 173.99], [557.94, 174.2], [502.68, 174.19]]}, {\"category\": \"car\", \"corners_3d\": [[851.83, 231.06], [917.26, 231.23], [1004.86, 249.14], [919.11, 248.86], [851.83, 174.34], [917.26, 174.35], [1004.86, 174.81], [919.11, 174.8]]}, {\"category\": \"car\", \"corners_3d\": [[676.36, 199.4], [712.35, 199.53], [716.56, 203.29], [675.49, 203.12], [676.36, 165.71], [712.35, 165.68], [716.56, 164.67], [675.49, 164.71]]}]\n```" - } - ] - }, - { - "id": 928, - "data_source": "Mono3DRefer", - "scene_name": "000531", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[939.55, 227.59], [1005.56, 227.88], [1109.49, 244.13], [1023.41, 243.63], [939.55, 166.92], [1005.56, 166.89], [1109.49, 165.13], [1023.41, 165.18]]}, {\"category\": \"car\", \"corners_3d\": [[1048.46, 227.51], [1120.04, 227.88], [1262.9, 244.99], [1168.0, 244.36], [1048.46, 175.12], [1120.04, 175.13], [1262.9, 175.84], [1168.0, 175.82]]}, {\"category\": \"truck\", \"corners_3d\": [[712.92, 197.56], [670.03, 197.49], [665.71, 193.23], [701.16, 193.28], [712.92, 129.7], [670.03, 129.82], [665.71, 137.26], [701.16, 137.18]]}, {\"category\": \"car\", \"corners_3d\": [[57.8, 245.81], [111.91, 246.22], [-54.75, 268.51], [-123.96, 267.81], [57.8, 191.62], [111.91, 191.72], [-54.75, 197.45], [-123.96, 197.27]]}, {\"category\": \"car\", \"corners_3d\": [[346.73, 242.11], [409.98, 242.36], [336.87, 264.55], [253.87, 264.11], [346.73, 186.69], [409.98, 186.74], [336.87, 191.18], [253.87, 191.09]]}, {\"category\": \"car\", \"corners_3d\": [[244.55, 242.6], [302.27, 243.2], [185.67, 264.77], [111.73, 263.76], [244.55, 168.09], [302.27, 168.05], [185.67, 166.58], [111.73, 166.65]]}, {\"category\": \"car\", \"corners_3d\": [[523.96, 214.52], [570.63, 214.63], [557.94, 222.38], [502.68, 222.22], [523.96, 173.99], [570.63, 173.99], [557.94, 174.2], [502.68, 174.19]]}, {\"category\": \"car\", \"corners_3d\": [[851.83, 231.06], [917.26, 231.23], [1004.86, 249.14], [919.11, 248.86], [851.83, 174.34], [917.26, 174.35], [1004.86, 174.81], [919.11, 174.8]]}, {\"category\": \"car\", \"corners_3d\": [[676.36, 199.4], [712.35, 199.53], [716.56, 203.29], [675.49, 203.12], [676.36, 165.71], [712.35, 165.68], [716.56, 164.67], [675.49, 164.71]]}]\n```" - } - ] - }, - { - "id": 929, - "data_source": "Mono3DRefer", - "scene_name": "000531", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[771.79, 195.3], [736.17, 195.28], [721.83, 192.42], [752.9, 192.44], [771.79, 147.44], [736.17, 147.46], [721.83, 150.69], [752.9, 150.67]]}, {\"category\": \"car\", \"corners_3d\": [[939.55, 227.59], [1005.56, 227.88], [1109.49, 244.13], [1023.41, 243.63], [939.55, 166.92], [1005.56, 166.89], [1109.49, 165.13], [1023.41, 165.18]]}, {\"category\": \"car\", \"corners_3d\": [[1048.46, 227.51], [1120.04, 227.88], [1262.9, 244.99], [1168.0, 244.36], [1048.46, 175.12], [1120.04, 175.13], [1262.9, 175.84], [1168.0, 175.82]]}, {\"category\": \"truck\", \"corners_3d\": [[712.92, 197.56], [670.03, 197.49], [665.71, 193.23], [701.16, 193.28], [712.92, 129.7], [670.03, 129.82], [665.71, 137.26], [701.16, 137.18]]}, {\"category\": \"car\", \"corners_3d\": [[57.8, 245.81], [111.91, 246.22], [-54.75, 268.51], [-123.96, 267.81], [57.8, 191.62], [111.91, 191.72], [-54.75, 197.45], [-123.96, 197.27]]}, {\"category\": \"van\", \"corners_3d\": [[610.94, 237.33], [680.35, 237.83], [684.32, 258.64], [592.86, 257.77], [610.94, 161.33], [680.35, 161.24], [684.32, 157.51], [592.86, 157.67]]}, {\"category\": \"van\", \"corners_3d\": [[1622.21, 266.2], [1855.02, 288.46], [1293.35, 284.31], [1172.3, 263.48], [1622.21, 92.52], [1855.02, 73.37], [1293.35, 76.93], [1172.3, 94.86]]}, {\"category\": \"car\", \"corners_3d\": [[346.73, 242.11], [409.98, 242.36], [336.87, 264.55], [253.87, 264.11], [346.73, 186.69], [409.98, 186.74], [336.87, 191.18], [253.87, 191.09]]}, {\"category\": \"car\", \"corners_3d\": [[244.55, 242.6], [302.27, 243.2], [185.67, 264.77], [111.73, 263.76], [244.55, 168.09], [302.27, 168.05], [185.67, 166.58], [111.73, 166.65]]}, {\"category\": \"car\", \"corners_3d\": [[523.96, 214.52], [570.63, 214.63], [557.94, 222.38], [502.68, 222.22], [523.96, 173.99], [570.63, 173.99], [557.94, 174.2], [502.68, 174.19]]}, {\"category\": \"car\", \"corners_3d\": [[851.83, 231.06], [917.26, 231.23], [1004.86, 249.14], [919.11, 248.86], [851.83, 174.34], [917.26, 174.35], [1004.86, 174.81], [919.11, 174.8]]}, {\"category\": \"car\", \"corners_3d\": [[676.36, 199.4], [712.35, 199.53], [716.56, 203.29], [675.49, 203.12], [676.36, 165.71], [712.35, 165.68], [716.56, 164.67], [675.49, 164.71]]}]\n```" - } - ] - }, - { - "id": 930, - "data_source": "Mono3DRefer", - "scene_name": "000532", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[774.07, 296.35], [766.5, 290.11], [853.03, 290.66], [865.59, 296.97], [774.07, 157.01], [766.5, 158.27], [853.03, 158.16], [865.59, 156.88]]}]\n```" - } - ] - }, - { - "id": 931, - "data_source": "Mono3DRefer", - "scene_name": "000537", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[380.15, 215.29], [352.06, 215.29], [373.8, 211.7], [399.51, 211.7], [380.15, 190.4], [352.06, 190.4], [373.8, 188.91], [399.51, 188.91]]}, {\"category\": \"car\", \"corners_3d\": [[421.44, 205.67], [400.86, 205.66], [416.15, 203.34], [435.27, 203.34], [421.44, 183.68], [400.86, 183.68], [416.15, 182.91], [435.27, 182.91]]}]\n```" - } - ] - }, - { - "id": 932, - "data_source": "Mono3DRefer", - "scene_name": "000537", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[585.22, 204.22], [629.68, 204.26], [631.22, 213.68], [573.45, 213.61], [585.22, 150.09], [629.68, 150.06], [631.22, 143.22], [573.45, 143.27]]}]\n```" - } - ] - }, - { - "id": 933, - "data_source": "Mono3DRefer", - "scene_name": "000537", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[585.22, 204.22], [629.68, 204.26], [631.22, 213.68], [573.45, 213.61], [585.22, 150.09], [629.68, 150.06], [631.22, 143.22], [573.45, 143.27]]}, {\"category\": \"car\", \"corners_3d\": [[380.15, 215.29], [352.06, 215.29], [373.8, 211.7], [399.51, 211.7], [380.15, 190.4], [352.06, 190.4], [373.8, 188.91], [399.51, 188.91]]}, {\"category\": \"car\", \"corners_3d\": [[421.44, 205.67], [400.86, 205.66], [416.15, 203.34], [435.27, 203.34], [421.44, 183.68], [400.86, 183.68], [416.15, 182.91], [435.27, 182.91]]}]\n```" - } - ] - }, - { - "id": 934, - "data_source": "Mono3DRefer", - "scene_name": "000539", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[621.84, 197.5], [667.89, 197.55], [679.2, 205.29], [618.73, 205.2], [621.84, 141.52], [667.89, 141.46], [679.2, 131.62], [618.73, 131.72]]}]\n```" - } - ] - }, - { - "id": 935, - "data_source": "Mono3DRefer", - "scene_name": "000539", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-63.75, 244.85], [-126.52, 244.98], [-8.28, 233.15], [44.04, 233.06], [-63.75, 196.28], [-126.52, 196.32], [-8.28, 192.47], [44.04, 192.44]]}, {\"category\": \"car\", \"corners_3d\": [[101.51, 226.25], [58.47, 226.34], [127.72, 219.34], [165.02, 219.27], [101.51, 189.22], [58.47, 189.25], [127.72, 187.1], [165.02, 187.08]]}, {\"category\": \"car\", \"corners_3d\": [[184.41, 217.7], [153.84, 217.74], [195.48, 213.5], [223.14, 213.47], [184.41, 186.68], [153.84, 186.69], [195.48, 185.38], [223.14, 185.38]]}, {\"category\": \"car\", \"corners_3d\": [[244.11, 210.4], [214.1, 210.37], [245.99, 207.45], [273.68, 207.47], [244.11, 184.76], [214.1, 184.75], [245.99, 183.83], [273.68, 183.83]]}, {\"category\": \"car\", \"corners_3d\": [[278.44, 207.32], [254.74, 207.24], [287.92, 204.42], [309.75, 204.49], [278.44, 184.99], [254.74, 184.96], [287.92, 183.97], [309.75, 183.99]]}, {\"category\": \"car\", \"corners_3d\": [[320.24, 203.03], [299.68, 203.0], [320.61, 201.13], [339.92, 201.16], [320.24, 180.68], [299.68, 180.67], [320.61, 180.19], [339.92, 180.19]]}, {\"category\": \"car\", \"corners_3d\": [[345.11, 199.87], [325.88, 199.87], [341.89, 198.34], [360.04, 198.34], [345.11, 181.29], [325.88, 181.29], [341.89, 180.82], [360.04, 180.82]]}]\n```" - } - ] - }, - { - "id": 936, - "data_source": "Mono3DRefer", - "scene_name": "000539", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[621.84, 197.5], [667.89, 197.55], [679.2, 205.29], [618.73, 205.2], [621.84, 141.52], [667.89, 141.46], [679.2, 131.62], [618.73, 131.72]]}, {\"category\": \"car\", \"corners_3d\": [[-63.75, 244.85], [-126.52, 244.98], [-8.28, 233.15], [44.04, 233.06], [-63.75, 196.28], [-126.52, 196.32], [-8.28, 192.47], [44.04, 192.44]]}, {\"category\": \"car\", \"corners_3d\": [[101.51, 226.25], [58.47, 226.34], [127.72, 219.34], [165.02, 219.27], [101.51, 189.22], [58.47, 189.25], [127.72, 187.1], [165.02, 187.08]]}, {\"category\": \"car\", \"corners_3d\": [[184.41, 217.7], [153.84, 217.74], [195.48, 213.5], [223.14, 213.47], [184.41, 186.68], [153.84, 186.69], [195.48, 185.38], [223.14, 185.38]]}, {\"category\": \"car\", \"corners_3d\": [[244.11, 210.4], [214.1, 210.37], [245.99, 207.45], [273.68, 207.47], [244.11, 184.76], [214.1, 184.75], [245.99, 183.83], [273.68, 183.83]]}, {\"category\": \"car\", \"corners_3d\": [[278.44, 207.32], [254.74, 207.24], [287.92, 204.42], [309.75, 204.49], [278.44, 184.99], [254.74, 184.96], [287.92, 183.97], [309.75, 183.99]]}, {\"category\": \"car\", \"corners_3d\": [[320.24, 203.03], [299.68, 203.0], [320.61, 201.13], [339.92, 201.16], [320.24, 180.68], [299.68, 180.67], [320.61, 180.19], [339.92, 180.19]]}, {\"category\": \"car\", \"corners_3d\": [[345.11, 199.87], [325.88, 199.87], [341.89, 198.34], [360.04, 198.34], [345.11, 181.29], [325.88, 181.29], [341.89, 180.82], [360.04, 180.82]]}]\n```" - } - ] - }, - { - "id": 937, - "data_source": "Mono3DRefer", - "scene_name": "000541", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[690.7, 191.98], [709.32, 192.22], [690.5, 193.59], [670.83, 193.31], [690.7, 166.44], [709.32, 166.36], [690.5, 165.9], [670.83, 165.99]]}]\n```" - } - ] - }, - { - "id": 938, - "data_source": "Mono3DRefer", - "scene_name": "000541", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[235.25, 338.21], [404.34, 336.72], [188.7, 536.47], [-195.1, 543.89], [235.25, 208.05], [404.34, 207.73], [188.7, 250.25], [-195.1, 251.83]]}, {\"category\": \"car\", \"corners_3d\": [[661.35, 232.16], [713.29, 231.99], [746.48, 246.85], [681.55, 247.12], [661.35, 182.23], [713.29, 182.2], [746.48, 184.55], [681.55, 184.59]]}, {\"category\": \"car\", \"corners_3d\": [[431.83, 252.51], [499.9, 252.22], [472.55, 278.9], [381.34, 279.43], [431.83, 189.66], [499.9, 189.6], [472.55, 195.23], [381.34, 195.34]]}, {\"category\": \"car\", \"corners_3d\": [[749.43, 307.22], [877.22, 306.76], [1093.89, 409.82], [868.47, 411.26], [749.43, 178.49], [877.22, 178.47], [1093.89, 182.8], [868.47, 182.86]]}, {\"category\": \"car\", \"corners_3d\": [[632.78, 209.22], [663.94, 209.33], [664.26, 213.03], [629.96, 212.9], [632.78, 183.08], [663.94, 183.11], [664.26, 184.15], [629.96, 184.11]]}]\n```" - } - ] - }, - { - "id": 939, - "data_source": "Mono3DRefer", - "scene_name": "000541", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[235.25, 338.21], [404.34, 336.72], [188.7, 536.47], [-195.1, 543.89], [235.25, 208.05], [404.34, 207.73], [188.7, 250.25], [-195.1, 251.83]]}, {\"category\": \"car\", \"corners_3d\": [[661.35, 232.16], [713.29, 231.99], [746.48, 246.85], [681.55, 247.12], [661.35, 182.23], [713.29, 182.2], [746.48, 184.55], [681.55, 184.59]]}, {\"category\": \"car\", \"corners_3d\": [[431.83, 252.51], [499.9, 252.22], [472.55, 278.9], [381.34, 279.43], [431.83, 189.66], [499.9, 189.6], [472.55, 195.23], [381.34, 195.34]]}, {\"category\": \"car\", \"corners_3d\": [[749.43, 307.22], [877.22, 306.76], [1093.89, 409.82], [868.47, 411.26], [749.43, 178.49], [877.22, 178.47], [1093.89, 182.8], [868.47, 182.86]]}, {\"category\": \"car\", \"corners_3d\": [[632.78, 209.22], [663.94, 209.33], [664.26, 213.03], [629.96, 212.9], [632.78, 183.08], [663.94, 183.11], [664.26, 184.15], [629.96, 184.11]]}, {\"category\": \"van\", \"corners_3d\": [[690.7, 191.98], [709.32, 192.22], [690.5, 193.59], [670.83, 193.31], [690.7, 166.44], [709.32, 166.36], [690.5, 165.9], [670.83, 165.99]]}]\n```" - } - ] - }, - { - "id": 940, - "data_source": "Mono3DRefer", - "scene_name": "000543", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[528.81, 195.29], [556.7, 194.81], [596.15, 197.16], [566.14, 197.75], [528.81, 153.76], [556.7, 154.17], [596.15, 152.17], [566.14, 151.68]]}]\n```" - } - ] - }, - { - "id": 941, - "data_source": "Mono3DRefer", - "scene_name": "000543", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[432.09, 256.21], [352.21, 257.03], [396.63, 237.52], [457.8, 237.03], [432.09, 192.73], [352.21, 192.93], [396.63, 188.28], [457.8, 188.16]]}, {\"category\": \"car\", \"corners_3d\": [[-268.04, 338.79], [-190.93, 315.69], [188.36, 293.2], [175.4, 309.19], [-268.04, 200.94], [-190.93, 197.03], [188.36, 193.22], [175.4, 195.93]]}]\n```" - } - ] - }, - { - "id": 942, - "data_source": "Mono3DRefer", - "scene_name": "000543", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[432.09, 256.21], [352.21, 257.03], [396.63, 237.52], [457.8, 237.03], [432.09, 192.73], [352.21, 192.93], [396.63, 188.28], [457.8, 188.16]]}, {\"category\": \"van\", \"corners_3d\": [[528.81, 195.29], [556.7, 194.81], [596.15, 197.16], [566.14, 197.75], [528.81, 153.76], [556.7, 154.17], [596.15, 152.17], [566.14, 151.68]]}, {\"category\": \"car\", \"corners_3d\": [[-268.04, 338.79], [-190.93, 315.69], [188.36, 293.2], [175.4, 309.19], [-268.04, 200.94], [-190.93, 197.03], [188.36, 193.22], [175.4, 195.93]]}]\n```" - } - ] - }, - { - "id": 943, - "data_source": "Mono3DRefer", - "scene_name": "000545", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1015.92, 244.22], [1011.76, 237.97], [1235.85, 245.85], [1264.96, 253.79], [1015.92, 173.3], [1011.76, 173.26], [1235.85, 173.31], [1264.96, 173.36]]}, {\"category\": \"car\", \"corners_3d\": [[752.14, 224.94], [778.27, 226.98], [710.24, 231.46], [684.8, 229.08], [752.14, 195.45], [778.27, 196.33], [710.24, 198.28], [684.8, 197.25]]}]\n```" - } - ] - }, - { - "id": 944, - "data_source": "Mono3DRefer", - "scene_name": "000545", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[593.12, 221.44], [632.45, 221.21], [645.72, 228.55], [600.47, 228.86], [593.12, 176.81], [632.45, 176.79], [645.72, 177.39], [600.47, 177.41]]}, {\"category\": \"van\", \"corners_3d\": [[508.84, 199.84], [530.6, 199.69], [534.37, 201.65], [511.04, 201.81], [508.84, 173.08], [530.6, 173.07], [534.37, 173.09], [511.04, 173.09]]}]\n```" - } - ] - }, - { - "id": 945, - "data_source": "Mono3DRefer", - "scene_name": "000545", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1015.92, 244.22], [1011.76, 237.97], [1235.85, 245.85], [1264.96, 253.79], [1015.92, 173.3], [1011.76, 173.26], [1235.85, 173.31], [1264.96, 173.36]]}, {\"category\": \"van\", \"corners_3d\": [[593.12, 221.44], [632.45, 221.21], [645.72, 228.55], [600.47, 228.86], [593.12, 176.81], [632.45, 176.79], [645.72, 177.39], [600.47, 177.41]]}, {\"category\": \"car\", \"corners_3d\": [[752.14, 224.94], [778.27, 226.98], [710.24, 231.46], [684.8, 229.08], [752.14, 195.45], [778.27, 196.33], [710.24, 198.28], [684.8, 197.25]]}, {\"category\": \"van\", \"corners_3d\": [[508.84, 199.84], [530.6, 199.69], [534.37, 201.65], [511.04, 201.81], [508.84, 173.08], [530.6, 173.07], [534.37, 173.09], [511.04, 173.09]]}]\n```" - } - ] - }, - { - "id": 946, - "data_source": "Mono3DRefer", - "scene_name": "000546", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1089.67, 274.72], [1214.26, 284.1], [1217.05, 321.85], [1054.53, 305.49], [1089.67, 193.21], [1214.26, 195.09], [1217.05, 202.63], [1054.53, 199.36]]}]\n```" - } - ] - }, - { - "id": 947, - "data_source": "Mono3DRefer", - "scene_name": "000547", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[410.77, 207.54], [389.22, 207.55], [402.95, 205.31], [423.1, 205.3], [410.77, 189.77], [389.22, 189.78], [402.95, 188.69], [423.1, 188.68]]}]\n```" - } - ] - }, - { - "id": 948, - "data_source": "Mono3DRefer", - "scene_name": "000547", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[588.18, 200.39], [627.77, 200.39], [632.32, 207.5], [582.51, 207.49], [588.18, 152.16], [627.77, 152.16], [632.32, 146.82], [582.51, 146.82]]}]\n```" - } - ] - }, - { - "id": 949, - "data_source": "Mono3DRefer", - "scene_name": "000547", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[588.18, 200.39], [627.77, 200.39], [632.32, 207.5], [582.51, 207.49], [588.18, 152.16], [627.77, 152.16], [632.32, 146.82], [582.51, 146.82]]}, {\"category\": \"car\", \"corners_3d\": [[410.77, 207.54], [389.22, 207.55], [402.95, 205.31], [423.1, 205.3], [410.77, 189.77], [389.22, 189.78], [402.95, 188.69], [423.1, 188.68]]}]\n```" - } - ] - }, - { - "id": 950, - "data_source": "Mono3DRefer", - "scene_name": "000550", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[440.82, 202.22], [421.26, 202.22], [432.15, 200.45], [450.52, 200.44], [440.82, 184.93], [421.26, 184.93], [432.15, 184.2], [450.52, 184.2]]}]\n```" - } - ] - }, - { - "id": 951, - "data_source": "Mono3DRefer", - "scene_name": "000550", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[574.47, 202.24], [620.13, 202.24], [623.23, 211.35], [563.41, 211.35], [574.47, 146.62], [620.13, 146.62], [623.23, 138.48], [563.41, 138.49]]}]\n```" - } - ] - }, - { - "id": 952, - "data_source": "Mono3DRefer", - "scene_name": "000550", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[440.82, 202.22], [421.26, 202.22], [432.15, 200.45], [450.52, 200.44], [440.82, 184.93], [421.26, 184.93], [432.15, 184.2], [450.52, 184.2]]}, {\"category\": \"truck\", \"corners_3d\": [[574.47, 202.24], [620.13, 202.24], [623.23, 211.35], [563.41, 211.35], [574.47, 146.62], [620.13, 146.62], [623.23, 138.48], [563.41, 138.49]]}]\n```" - } - ] - }, - { - "id": 953, - "data_source": "Mono3DRefer", - "scene_name": "000552", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[721.51, 247.89], [802.15, 247.64], [862.15, 268.64], [759.03, 269.05], [721.51, 168.99], [802.15, 169.0], [862.15, 167.92], [759.03, 167.9]]}]\n```" - } - ] - }, - { - "id": 954, - "data_source": "Mono3DRefer", - "scene_name": "000555", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[542.72, 193.93], [562.56, 193.92], [560.87, 194.99], [540.02, 195.0], [542.72, 174.91], [562.56, 174.91], [560.87, 175.02], [540.02, 175.02]]}, {\"category\": \"car\", \"corners_3d\": [[144.54, 240.76], [99.44, 241.0], [155.6, 232.79], [195.11, 232.6], [144.54, 200.51], [99.44, 200.61], [155.6, 197.27], [195.11, 197.19]]}, {\"category\": \"car\", \"corners_3d\": [[241.85, 214.14], [218.51, 214.03], [251.89, 210.96], [273.57, 211.06], [241.85, 193.01], [218.51, 192.95], [251.89, 191.46], [273.57, 191.5]]}]\n```" - } - ] - }, - { - "id": 955, - "data_source": "Mono3DRefer", - "scene_name": "000555", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[574.31, 189.05], [597.18, 189.01], [603.73, 191.82], [576.88, 191.87], [574.31, 164.23], [597.18, 164.25], [603.73, 162.75], [576.88, 162.73]]}]\n```" - } - ] - }, - { - "id": 956, - "data_source": "Mono3DRefer", - "scene_name": "000555", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[542.72, 193.93], [562.56, 193.92], [560.87, 194.99], [540.02, 195.0], [542.72, 174.91], [562.56, 174.91], [560.87, 175.02], [540.02, 175.02]]}, {\"category\": \"bus\", \"corners_3d\": [[574.31, 189.05], [597.18, 189.01], [603.73, 191.82], [576.88, 191.87], [574.31, 164.23], [597.18, 164.25], [603.73, 162.75], [576.88, 162.73]]}, {\"category\": \"car\", \"corners_3d\": [[144.54, 240.76], [99.44, 241.0], [155.6, 232.79], [195.11, 232.6], [144.54, 200.51], [99.44, 200.61], [155.6, 197.27], [195.11, 197.19]]}, {\"category\": \"car\", \"corners_3d\": [[241.85, 214.14], [218.51, 214.03], [251.89, 210.96], [273.57, 211.06], [241.85, 193.01], [218.51, 192.95], [251.89, 191.46], [273.57, 191.5]]}]\n```" - } - ] - }, - { - "id": 957, - "data_source": "Mono3DRefer", - "scene_name": "000556", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[556.41, 210.0], [522.89, 209.91], [534.84, 206.29], [565.09, 206.36], [556.41, 184.16], [522.89, 184.14], [534.84, 183.03], [565.09, 183.05]]}, {\"category\": \"car\", \"corners_3d\": [[504.74, 212.5], [499.45, 214.16], [420.67, 214.02], [429.09, 212.38], [504.74, 186.07], [499.45, 186.62], [420.67, 186.57], [429.09, 186.02]]}]\n```" - } - ] - }, - { - "id": 958, - "data_source": "Mono3DRefer", - "scene_name": "000557", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[141.72, 406.91], [363.08, 406.96], [-487.54, 1213.14], [-1498.64, 1211.99], [141.72, 202.19], [363.08, 202.2], [-487.54, 276.17], [-1498.64, 276.07]]}, {\"category\": \"car\", \"corners_3d\": [[427.34, 254.11], [505.05, 254.12], [475.41, 276.56], [373.68, 276.55], [427.34, 182.95], [505.05, 182.95], [475.41, 183.4], [373.68, 183.4]]}, {\"category\": \"car\", \"corners_3d\": [[591.79, 214.58], [637.9, 214.61], [643.92, 221.31], [588.48, 221.27], [591.79, 172.64], [637.9, 172.63], [643.92, 170.84], [588.48, 170.85]]}, {\"category\": \"car\", \"corners_3d\": [[531.75, 212.3], [571.87, 212.33], [565.47, 216.88], [519.44, 216.83], [531.75, 172.88], [571.87, 172.87], [565.47, 171.6], [519.44, 171.61]]}, {\"category\": \"car\", \"corners_3d\": [[512.79, 201.19], [483.43, 201.16], [497.76, 199.22], [524.23, 199.25], [512.79, 174.28], [483.43, 174.29], [497.76, 175.01], [524.23, 175.0]]}]\n```" - } - ] - }, - { - "id": 959, - "data_source": "Mono3DRefer", - "scene_name": "000557", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[332.7, 252.75], [250.38, 252.68], [332.44, 236.3], [395.87, 236.35], [332.7, 163.61], [250.38, 163.63], [332.44, 167.74], [395.87, 167.73]]}]\n```" - } - ] - }, - { - "id": 960, - "data_source": "Mono3DRefer", - "scene_name": "000557", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[141.72, 406.91], [363.08, 406.96], [-487.54, 1213.14], [-1498.64, 1211.99], [141.72, 202.19], [363.08, 202.2], [-487.54, 276.17], [-1498.64, 276.07]]}, {\"category\": \"car\", \"corners_3d\": [[427.34, 254.11], [505.05, 254.12], [475.41, 276.56], [373.68, 276.55], [427.34, 182.95], [505.05, 182.95], [475.41, 183.4], [373.68, 183.4]]}, {\"category\": \"car\", \"corners_3d\": [[591.79, 214.58], [637.9, 214.61], [643.92, 221.31], [588.48, 221.27], [591.79, 172.64], [637.9, 172.63], [643.92, 170.84], [588.48, 170.85]]}, {\"category\": \"car\", \"corners_3d\": [[531.75, 212.3], [571.87, 212.33], [565.47, 216.88], [519.44, 216.83], [531.75, 172.88], [571.87, 172.87], [565.47, 171.6], [519.44, 171.61]]}, {\"category\": \"van\", \"corners_3d\": [[332.7, 252.75], [250.38, 252.68], [332.44, 236.3], [395.87, 236.35], [332.7, 163.61], [250.38, 163.63], [332.44, 167.74], [395.87, 167.73]]}, {\"category\": \"car\", \"corners_3d\": [[512.79, 201.19], [483.43, 201.16], [497.76, 199.22], [524.23, 199.25], [512.79, 174.28], [483.43, 174.29], [497.76, 175.01], [524.23, 175.0]]}]\n```" - } - ] - }, - { - "id": 961, - "data_source": "Mono3DRefer", - "scene_name": "000559", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[310.77, 425.55], [190.07, 425.59], [297.79, 361.42], [386.75, 361.4], [310.77, 172.35], [190.07, 172.35], [297.79, 174.76], [386.75, 174.76]]}, {\"category\": \"cyclist\", \"corners_3d\": [[419.17, 295.5], [374.41, 295.22], [415.5, 276.84], [453.09, 277.04], [419.17, 164.58], [374.41, 164.62], [415.5, 167.35], [453.09, 167.32]]}]\n```" - } - ] - }, - { - "id": 962, - "data_source": "Mono3DRefer", - "scene_name": "000559", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[610.55, 319.11], [684.15, 318.98], [694.75, 335.09], [612.52, 335.25], [610.55, 167.35], [684.15, 167.36], [694.75, 165.7], [612.52, 165.68]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[664.24, 314.03], [740.78, 313.61], [761.29, 330.4], [675.06, 330.93], [664.24, 159.14], [740.78, 159.21], [761.29, 156.37], [675.06, 156.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[205.59, 321.77], [290.15, 320.36], [255.59, 339.86], [158.61, 341.7], [205.59, 163.61], [290.15, 163.79], [255.59, 161.3], [158.61, 161.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[727.84, 231.26], [757.52, 230.78], [774.57, 233.43], [743.46, 233.96], [727.84, 164.73], [757.52, 164.89], [774.57, 164.0], [743.46, 163.82]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[767.61, 212.86], [751.55, 212.25], [760.58, 211.81], [776.59, 212.41], [767.61, 164.37], [751.55, 164.71], [760.58, 164.95], [776.59, 164.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[767.5, 240.55], [796.25, 241.17], [796.06, 244.34], [765.8, 243.66], [767.5, 166.44], [796.25, 166.28], [796.06, 165.47], [765.8, 165.65]]}]\n```" - } - ] - }, - { - "id": 963, - "data_source": "Mono3DRefer", - "scene_name": "000559", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[528.25, 209.25], [489.56, 209.29], [499.18, 206.28], [533.67, 206.24], [528.25, 169.4], [489.56, 169.38], [499.18, 170.69], [533.67, 170.71]]}]\n```" - } - ] - }, - { - "id": 964, - "data_source": "Mono3DRefer", - "scene_name": "000559", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[610.55, 319.11], [684.15, 318.98], [694.75, 335.09], [612.52, 335.25], [610.55, 167.35], [684.15, 167.36], [694.75, 165.7], [612.52, 165.68]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[664.24, 314.03], [740.78, 313.61], [761.29, 330.4], [675.06, 330.93], [664.24, 159.14], [740.78, 159.21], [761.29, 156.37], [675.06, 156.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[205.59, 321.77], [290.15, 320.36], [255.59, 339.86], [158.61, 341.7], [205.59, 163.61], [290.15, 163.79], [255.59, 161.3], [158.61, 161.07]]}, {\"category\": \"cyclist\", \"corners_3d\": [[310.77, 425.55], [190.07, 425.59], [297.79, 361.42], [386.75, 361.4], [310.77, 172.35], [190.07, 172.35], [297.79, 174.76], [386.75, 174.76]]}, {\"category\": \"cyclist\", \"corners_3d\": [[419.17, 295.5], [374.41, 295.22], [415.5, 276.84], [453.09, 277.04], [419.17, 164.58], [374.41, 164.62], [415.5, 167.35], [453.09, 167.32]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[727.84, 231.26], [757.52, 230.78], [774.57, 233.43], [743.46, 233.96], [727.84, 164.73], [757.52, 164.89], [774.57, 164.0], [743.46, 163.82]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[767.61, 212.86], [751.55, 212.25], [760.58, 211.81], [776.59, 212.41], [767.61, 164.37], [751.55, 164.71], [760.58, 164.95], [776.59, 164.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[767.5, 240.55], [796.25, 241.17], [796.06, 244.34], [765.8, 243.66], [767.5, 166.44], [796.25, 166.28], [796.06, 165.47], [765.8, 165.65]]}]\n```" - } - ] - }, - { - "id": 965, - "data_source": "Mono3DRefer", - "scene_name": "000559", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[310.77, 425.55], [190.07, 425.59], [297.79, 361.42], [386.75, 361.4], [310.77, 172.35], [190.07, 172.35], [297.79, 174.76], [386.75, 174.76]]}, {\"category\": \"cyclist\", \"corners_3d\": [[419.17, 295.5], [374.41, 295.22], [415.5, 276.84], [453.09, 277.04], [419.17, 164.58], [374.41, 164.62], [415.5, 167.35], [453.09, 167.32]]}, {\"category\": \"car\", \"corners_3d\": [[528.25, 209.25], [489.56, 209.29], [499.18, 206.28], [533.67, 206.24], [528.25, 169.4], [489.56, 169.38], [499.18, 170.69], [533.67, 170.71]]}]\n```" - } - ] - }, - { - "id": 966, - "data_source": "Mono3DRefer", - "scene_name": "000559", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[610.55, 319.11], [684.15, 318.98], [694.75, 335.09], [612.52, 335.25], [610.55, 167.35], [684.15, 167.36], [694.75, 165.7], [612.52, 165.68]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[664.24, 314.03], [740.78, 313.61], [761.29, 330.4], [675.06, 330.93], [664.24, 159.14], [740.78, 159.21], [761.29, 156.37], [675.06, 156.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[205.59, 321.77], [290.15, 320.36], [255.59, 339.86], [158.61, 341.7], [205.59, 163.61], [290.15, 163.79], [255.59, 161.3], [158.61, 161.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[727.84, 231.26], [757.52, 230.78], [774.57, 233.43], [743.46, 233.96], [727.84, 164.73], [757.52, 164.89], [774.57, 164.0], [743.46, 163.82]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[767.61, 212.86], [751.55, 212.25], [760.58, 211.81], [776.59, 212.41], [767.61, 164.37], [751.55, 164.71], [760.58, 164.95], [776.59, 164.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[767.5, 240.55], [796.25, 241.17], [796.06, 244.34], [765.8, 243.66], [767.5, 166.44], [796.25, 166.28], [796.06, 165.47], [765.8, 165.65]]}, {\"category\": \"car\", \"corners_3d\": [[528.25, 209.25], [489.56, 209.29], [499.18, 206.28], [533.67, 206.24], [528.25, 169.4], [489.56, 169.38], [499.18, 170.69], [533.67, 170.71]]}]\n```" - } - ] - }, - { - "id": 967, - "data_source": "Mono3DRefer", - "scene_name": "000559", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[610.55, 319.11], [684.15, 318.98], [694.75, 335.09], [612.52, 335.25], [610.55, 167.35], [684.15, 167.36], [694.75, 165.7], [612.52, 165.68]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[664.24, 314.03], [740.78, 313.61], [761.29, 330.4], [675.06, 330.93], [664.24, 159.14], [740.78, 159.21], [761.29, 156.37], [675.06, 156.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[205.59, 321.77], [290.15, 320.36], [255.59, 339.86], [158.61, 341.7], [205.59, 163.61], [290.15, 163.79], [255.59, 161.3], [158.61, 161.07]]}, {\"category\": \"cyclist\", \"corners_3d\": [[310.77, 425.55], [190.07, 425.59], [297.79, 361.42], [386.75, 361.4], [310.77, 172.35], [190.07, 172.35], [297.79, 174.76], [386.75, 174.76]]}, {\"category\": \"cyclist\", \"corners_3d\": [[419.17, 295.5], [374.41, 295.22], [415.5, 276.84], [453.09, 277.04], [419.17, 164.58], [374.41, 164.62], [415.5, 167.35], [453.09, 167.32]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[727.84, 231.26], [757.52, 230.78], [774.57, 233.43], [743.46, 233.96], [727.84, 164.73], [757.52, 164.89], [774.57, 164.0], [743.46, 163.82]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[767.61, 212.86], [751.55, 212.25], [760.58, 211.81], [776.59, 212.41], [767.61, 164.37], [751.55, 164.71], [760.58, 164.95], [776.59, 164.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[767.5, 240.55], [796.25, 241.17], [796.06, 244.34], [765.8, 243.66], [767.5, 166.44], [796.25, 166.28], [796.06, 165.47], [765.8, 165.65]]}, {\"category\": \"car\", \"corners_3d\": [[528.25, 209.25], [489.56, 209.29], [499.18, 206.28], [533.67, 206.24], [528.25, 169.4], [489.56, 169.38], [499.18, 170.69], [533.67, 170.71]]}]\n```" - } - ] - }, - { - "id": 968, - "data_source": "Mono3DRefer", - "scene_name": "000560", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[582.6, 244.72], [652.85, 244.72], [661.71, 259.61], [576.89, 259.61], [582.6, 177.31], [652.85, 177.31], [661.71, 178.23], [576.89, 178.23]]}]\n```" - } - ] - }, - { - "id": 969, - "data_source": "Mono3DRefer", - "scene_name": "000560", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[816.36, 229.89], [836.86, 229.81], [868.94, 236.76], [845.98, 236.86], [816.36, 158.28], [836.86, 158.3], [868.94, 156.53], [845.98, 156.5]]}]\n```" - } - ] - }, - { - "id": 970, - "data_source": "Mono3DRefer", - "scene_name": "000560", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[458.72, 221.82], [532.45, 221.82], [469.7, 261.38], [336.44, 261.36], [458.72, 118.49], [532.45, 118.48], [469.7, 74.57], [336.44, 74.58]]}]\n```" - } - ] - }, - { - "id": 971, - "data_source": "Mono3DRefer", - "scene_name": "000560", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[582.6, 244.72], [652.85, 244.72], [661.71, 259.61], [576.89, 259.61], [582.6, 177.31], [652.85, 177.31], [661.71, 178.23], [576.89, 178.23]]}, {\"category\": \"cyclist\", \"corners_3d\": [[816.36, 229.89], [836.86, 229.81], [868.94, 236.76], [845.98, 236.86], [816.36, 158.28], [836.86, 158.3], [868.94, 156.53], [845.98, 156.5]]}]\n```" - } - ] - }, - { - "id": 972, - "data_source": "Mono3DRefer", - "scene_name": "000560", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[582.6, 244.72], [652.85, 244.72], [661.71, 259.61], [576.89, 259.61], [582.6, 177.31], [652.85, 177.31], [661.71, 178.23], [576.89, 178.23]]}, {\"category\": \"tram\", \"corners_3d\": [[458.72, 221.82], [532.45, 221.82], [469.7, 261.38], [336.44, 261.36], [458.72, 118.49], [532.45, 118.48], [469.7, 74.57], [336.44, 74.58]]}]\n```" - } - ] - }, - { - "id": 973, - "data_source": "Mono3DRefer", - "scene_name": "000560", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[816.36, 229.89], [836.86, 229.81], [868.94, 236.76], [845.98, 236.86], [816.36, 158.28], [836.86, 158.3], [868.94, 156.53], [845.98, 156.5]]}, {\"category\": \"tram\", \"corners_3d\": [[458.72, 221.82], [532.45, 221.82], [469.7, 261.38], [336.44, 261.36], [458.72, 118.49], [532.45, 118.48], [469.7, 74.57], [336.44, 74.58]]}]\n```" - } - ] - }, - { - "id": 974, - "data_source": "Mono3DRefer", - "scene_name": "000560", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[582.6, 244.72], [652.85, 244.72], [661.71, 259.61], [576.89, 259.61], [582.6, 177.31], [652.85, 177.31], [661.71, 178.23], [576.89, 178.23]]}, {\"category\": \"cyclist\", \"corners_3d\": [[816.36, 229.89], [836.86, 229.81], [868.94, 236.76], [845.98, 236.86], [816.36, 158.28], [836.86, 158.3], [868.94, 156.53], [845.98, 156.5]]}, {\"category\": \"tram\", \"corners_3d\": [[458.72, 221.82], [532.45, 221.82], [469.7, 261.38], [336.44, 261.36], [458.72, 118.49], [532.45, 118.48], [469.7, 74.57], [336.44, 74.58]]}]\n```" - } - ] - }, - { - "id": 975, - "data_source": "Mono3DRefer", - "scene_name": "000561", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[601.14, 190.38], [624.43, 190.38], [626.99, 193.51], [599.54, 193.51], [601.14, 165.15], [624.43, 165.15], [626.99, 163.77], [599.54, 163.77]]}]\n```" - } - ] - }, - { - "id": 976, - "data_source": "Mono3DRefer", - "scene_name": "000561", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[766.04, 281.41], [869.55, 279.82], [1027.83, 330.85], [877.33, 334.33], [766.04, 184.44], [869.55, 184.27], [1027.83, 189.72], [877.33, 190.09]]}]\n```" - } - ] - }, - { - "id": 977, - "data_source": "Mono3DRefer", - "scene_name": "000561", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[766.04, 281.41], [869.55, 279.82], [1027.83, 330.85], [877.33, 334.33], [766.04, 184.44], [869.55, 184.27], [1027.83, 189.72], [877.33, 190.09]]}, {\"category\": \"bus\", \"corners_3d\": [[601.14, 190.38], [624.43, 190.38], [626.99, 193.51], [599.54, 193.51], [601.14, 165.15], [624.43, 165.15], [626.99, 163.77], [599.54, 163.77]]}]\n```" - } - ] - }, - { - "id": 978, - "data_source": "Mono3DRefer", - "scene_name": "000563", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[612.4, 206.09], [592.67, 206.1], [593.32, 203.73], [611.64, 203.72], [612.4, 188.6], [592.67, 188.61], [593.32, 187.48], [611.64, 187.48]]}]\n```" - } - ] - }, - { - "id": 979, - "data_source": "Mono3DRefer", - "scene_name": "000570", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[616.81, 201.89], [597.88, 201.9], [598.15, 199.9], [615.77, 199.9], [616.81, 185.11], [597.88, 185.11], [598.15, 184.27], [615.77, 184.27]]}]\n```" - } - ] - }, - { - "id": 980, - "data_source": "Mono3DRefer", - "scene_name": "000572", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[515.15, 200.95], [542.34, 200.94], [537.29, 202.49], [507.4, 202.51], [515.15, 174.25], [542.34, 174.26], [537.29, 173.17], [507.4, 173.16]]}, {\"category\": \"car\", \"corners_3d\": [[575.38, 207.26], [611.34, 207.24], [613.53, 209.94], [573.17, 209.96], [575.38, 174.57], [611.34, 174.58], [613.53, 173.27], [573.17, 173.26]]}, {\"category\": \"car\", \"corners_3d\": [[-2.17, 289.92], [107.86, 289.94], [-192.76, 352.91], [-368.88, 352.88], [-2.17, 197.08], [107.86, 197.08], [-192.76, 204.21], [-368.88, 204.21]]}, {\"category\": \"car\", \"corners_3d\": [[305.41, 274.09], [397.86, 274.1], [305.81, 313.08], [172.84, 313.06], [305.41, 194.57], [397.86, 194.57], [305.81, 198.68], [172.84, 198.68]]}, {\"category\": \"car\", \"corners_3d\": [[264.5, 236.59], [320.89, 236.59], [257.15, 248.01], [188.23, 248.0], [264.5, 185.18], [320.89, 185.18], [257.15, 185.17], [188.23, 185.17]]}, {\"category\": \"car\", \"corners_3d\": [[394.98, 234.35], [452.99, 234.35], [415.64, 246.21], [343.63, 246.21], [394.98, 181.64], [452.99, 181.64], [415.64, 180.77], [343.63, 180.77]]}]\n```" - } - ] - }, - { - "id": 981, - "data_source": "Mono3DRefer", - "scene_name": "000572", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[486.88, 211.0], [530.22, 210.98], [518.64, 215.23], [468.14, 215.25], [486.88, 158.95], [530.22, 158.96], [518.64, 154.64], [468.14, 154.62]]}, {\"category\": \"van\", \"corners_3d\": [[539.82, 192.11], [559.77, 192.1], [557.33, 192.74], [535.53, 192.75], [539.82, 167.25], [559.77, 167.26], [557.33, 165.61], [535.53, 165.59]]}, {\"category\": \"van\", \"corners_3d\": [[386.9, 215.71], [426.88, 215.69], [401.32, 220.17], [355.44, 220.2], [386.9, 172.93], [426.88, 172.94], [401.32, 171.13], [355.44, 171.13]]}]\n```" - } - ] - }, - { - "id": 982, - "data_source": "Mono3DRefer", - "scene_name": "000572", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[486.88, 211.0], [530.22, 210.98], [518.64, 215.23], [468.14, 215.25], [486.88, 158.95], [530.22, 158.96], [518.64, 154.64], [468.14, 154.62]]}, {\"category\": \"car\", \"corners_3d\": [[515.15, 200.95], [542.34, 200.94], [537.29, 202.49], [507.4, 202.51], [515.15, 174.25], [542.34, 174.26], [537.29, 173.17], [507.4, 173.16]]}, {\"category\": \"van\", \"corners_3d\": [[539.82, 192.11], [559.77, 192.1], [557.33, 192.74], [535.53, 192.75], [539.82, 167.25], [559.77, 167.26], [557.33, 165.61], [535.53, 165.59]]}, {\"category\": \"car\", \"corners_3d\": [[575.38, 207.26], [611.34, 207.24], [613.53, 209.94], [573.17, 209.96], [575.38, 174.57], [611.34, 174.58], [613.53, 173.27], [573.17, 173.26]]}, {\"category\": \"car\", \"corners_3d\": [[-2.17, 289.92], [107.86, 289.94], [-192.76, 352.91], [-368.88, 352.88], [-2.17, 197.08], [107.86, 197.08], [-192.76, 204.21], [-368.88, 204.21]]}, {\"category\": \"car\", \"corners_3d\": [[305.41, 274.09], [397.86, 274.1], [305.81, 313.08], [172.84, 313.06], [305.41, 194.57], [397.86, 194.57], [305.81, 198.68], [172.84, 198.68]]}, {\"category\": \"car\", \"corners_3d\": [[264.5, 236.59], [320.89, 236.59], [257.15, 248.01], [188.23, 248.0], [264.5, 185.18], [320.89, 185.18], [257.15, 185.17], [188.23, 185.17]]}, {\"category\": \"car\", \"corners_3d\": [[394.98, 234.35], [452.99, 234.35], [415.64, 246.21], [343.63, 246.21], [394.98, 181.64], [452.99, 181.64], [415.64, 180.77], [343.63, 180.77]]}, {\"category\": \"van\", \"corners_3d\": [[386.9, 215.71], [426.88, 215.69], [401.32, 220.17], [355.44, 220.2], [386.9, 172.93], [426.88, 172.94], [401.32, 171.13], [355.44, 171.13]]}]\n```" - } - ] - }, - { - "id": 983, - "data_source": "Mono3DRefer", - "scene_name": "000574", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[517.52, 205.51], [511.8, 206.76], [439.52, 206.48], [447.84, 205.25], [517.52, 183.95], [511.8, 184.37], [439.52, 184.28], [447.84, 183.86]]}]\n```" - } - ] - }, - { - "id": 984, - "data_source": "Mono3DRefer", - "scene_name": "000578", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[836.92, 202.83], [802.12, 202.41], [806.11, 199.3], [837.25, 199.63], [836.92, 162.43], [802.12, 162.57], [806.11, 163.66], [837.25, 163.54]]}]\n```" - } - ] - }, - { - "id": 985, - "data_source": "Mono3DRefer", - "scene_name": "000578", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[643.37, 276.8], [564.59, 272.73], [637.27, 249.14], [699.1, 251.49], [643.37, 192.72], [564.59, 191.94], [637.27, 187.43], [699.1, 187.88]]}, {\"category\": \"car\", \"corners_3d\": [[837.02, 237.22], [775.57, 235.38], [791.26, 226.03], [843.69, 227.35], [837.02, 184.35], [775.57, 184.02], [791.26, 182.35], [843.69, 182.59]]}, {\"category\": \"car\", \"corners_3d\": [[789.49, 282.36], [686.71, 276.25], [740.39, 252.71], [821.13, 256.31], [789.49, 192.52], [686.71, 191.43], [740.39, 187.2], [821.13, 187.84]]}, {\"category\": \"car\", \"corners_3d\": [[714.73, 246.29], [650.33, 243.82], [694.06, 231.24], [747.97, 232.9], [714.73, 188.98], [650.33, 188.44], [694.06, 185.68], [747.97, 186.04]]}, {\"category\": \"car\", \"corners_3d\": [[768.37, 224.5], [729.53, 223.46], [750.32, 217.48], [784.87, 218.29], [768.37, 186.32], [729.53, 186.05], [750.32, 184.49], [784.87, 184.7]]}, {\"category\": \"car\", \"corners_3d\": [[819.23, 208.45], [786.93, 208.0], [790.03, 205.59], [820.13, 205.98], [819.23, 181.03], [786.93, 180.93], [790.03, 180.38], [820.13, 180.47]]}]\n```" - } - ] - }, - { - "id": 986, - "data_source": "Mono3DRefer", - "scene_name": "000578", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[643.37, 276.8], [564.59, 272.73], [637.27, 249.14], [699.1, 251.49], [643.37, 192.72], [564.59, 191.94], [637.27, 187.43], [699.1, 187.88]]}, {\"category\": \"car\", \"corners_3d\": [[837.02, 237.22], [775.57, 235.38], [791.26, 226.03], [843.69, 227.35], [837.02, 184.35], [775.57, 184.02], [791.26, 182.35], [843.69, 182.59]]}, {\"category\": \"car\", \"corners_3d\": [[789.49, 282.36], [686.71, 276.25], [740.39, 252.71], [821.13, 256.31], [789.49, 192.52], [686.71, 191.43], [740.39, 187.2], [821.13, 187.84]]}, {\"category\": \"car\", \"corners_3d\": [[714.73, 246.29], [650.33, 243.82], [694.06, 231.24], [747.97, 232.9], [714.73, 188.98], [650.33, 188.44], [694.06, 185.68], [747.97, 186.04]]}, {\"category\": \"car\", \"corners_3d\": [[768.37, 224.5], [729.53, 223.46], [750.32, 217.48], [784.87, 218.29], [768.37, 186.32], [729.53, 186.05], [750.32, 184.49], [784.87, 184.7]]}, {\"category\": \"van\", \"corners_3d\": [[836.92, 202.83], [802.12, 202.41], [806.11, 199.3], [837.25, 199.63], [836.92, 162.43], [802.12, 162.57], [806.11, 163.66], [837.25, 163.54]]}, {\"category\": \"car\", \"corners_3d\": [[819.23, 208.45], [786.93, 208.0], [790.03, 205.59], [820.13, 205.98], [819.23, 181.03], [786.93, 180.93], [790.03, 180.38], [820.13, 180.47]]}]\n```" - } - ] - }, - { - "id": 987, - "data_source": "Mono3DRefer", - "scene_name": "000580", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[491.45, 210.13], [483.48, 211.9], [394.29, 211.55], [406.18, 209.82], [491.45, 183.72], [483.48, 184.24], [394.29, 184.14], [406.18, 183.63]]}]\n```" - } - ] - }, - { - "id": 988, - "data_source": "Mono3DRefer", - "scene_name": "000581", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[586.66, 240.46], [551.84, 240.53], [556.58, 233.02], [587.54, 232.97], [586.66, 208.26], [551.84, 208.3], [556.58, 204.37], [587.54, 204.34]]}]\n```" - } - ] - }, - { - "id": 989, - "data_source": "Mono3DRefer", - "scene_name": "000583", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[578.04, 211.54], [615.01, 211.5], [616.93, 215.34], [576.29, 215.38], [578.04, 176.05], [615.01, 176.05], [616.93, 176.37], [576.29, 176.37]]}, {\"category\": \"car\", \"corners_3d\": [[359.58, 221.54], [324.61, 221.59], [354.11, 216.27], [385.24, 216.23], [359.58, 191.72], [324.61, 191.74], [354.11, 189.67], [385.24, 189.66]]}, {\"category\": \"car\", \"corners_3d\": [[432.88, 205.0], [410.0, 204.98], [424.42, 202.81], [445.76, 202.83], [432.88, 184.55], [410.0, 184.55], [424.42, 183.76], [445.76, 183.76]]}]\n```" - } - ] - }, - { - "id": 990, - "data_source": "Mono3DRefer", - "scene_name": "000583", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1025.47, 291.08], [1068.3, 291.36], [1188.57, 324.55], [1133.42, 324.1], [1025.47, 147.6], [1068.3, 147.54], [1188.57, 140.45], [1133.42, 140.55]]}]\n```" - } - ] - }, - { - "id": 991, - "data_source": "Mono3DRefer", - "scene_name": "000583", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[520.82, 199.07], [552.01, 199.04], [539.73, 207.86], [498.0, 207.91], [520.82, 155.73], [552.01, 155.75], [539.73, 149.99], [498.0, 149.95]]}]\n```" - } - ] - }, - { - "id": 992, - "data_source": "Mono3DRefer", - "scene_name": "000583", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[578.04, 211.54], [615.01, 211.5], [616.93, 215.34], [576.29, 215.38], [578.04, 176.05], [615.01, 176.05], [616.93, 176.37], [576.29, 176.37]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1025.47, 291.08], [1068.3, 291.36], [1188.57, 324.55], [1133.42, 324.1], [1025.47, 147.6], [1068.3, 147.54], [1188.57, 140.45], [1133.42, 140.55]]}, {\"category\": \"car\", \"corners_3d\": [[359.58, 221.54], [324.61, 221.59], [354.11, 216.27], [385.24, 216.23], [359.58, 191.72], [324.61, 191.74], [354.11, 189.67], [385.24, 189.66]]}, {\"category\": \"car\", \"corners_3d\": [[432.88, 205.0], [410.0, 204.98], [424.42, 202.81], [445.76, 202.83], [432.88, 184.55], [410.0, 184.55], [424.42, 183.76], [445.76, 183.76]]}]\n```" - } - ] - }, - { - "id": 993, - "data_source": "Mono3DRefer", - "scene_name": "000583", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[578.04, 211.54], [615.01, 211.5], [616.93, 215.34], [576.29, 215.38], [578.04, 176.05], [615.01, 176.05], [616.93, 176.37], [576.29, 176.37]]}, {\"category\": \"tram\", \"corners_3d\": [[520.82, 199.07], [552.01, 199.04], [539.73, 207.86], [498.0, 207.91], [520.82, 155.73], [552.01, 155.75], [539.73, 149.99], [498.0, 149.95]]}, {\"category\": \"car\", \"corners_3d\": [[359.58, 221.54], [324.61, 221.59], [354.11, 216.27], [385.24, 216.23], [359.58, 191.72], [324.61, 191.74], [354.11, 189.67], [385.24, 189.66]]}, {\"category\": \"car\", \"corners_3d\": [[432.88, 205.0], [410.0, 204.98], [424.42, 202.81], [445.76, 202.83], [432.88, 184.55], [410.0, 184.55], [424.42, 183.76], [445.76, 183.76]]}]\n```" - } - ] - }, - { - "id": 994, - "data_source": "Mono3DRefer", - "scene_name": "000583", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1025.47, 291.08], [1068.3, 291.36], [1188.57, 324.55], [1133.42, 324.1], [1025.47, 147.6], [1068.3, 147.54], [1188.57, 140.45], [1133.42, 140.55]]}, {\"category\": \"tram\", \"corners_3d\": [[520.82, 199.07], [552.01, 199.04], [539.73, 207.86], [498.0, 207.91], [520.82, 155.73], [552.01, 155.75], [539.73, 149.99], [498.0, 149.95]]}]\n```" - } - ] - }, - { - "id": 995, - "data_source": "Mono3DRefer", - "scene_name": "000583", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[578.04, 211.54], [615.01, 211.5], [616.93, 215.34], [576.29, 215.38], [578.04, 176.05], [615.01, 176.05], [616.93, 176.37], [576.29, 176.37]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1025.47, 291.08], [1068.3, 291.36], [1188.57, 324.55], [1133.42, 324.1], [1025.47, 147.6], [1068.3, 147.54], [1188.57, 140.45], [1133.42, 140.55]]}, {\"category\": \"tram\", \"corners_3d\": [[520.82, 199.07], [552.01, 199.04], [539.73, 207.86], [498.0, 207.91], [520.82, 155.73], [552.01, 155.75], [539.73, 149.99], [498.0, 149.95]]}, {\"category\": \"car\", \"corners_3d\": [[359.58, 221.54], [324.61, 221.59], [354.11, 216.27], [385.24, 216.23], [359.58, 191.72], [324.61, 191.74], [354.11, 189.67], [385.24, 189.66]]}, {\"category\": \"car\", \"corners_3d\": [[432.88, 205.0], [410.0, 204.98], [424.42, 202.81], [445.76, 202.83], [432.88, 184.55], [410.0, 184.55], [424.42, 183.76], [445.76, 183.76]]}]\n```" - } - ] - }, - { - "id": 996, - "data_source": "Mono3DRefer", - "scene_name": "000584", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[503.2, 240.93], [440.27, 241.75], [453.25, 228.41], [503.98, 227.87], [503.2, 186.38], [440.27, 186.54], [453.25, 183.89], [503.98, 183.79]]}]\n```" - } - ] - }, - { - "id": 997, - "data_source": "Mono3DRefer", - "scene_name": "000587", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[559.43, 206.89], [523.75, 206.94], [530.54, 203.3], [562.4, 203.26], [559.43, 169.42], [523.75, 169.42], [530.54, 169.79], [562.4, 169.79]]}, {\"category\": \"car\", \"corners_3d\": [[573.67, 192.98], [555.99, 192.97], [559.88, 191.81], [576.54, 191.82], [573.67, 173.31], [555.99, 173.31], [559.88, 173.29], [576.54, 173.29]]}]\n```" - } - ] - }, - { - "id": 998, - "data_source": "Mono3DRefer", - "scene_name": "000593", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[602.88, 208.65], [569.43, 208.44], [583.91, 204.59], [613.81, 204.76], [602.88, 179.52], [569.43, 179.48], [583.91, 178.77], [613.81, 178.8]]}]\n```" - } - ] - }, - { - "id": 999, - "data_source": "Mono3DRefer", - "scene_name": "000594", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[586.1, 221.79], [634.19, 221.82], [637.14, 230.4], [580.63, 230.35], [586.1, 176.34], [634.19, 176.34], [637.14, 176.95], [580.63, 176.95]]}, {\"category\": \"car\", \"corners_3d\": [[399.75, 212.53], [367.57, 212.51], [395.14, 208.11], [423.76, 208.12], [399.75, 186.0], [367.57, 186.0], [395.14, 184.54], [423.76, 184.55]]}, {\"category\": \"car\", \"corners_3d\": [[478.67, 199.02], [459.32, 199.0], [469.66, 197.42], [487.84, 197.44], [478.67, 181.32], [459.32, 181.31], [469.66, 180.8], [487.84, 180.81]]}, {\"category\": \"car\", \"corners_3d\": [[665.29, 188.89], [678.91, 189.19], [650.06, 189.86], [636.44, 189.53], [665.29, 171.23], [678.91, 171.2], [650.06, 171.14], [636.44, 171.17]]}]\n```" - } - ] - }, - { - "id": 1000, - "data_source": "Mono3DRefer", - "scene_name": "000597", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1505.69, 368.74], [1243.32, 366.07], [956.48, 274.53], [1091.58, 275.27], [1505.69, 146.08], [1243.32, 146.45], [956.48, 158.96], [1091.58, 158.86]]}, {\"category\": \"car\", \"corners_3d\": [[282.98, 379.98], [98.04, 377.31], [298.29, 302.47], [416.63, 303.54], [282.98, 198.71], [98.04, 198.37], [298.29, 189.03], [416.63, 189.17]]}, {\"category\": \"car\", \"corners_3d\": [[467.96, 274.79], [384.31, 274.21], [443.1, 251.36], [508.07, 251.71], [467.96, 192.06], [384.31, 191.95], [443.1, 187.64], [508.07, 187.71]]}, {\"category\": \"car\", \"corners_3d\": [[522.88, 244.74], [452.82, 244.67], [485.01, 230.52], [541.28, 230.56], [522.88, 184.26], [452.82, 184.25], [485.01, 182.0], [541.28, 182.01]]}, {\"category\": \"car\", \"corners_3d\": [[701.77, 228.81], [762.92, 229.71], [766.02, 242.3], [691.52, 240.97], [701.77, 173.54], [762.92, 173.55], [766.02, 173.71], [691.52, 173.69]]}, {\"category\": \"car\", \"corners_3d\": [[1150.96, 252.74], [1226.37, 264.84], [922.25, 262.62], [887.57, 251.07], [1150.96, 162.54], [1226.37, 160.98], [922.25, 161.27], [887.57, 162.76]]}, {\"category\": \"car\", \"corners_3d\": [[540.85, 220.75], [501.78, 220.6], [518.06, 215.43], [552.94, 215.56], [540.85, 176.0], [501.78, 175.99], [518.06, 175.65], [552.94, 175.66]]}, {\"category\": \"car\", \"corners_3d\": [[566.0, 209.09], [529.47, 209.82], [507.32, 206.05], [540.6, 205.46], [566.0, 176.38], [529.47, 176.45], [507.32, 176.09], [540.6, 176.03]]}, {\"category\": \"car\", \"corners_3d\": [[442.69, 193.82], [445.25, 194.43], [383.72, 195.03], [382.87, 194.39], [442.69, 174.08], [445.25, 174.12], [383.72, 174.15], [382.87, 174.12]]}, {\"category\": \"car\", \"corners_3d\": [[740.67, 219.29], [786.65, 220.67], [760.18, 227.32], [708.87, 225.54], [740.67, 175.46], [786.65, 175.53], [760.18, 175.91], [708.87, 175.81]]}, {\"category\": \"car\", \"corners_3d\": [[793.73, 210.11], [831.91, 211.59], [791.06, 215.1], [751.27, 213.34], [793.73, 169.3], [831.91, 169.15], [791.06, 168.82], [751.27, 168.99]]}, {\"category\": \"car\", \"corners_3d\": [[921.23, 187.79], [898.39, 187.21], [952.13, 186.73], [976.29, 187.28], [921.23, 161.41], [898.39, 161.86], [952.13, 162.22], [976.29, 161.8]]}]\n```" - } - ] - }, - { - "id": 1001, - "data_source": "Mono3DRefer", - "scene_name": "000599", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[715.32, 204.87], [703.66, 204.86], [701.74, 204.16], [713.14, 204.16], [715.32, 170.77], [703.66, 170.77], [701.74, 170.81], [713.14, 170.81]]}]\n```" - } - ] - }, - { - "id": 1002, - "data_source": "Mono3DRefer", - "scene_name": "000599", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[605.56, 199.37], [630.13, 199.38], [631.19, 201.59], [604.58, 201.57], [605.56, 176.16], [630.13, 176.17], [631.19, 176.44], [604.58, 176.44]]}, {\"category\": \"car\", \"corners_3d\": [[697.22, 235.22], [757.68, 235.22], [798.62, 252.53], [721.38, 252.52], [697.22, 177.85], [757.68, 177.85], [798.62, 179.23], [721.38, 179.23]]}, {\"category\": \"car\", \"corners_3d\": [[806.22, 315.92], [946.72, 315.94], [1232.47, 437.42], [972.64, 437.35], [806.22, 177.24], [946.72, 177.24], [1232.47, 180.96], [972.64, 180.96]]}, {\"category\": \"car\", \"corners_3d\": [[297.04, 244.85], [276.6, 250.04], [137.38, 250.62], [167.39, 245.36], [297.04, 195.99], [276.6, 197.65], [137.38, 197.84], [167.39, 196.15]]}, {\"category\": \"car\", \"corners_3d\": [[121.29, 248.83], [83.69, 254.51], [-60.28, 254.18], [-12.84, 248.54], [121.29, 197.96], [83.69, 199.83], [-60.28, 199.72], [-12.84, 197.86]]}, {\"category\": \"car\", \"corners_3d\": [[687.15, 225.06], [734.09, 225.06], [756.7, 234.59], [701.19, 234.58], [687.15, 177.19], [734.09, 177.19], [756.7, 177.99], [701.19, 177.99]]}, {\"category\": \"car\", \"corners_3d\": [[659.22, 205.22], [689.42, 205.21], [699.69, 209.05], [665.91, 209.06], [659.22, 176.16], [689.42, 176.16], [699.69, 176.55], [665.91, 176.56]]}]\n```" - } - ] - }, - { - "id": 1003, - "data_source": "Mono3DRefer", - "scene_name": "000599", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[605.56, 199.37], [630.13, 199.38], [631.19, 201.59], [604.58, 201.57], [605.56, 176.16], [630.13, 176.17], [631.19, 176.44], [604.58, 176.44]]}, {\"category\": \"car\", \"corners_3d\": [[697.22, 235.22], [757.68, 235.22], [798.62, 252.53], [721.38, 252.52], [697.22, 177.85], [757.68, 177.85], [798.62, 179.23], [721.38, 179.23]]}, {\"category\": \"car\", \"corners_3d\": [[806.22, 315.92], [946.72, 315.94], [1232.47, 437.42], [972.64, 437.35], [806.22, 177.24], [946.72, 177.24], [1232.47, 180.96], [972.64, 180.96]]}, {\"category\": \"car\", \"corners_3d\": [[297.04, 244.85], [276.6, 250.04], [137.38, 250.62], [167.39, 245.36], [297.04, 195.99], [276.6, 197.65], [137.38, 197.84], [167.39, 196.15]]}, {\"category\": \"car\", \"corners_3d\": [[121.29, 248.83], [83.69, 254.51], [-60.28, 254.18], [-12.84, 248.54], [121.29, 197.96], [83.69, 199.83], [-60.28, 199.72], [-12.84, 197.86]]}, {\"category\": \"car\", \"corners_3d\": [[687.15, 225.06], [734.09, 225.06], [756.7, 234.59], [701.19, 234.58], [687.15, 177.19], [734.09, 177.19], [756.7, 177.99], [701.19, 177.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[715.32, 204.87], [703.66, 204.86], [701.74, 204.16], [713.14, 204.16], [715.32, 170.77], [703.66, 170.77], [701.74, 170.81], [713.14, 170.81]]}, {\"category\": \"car\", \"corners_3d\": [[659.22, 205.22], [689.42, 205.21], [699.69, 209.05], [665.91, 209.06], [659.22, 176.16], [689.42, 176.16], [699.69, 176.55], [665.91, 176.56]]}]\n```" - } - ] - }, - { - "id": 1004, - "data_source": "Mono3DRefer", - "scene_name": "000603", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[417.15, 211.18], [387.28, 210.99], [417.71, 208.06], [444.85, 208.21], [417.15, 181.47], [387.28, 181.46], [417.71, 181.37], [444.85, 181.38]]}, {\"category\": \"car\", \"corners_3d\": [[737.24, 247.91], [813.57, 248.84], [849.58, 269.97], [749.44, 268.38], [737.24, 175.23], [813.57, 175.16], [849.58, 173.51], [749.44, 173.63]]}, {\"category\": \"car\", \"corners_3d\": [[702.52, 229.22], [760.69, 229.26], [793.37, 239.96], [722.42, 239.9], [702.52, 175.82], [760.69, 175.81], [793.37, 174.78], [722.42, 174.79]]}, {\"category\": \"car\", \"corners_3d\": [[682.85, 217.12], [726.67, 217.23], [740.91, 223.27], [689.86, 223.11], [682.85, 176.9], [726.67, 176.89], [740.91, 176.3], [689.86, 176.31]]}, {\"category\": \"car\", \"corners_3d\": [[668.19, 209.99], [707.6, 210.06], [718.81, 214.49], [673.47, 214.4], [668.19, 176.41], [707.6, 176.4], [718.81, 175.78], [673.47, 175.8]]}, {\"category\": \"car\", \"corners_3d\": [[655.06, 201.31], [681.49, 201.35], [685.56, 203.09], [656.91, 203.06], [655.06, 177.31], [681.49, 177.31], [685.56, 177.04], [656.91, 177.04]]}]\n```" - } - ] - }, - { - "id": 1005, - "data_source": "Mono3DRefer", - "scene_name": "000606", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[520.64, 205.29], [515.18, 206.5], [444.77, 206.23], [452.7, 205.05], [520.64, 184.28], [515.18, 184.71], [444.77, 184.61], [452.7, 184.19]]}]\n```" - } - ] - }, - { - "id": 1006, - "data_source": "Mono3DRefer", - "scene_name": "000607", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[612.5, 205.0], [582.04, 204.83], [594.08, 201.65], [621.56, 201.79], [612.5, 178.54], [582.04, 178.51], [594.08, 177.95], [621.56, 177.98]]}]\n```" - } - ] - }, - { - "id": 1007, - "data_source": "Mono3DRefer", - "scene_name": "000609", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[298.13, 227.95], [259.32, 227.96], [298.91, 221.72], [333.33, 221.71], [298.13, 193.57], [259.32, 193.57], [298.91, 191.22], [333.33, 191.22]]}, {\"category\": \"car\", \"corners_3d\": [[380.51, 210.32], [354.53, 210.33], [376.16, 207.06], [399.85, 207.04], [380.51, 182.75], [354.53, 182.75], [376.16, 181.89], [399.85, 181.88]]}]\n```" - } - ] - }, - { - "id": 1008, - "data_source": "Mono3DRefer", - "scene_name": "000609", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[587.6, 201.38], [631.48, 201.4], [635.64, 209.81], [578.85, 209.77], [587.6, 147.96], [631.48, 147.94], [635.64, 140.61], [578.85, 140.63]]}]\n```" - } - ] - }, - { - "id": 1009, - "data_source": "Mono3DRefer", - "scene_name": "000609", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[298.13, 227.95], [259.32, 227.96], [298.91, 221.72], [333.33, 221.71], [298.13, 193.57], [259.32, 193.57], [298.91, 191.22], [333.33, 191.22]]}, {\"category\": \"car\", \"corners_3d\": [[380.51, 210.32], [354.53, 210.33], [376.16, 207.06], [399.85, 207.04], [380.51, 182.75], [354.53, 182.75], [376.16, 181.89], [399.85, 181.88]]}, {\"category\": \"truck\", \"corners_3d\": [[587.6, 201.38], [631.48, 201.4], [635.64, 209.81], [578.85, 209.77], [587.6, 147.96], [631.48, 147.94], [635.64, 140.61], [578.85, 140.63]]}]\n```" - } - ] - }, - { - "id": 1010, - "data_source": "Mono3DRefer", - "scene_name": "000610", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[583.83, 189.39], [599.96, 189.38], [600.56, 190.24], [583.59, 190.25], [583.83, 176.68], [599.96, 176.68], [600.56, 176.88], [583.59, 176.88]]}, {\"category\": \"car\", \"corners_3d\": [[442.57, 218.52], [399.73, 218.6], [423.27, 212.86], [460.7, 212.8], [442.57, 180.21], [399.73, 180.22], [423.27, 179.3], [460.7, 179.29]]}]\n```" - } - ] - }, - { - "id": 1011, - "data_source": "Mono3DRefer", - "scene_name": "000611", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[600.6, 203.14], [627.61, 203.14], [628.85, 205.28], [599.92, 205.28], [600.6, 177.22], [627.61, 177.22], [628.85, 177.53], [599.92, 177.53]]}, {\"category\": \"car\", \"corners_3d\": [[345.7, 220.98], [309.52, 220.96], [338.88, 216.35], [371.61, 216.37], [345.7, 188.89], [309.52, 188.88], [338.88, 187.35], [371.61, 187.35]]}]\n```" - } - ] - }, - { - "id": 1012, - "data_source": "Mono3DRefer", - "scene_name": "000611", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[639.65, 191.25], [663.27, 191.2], [683.49, 194.54], [655.62, 194.61], [639.65, 165.42], [663.27, 165.44], [683.49, 164.09], [655.62, 164.06]]}]\n```" - } - ] - }, - { - "id": 1013, - "data_source": "Mono3DRefer", - "scene_name": "000611", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[600.6, 203.14], [627.61, 203.14], [628.85, 205.28], [599.92, 205.28], [600.6, 177.22], [627.61, 177.22], [628.85, 177.53], [599.92, 177.53]]}, {\"category\": \"bus\", \"corners_3d\": [[639.65, 191.25], [663.27, 191.2], [683.49, 194.54], [655.62, 194.61], [639.65, 165.42], [663.27, 165.44], [683.49, 164.09], [655.62, 164.06]]}, {\"category\": \"car\", \"corners_3d\": [[345.7, 220.98], [309.52, 220.96], [338.88, 216.35], [371.61, 216.37], [345.7, 188.89], [309.52, 188.88], [338.88, 187.35], [371.61, 187.35]]}]\n```" - } - ] - }, - { - "id": 1014, - "data_source": "Mono3DRefer", - "scene_name": "000616", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[433.9, 293.39], [330.24, 293.4], [403.82, 261.58], [480.11, 261.57], [433.9, 178.25], [330.24, 178.25], [403.82, 176.82], [480.11, 176.82]]}, {\"category\": \"car\", \"corners_3d\": [[530.79, 233.17], [479.34, 233.26], [494.99, 225.05], [539.44, 224.99], [530.79, 182.9], [479.34, 182.91], [494.99, 181.55], [539.44, 181.53]]}, {\"category\": \"car\", \"corners_3d\": [[565.49, 206.09], [535.93, 205.72], [559.16, 202.84], [586.33, 203.15], [565.49, 174.66], [535.93, 174.64], [559.16, 174.48], [586.33, 174.5]]}]\n```" - } - ] - }, - { - "id": 1015, - "data_source": "Mono3DRefer", - "scene_name": "000621", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[349.47, 362.09], [196.93, 362.52], [336.44, 297.16], [436.29, 296.98], [349.47, 193.63], [196.93, 193.68], [336.44, 186.5], [436.29, 186.48]]}, {\"category\": \"car\", \"corners_3d\": [[512.24, 254.09], [450.93, 254.3], [472.41, 241.47], [524.05, 241.32], [512.24, 194.35], [450.93, 194.4], [472.41, 191.01], [524.05, 190.97]]}, {\"category\": \"car\", \"corners_3d\": [[556.41, 221.98], [523.4, 221.42], [548.84, 216.71], [578.88, 217.16], [556.41, 186.98], [523.4, 186.82], [548.84, 185.46], [578.88, 185.6]]}]\n```" - } - ] - }, - { - "id": 1016, - "data_source": "Mono3DRefer", - "scene_name": "000622", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[463.88, 210.31], [466.3, 209.92], [481.65, 210.02], [479.38, 210.41], [463.88, 179.31], [466.3, 179.25], [481.65, 179.26], [479.38, 179.33]]}]\n```" - } - ] - }, - { - "id": 1017, - "data_source": "Mono3DRefer", - "scene_name": "000622", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[684.59, 207.1], [655.29, 206.86], [662.05, 203.9], [688.82, 204.09], [684.59, 180.67], [655.29, 180.61], [662.05, 179.94], [688.82, 179.98]]}, {\"category\": \"car\", \"corners_3d\": [[744.76, 198.23], [761.42, 198.32], [762.97, 199.76], [745.37, 199.66], [744.76, 184.81], [761.42, 184.86], [762.97, 185.54], [745.37, 185.49]]}]\n```" - } - ] - }, - { - "id": 1018, - "data_source": "Mono3DRefer", - "scene_name": "000622", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[684.59, 207.1], [655.29, 206.86], [662.05, 203.9], [688.82, 204.09], [684.59, 180.67], [655.29, 180.61], [662.05, 179.94], [688.82, 179.98]]}, {\"category\": \"car\", \"corners_3d\": [[744.76, 198.23], [761.42, 198.32], [762.97, 199.76], [745.37, 199.66], [744.76, 184.81], [761.42, 184.86], [762.97, 185.54], [745.37, 185.49]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[463.88, 210.31], [466.3, 209.92], [481.65, 210.02], [479.38, 210.41], [463.88, 179.31], [466.3, 179.25], [481.65, 179.26], [479.38, 179.33]]}]\n```" - } - ] - }, - { - "id": 1019, - "data_source": "Mono3DRefer", - "scene_name": "000623", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-376.09, 580.49], [-754.94, 586.82], [-13.86, 358.26], [152.15, 356.98], [-376.09, 265.13], [-754.94, 266.56], [-13.86, 214.82], [152.15, 214.53]]}, {\"category\": \"car\", \"corners_3d\": [[666.5, 229.35], [722.93, 229.31], [749.25, 241.7], [680.46, 241.76], [666.5, 180.77], [722.93, 180.76], [749.25, 182.5], [680.46, 182.5]]}, {\"category\": \"car\", \"corners_3d\": [[534.04, 207.45], [503.93, 207.44], [513.6, 204.46], [541.12, 204.47], [534.04, 182.95], [503.93, 182.95], [513.6, 182.08], [541.12, 182.08]]}, {\"category\": \"car\", \"corners_3d\": [[460.13, 239.73], [406.04, 239.64], [439.85, 229.25], [485.56, 229.32], [460.13, 189.4], [406.04, 189.37], [439.85, 186.8], [485.56, 186.82]]}, {\"category\": \"car\", \"corners_3d\": [[487.75, 225.9], [443.04, 225.77], [467.94, 219.0], [506.98, 219.1], [487.75, 185.63], [443.04, 185.6], [467.94, 183.97], [506.98, 183.99]]}, {\"category\": \"car\", \"corners_3d\": [[349.64, 283.27], [263.65, 283.01], [351.13, 256.23], [416.33, 256.37], [349.64, 211.29], [263.65, 211.2], [351.13, 201.88], [416.33, 201.93]]}]\n```" - } - ] - }, - { - "id": 1020, - "data_source": "Mono3DRefer", - "scene_name": "000623", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[236.24, 553.67], [-226.48, 563.44], [271.13, 322.47], [446.32, 321.02], [236.24, 104.91], [-226.48, 103.17], [271.13, 146.16], [446.32, 146.42]]}]\n```" - } - ] - }, - { - "id": 1021, - "data_source": "Mono3DRefer", - "scene_name": "000623", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[236.24, 553.67], [-226.48, 563.44], [271.13, 322.47], [446.32, 321.02], [236.24, 104.91], [-226.48, 103.17], [271.13, 146.16], [446.32, 146.42]]}, {\"category\": \"car\", \"corners_3d\": [[-376.09, 580.49], [-754.94, 586.82], [-13.86, 358.26], [152.15, 356.98], [-376.09, 265.13], [-754.94, 266.56], [-13.86, 214.82], [152.15, 214.53]]}, {\"category\": \"car\", \"corners_3d\": [[666.5, 229.35], [722.93, 229.31], [749.25, 241.7], [680.46, 241.76], [666.5, 180.77], [722.93, 180.76], [749.25, 182.5], [680.46, 182.5]]}, {\"category\": \"car\", \"corners_3d\": [[534.04, 207.45], [503.93, 207.44], [513.6, 204.46], [541.12, 204.47], [534.04, 182.95], [503.93, 182.95], [513.6, 182.08], [541.12, 182.08]]}, {\"category\": \"car\", \"corners_3d\": [[460.13, 239.73], [406.04, 239.64], [439.85, 229.25], [485.56, 229.32], [460.13, 189.4], [406.04, 189.37], [439.85, 186.8], [485.56, 186.82]]}, {\"category\": \"car\", \"corners_3d\": [[487.75, 225.9], [443.04, 225.77], [467.94, 219.0], [506.98, 219.1], [487.75, 185.63], [443.04, 185.6], [467.94, 183.97], [506.98, 183.99]]}, {\"category\": \"car\", \"corners_3d\": [[349.64, 283.27], [263.65, 283.01], [351.13, 256.23], [416.33, 256.37], [349.64, 211.29], [263.65, 211.2], [351.13, 201.88], [416.33, 201.93]]}]\n```" - } - ] - }, - { - "id": 1022, - "data_source": "Mono3DRefer", - "scene_name": "000624", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[653.26, 233.23], [712.75, 233.39], [725.73, 243.09], [656.7, 242.89], [653.26, 173.6], [712.75, 173.6], [725.73, 173.72], [656.7, 173.72]]}, {\"category\": \"car\", \"corners_3d\": [[-450.82, 821.39], [-1002.42, 806.26], [86.62, 383.86], [276.02, 385.51], [-450.82, 318.49], [-1002.42, 315.09], [86.62, 220.24], [276.02, 220.61]]}, {\"category\": \"car\", \"corners_3d\": [[486.3, 248.72], [423.54, 247.91], [473.62, 234.12], [525.18, 234.65], [486.3, 185.2], [423.54, 185.07], [473.62, 182.82], [525.18, 182.91]]}]\n```" - } - ] - }, - { - "id": 1023, - "data_source": "Mono3DRefer", - "scene_name": "000625", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-1.18, 382.32], [267.68, 360.51], [54.41, 901.94], [-1803.74, 1453.0], [-1.18, 121.52], [267.68, 127.89], [54.41, -30.36], [-1803.74, -191.43]]}]\n```" - } - ] - }, - { - "id": 1024, - "data_source": "Mono3DRefer", - "scene_name": "000625", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[923.43, 333.12], [871.78, 335.48], [820.21, 315.71], [866.04, 313.91], [923.43, 144.3], [871.78, 143.74], [820.21, 148.43], [866.04, 148.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[797.14, 259.57], [768.44, 260.22], [745.16, 254.05], [771.84, 253.5], [797.14, 143.8], [768.44, 143.49], [745.16, 146.35], [771.84, 146.61]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[821.36, 258.06], [793.45, 258.76], [769.59, 253.28], [795.76, 252.67], [821.36, 146.16], [793.45, 145.85], [769.59, 148.28], [795.76, 148.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[648.85, 201.89], [665.16, 201.72], [674.03, 202.34], [657.31, 202.52], [648.85, 142.92], [665.16, 143.22], [674.03, 142.13], [657.31, 141.81]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[280.41, 255.48], [323.02, 253.77], [325.28, 257.67], [280.41, 259.57], [280.41, 140.16], [323.02, 141.08], [325.28, 138.98], [280.41, 137.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[501.92, 208.85], [482.7, 209.18], [476.52, 208.27], [495.2, 207.96], [501.92, 152.77], [482.7, 152.45], [476.52, 153.34], [495.2, 153.64]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[523.6, 209.24], [499.22, 209.62], [490.98, 208.27], [514.34, 207.93], [523.6, 149.46], [499.22, 149.05], [490.98, 150.5], [514.34, 150.88]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[452.19, 217.81], [425.15, 218.77], [412.56, 217.51], [439.04, 216.61], [452.19, 141.01], [425.15, 139.99], [412.56, 141.32], [439.04, 142.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[430.55, 208.19], [408.72, 208.72], [393.38, 207.37], [414.46, 206.89], [430.55, 141.45], [408.72, 140.7], [393.38, 142.6], [414.46, 143.28]]}]\n```" - } - ] - }, - { - "id": 1025, - "data_source": "Mono3DRefer", - "scene_name": "000625", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[923.43, 333.12], [871.78, 335.48], [820.21, 315.71], [866.04, 313.91], [923.43, 144.3], [871.78, 143.74], [820.21, 148.43], [866.04, 148.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[797.14, 259.57], [768.44, 260.22], [745.16, 254.05], [771.84, 253.5], [797.14, 143.8], [768.44, 143.49], [745.16, 146.35], [771.84, 146.61]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[821.36, 258.06], [793.45, 258.76], [769.59, 253.28], [795.76, 252.67], [821.36, 146.16], [793.45, 145.85], [769.59, 148.28], [795.76, 148.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[648.85, 201.89], [665.16, 201.72], [674.03, 202.34], [657.31, 202.52], [648.85, 142.92], [665.16, 143.22], [674.03, 142.13], [657.31, 141.81]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[280.41, 255.48], [323.02, 253.77], [325.28, 257.67], [280.41, 259.57], [280.41, 140.16], [323.02, 141.08], [325.28, 138.98], [280.41, 137.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[501.92, 208.85], [482.7, 209.18], [476.52, 208.27], [495.2, 207.96], [501.92, 152.77], [482.7, 152.45], [476.52, 153.34], [495.2, 153.64]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[523.6, 209.24], [499.22, 209.62], [490.98, 208.27], [514.34, 207.93], [523.6, 149.46], [499.22, 149.05], [490.98, 150.5], [514.34, 150.88]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[452.19, 217.81], [425.15, 218.77], [412.56, 217.51], [439.04, 216.61], [452.19, 141.01], [425.15, 139.99], [412.56, 141.32], [439.04, 142.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[430.55, 208.19], [408.72, 208.72], [393.38, 207.37], [414.46, 206.89], [430.55, 141.45], [408.72, 140.7], [393.38, 142.6], [414.46, 143.28]]}, {\"category\": \"van\", \"corners_3d\": [[-1.18, 382.32], [267.68, 360.51], [54.41, 901.94], [-1803.74, 1453.0], [-1.18, 121.52], [267.68, 127.89], [54.41, -30.36], [-1803.74, -191.43]]}]\n```" - } - ] - }, - { - "id": 1026, - "data_source": "Mono3DRefer", - "scene_name": "000626", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1034.87, 425.77], [1273.83, 425.0], [5303.73, 1939.52], [3686.36, 1978.29], [1034.87, 176.21], [1273.83, 176.2], [5303.73, 196.31], [3686.36, 196.82]]}, {\"category\": \"car\", \"corners_3d\": [[752.64, 240.22], [825.69, 240.22], [911.5, 267.04], [809.36, 267.03], [752.64, 173.31], [825.69, 173.31], [911.5, 173.49], [809.36, 173.49]]}, {\"category\": \"car\", \"corners_3d\": [[572.17, 212.67], [613.37, 212.81], [607.01, 219.15], [559.31, 218.96], [572.17, 180.57], [613.37, 180.6], [607.01, 181.82], [559.31, 181.79]]}, {\"category\": \"car\", \"corners_3d\": [[598.99, 204.15], [627.09, 204.27], [620.58, 207.96], [589.23, 207.8], [598.99, 178.0], [627.09, 178.02], [620.58, 178.63], [589.23, 178.6]]}]\n```" - } - ] - }, - { - "id": 1027, - "data_source": "Mono3DRefer", - "scene_name": "000626", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[710.35, 211.58], [762.0, 211.77], [794.7, 224.34], [726.27, 224.0], [710.35, 141.2], [762.0, 141.04], [794.7, 130.77], [726.27, 131.05]]}]\n```" - } - ] - }, - { - "id": 1028, - "data_source": "Mono3DRefer", - "scene_name": "000626", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1034.87, 425.77], [1273.83, 425.0], [5303.73, 1939.52], [3686.36, 1978.29], [1034.87, 176.21], [1273.83, 176.2], [5303.73, 196.31], [3686.36, 196.82]]}, {\"category\": \"car\", \"corners_3d\": [[752.64, 240.22], [825.69, 240.22], [911.5, 267.04], [809.36, 267.03], [752.64, 173.31], [825.69, 173.31], [911.5, 173.49], [809.36, 173.49]]}, {\"category\": \"car\", \"corners_3d\": [[572.17, 212.67], [613.37, 212.81], [607.01, 219.15], [559.31, 218.96], [572.17, 180.57], [613.37, 180.6], [607.01, 181.82], [559.31, 181.79]]}, {\"category\": \"van\", \"corners_3d\": [[710.35, 211.58], [762.0, 211.77], [794.7, 224.34], [726.27, 224.0], [710.35, 141.2], [762.0, 141.04], [794.7, 130.77], [726.27, 131.05]]}, {\"category\": \"car\", \"corners_3d\": [[598.99, 204.15], [627.09, 204.27], [620.58, 207.96], [589.23, 207.8], [598.99, 178.0], [627.09, 178.02], [620.58, 178.63], [589.23, 178.6]]}]\n```" - } - ] - }, - { - "id": 1029, - "data_source": "Mono3DRefer", - "scene_name": "000627", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[478.92, 243.3], [409.68, 244.23], [428.67, 230.34], [484.38, 229.73], [478.92, 185.11], [409.68, 185.27], [428.67, 182.86], [484.38, 182.75]]}, {\"category\": \"car\", \"corners_3d\": [[273.34, 435.55], [10.71, 441.21], [265.17, 318.16], [405.82, 316.48], [273.34, 177.41], [10.71, 177.51], [265.17, 175.37], [405.82, 175.34]]}, {\"category\": \"car\", \"corners_3d\": [[390.2, 210.11], [363.31, 210.71], [356.7, 208.48], [382.13, 207.94], [390.2, 185.91], [363.31, 186.12], [356.7, 185.33], [382.13, 185.15]]}]\n```" - } - ] - }, - { - "id": 1030, - "data_source": "Mono3DRefer", - "scene_name": "000629", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[440.56, 266.66], [357.51, 268.56], [385.96, 245.89], [449.21, 244.79], [440.56, 196.29], [357.51, 196.77], [385.96, 191.1], [449.21, 190.83]]}, {\"category\": \"car\", \"corners_3d\": [[528.83, 201.71], [547.97, 201.53], [553.65, 203.06], [533.51, 203.25], [528.83, 183.0], [547.97, 182.94], [553.65, 183.48], [533.51, 183.54]]}]\n```" - } - ] - }, - { - "id": 1031, - "data_source": "Mono3DRefer", - "scene_name": "000631", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[653.39, 204.8], [681.13, 204.74], [691.09, 207.71], [660.79, 207.78], [653.39, 177.47], [681.13, 177.46], [691.09, 177.89], [660.79, 177.9]]}, {\"category\": \"car\", \"corners_3d\": [[740.85, 264.05], [822.7, 263.64], [905.16, 294.65], [795.68, 295.39], [740.85, 187.3], [822.7, 187.23], [905.16, 192.14], [795.68, 192.26]]}, {\"category\": \"car\", \"corners_3d\": [[741.82, 207.01], [772.45, 206.93], [793.06, 210.47], [759.28, 210.56], [741.82, 177.04], [772.45, 177.03], [793.06, 177.46], [759.28, 177.47]]}, {\"category\": \"car\", \"corners_3d\": [[826.26, 226.47], [876.23, 226.17], [947.39, 237.88], [886.86, 238.33], [826.26, 178.8], [876.23, 178.77], [947.39, 180.07], [886.86, 180.12]]}]\n```" - } - ] - }, - { - "id": 1032, - "data_source": "Mono3DRefer", - "scene_name": "000633", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[251.94, 353.34], [280.32, 337.07], [401.78, 335.2], [386.23, 351.07], [251.94, 146.82], [280.32, 149.99], [401.78, 150.36], [386.23, 147.26]]}]\n```" - } - ] - }, - { - "id": 1033, - "data_source": "Mono3DRefer", - "scene_name": "000634", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[512.47, 225.44], [465.33, 225.41], [485.5, 218.38], [526.34, 218.41], [512.47, 176.52], [465.33, 176.52], [485.5, 176.03], [526.34, 176.03]]}, {\"category\": \"van\", \"corners_3d\": [[561.51, 194.6], [536.89, 194.62], [540.76, 192.95], [563.49, 192.93], [561.51, 167.82], [536.89, 167.81], [540.76, 168.2], [563.49, 168.2]]}]\n```" - } - ] - }, - { - "id": 1034, - "data_source": "Mono3DRefer", - "scene_name": "000634", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-1862.52, 1887.69], [-3344.01, 1815.3], [-25.36, 441.43], [231.36, 443.3], [-1862.52, 349.93], [-3344.01, 342.46], [-25.36, 200.59], [231.36, 200.78]]}, {\"category\": \"car\", \"corners_3d\": [[481.74, 251.9], [407.59, 251.66], [454.16, 235.21], [512.9, 235.36], [481.74, 190.79], [407.59, 190.73], [454.16, 187.0], [512.9, 187.04]]}, {\"category\": \"car\", \"corners_3d\": [[553.17, 198.48], [527.99, 198.5], [533.4, 196.42], [556.53, 196.4], [553.17, 174.89], [527.99, 174.89], [533.4, 174.72], [556.53, 174.72]]}]\n```" - } - ] - }, - { - "id": 1035, - "data_source": "Mono3DRefer", - "scene_name": "000634", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-1862.52, 1887.69], [-3344.01, 1815.3], [-25.36, 441.43], [231.36, 443.3], [-1862.52, 349.93], [-3344.01, 342.46], [-25.36, 200.59], [231.36, 200.78]]}, {\"category\": \"car\", \"corners_3d\": [[481.74, 251.9], [407.59, 251.66], [454.16, 235.21], [512.9, 235.36], [481.74, 190.79], [407.59, 190.73], [454.16, 187.0], [512.9, 187.04]]}, {\"category\": \"van\", \"corners_3d\": [[512.47, 225.44], [465.33, 225.41], [485.5, 218.38], [526.34, 218.41], [512.47, 176.52], [465.33, 176.52], [485.5, 176.03], [526.34, 176.03]]}, {\"category\": \"car\", \"corners_3d\": [[553.17, 198.48], [527.99, 198.5], [533.4, 196.42], [556.53, 196.4], [553.17, 174.89], [527.99, 174.89], [533.4, 174.72], [556.53, 174.72]]}, {\"category\": \"van\", \"corners_3d\": [[561.51, 194.6], [536.89, 194.62], [540.76, 192.95], [563.49, 192.93], [561.51, 167.82], [536.89, 167.81], [540.76, 168.2], [563.49, 168.2]]}]\n```" - } - ] - }, - { - "id": 1036, - "data_source": "Mono3DRefer", - "scene_name": "000636", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[585.17, 187.61], [603.91, 187.6], [604.86, 188.5], [584.97, 188.51], [585.17, 172.85], [603.91, 172.85], [604.86, 172.85], [584.97, 172.85]]}, {\"category\": \"car\", \"corners_3d\": [[562.42, 184.67], [578.49, 184.66], [577.97, 185.26], [561.09, 185.27], [562.42, 171.18], [578.49, 171.18], [577.97, 171.09], [561.09, 171.09]]}]\n```" - } - ] - }, - { - "id": 1037, - "data_source": "Mono3DRefer", - "scene_name": "000637", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[183.61, 241.69], [125.52, 241.04], [225.19, 230.33], [273.79, 230.77], [183.61, 181.64], [125.52, 181.63], [225.19, 181.43], [273.79, 181.44]]}, {\"category\": \"car\", \"corners_3d\": [[1034.73, 426.47], [1306.95, 437.18], [3834.04, 1488.38], [2190.67, 1251.01], [1034.73, 186.47], [1306.95, 186.73], [3834.04, 212.2], [2190.67, 206.44]]}, {\"category\": \"car\", \"corners_3d\": [[804.03, 290.59], [926.28, 290.6], [1122.25, 357.68], [925.48, 357.64], [804.03, 178.13], [926.28, 178.13], [1122.25, 176.67], [925.48, 176.67]]}, {\"category\": \"car\", \"corners_3d\": [[730.31, 245.86], [804.02, 246.2], [854.81, 266.55], [758.09, 265.96], [730.31, 178.6], [804.02, 178.59], [854.81, 177.99], [758.09, 178.01]]}, {\"category\": \"car\", \"corners_3d\": [[695.03, 228.36], [753.78, 228.52], [782.92, 240.1], [709.94, 239.85], [695.03, 178.44], [753.78, 178.43], [782.92, 177.93], [709.94, 177.94]]}, {\"category\": \"car\", \"corners_3d\": [[682.54, 214.31], [726.72, 214.42], [741.68, 220.27], [689.85, 220.12], [682.54, 176.36], [726.72, 176.35], [741.68, 175.63], [689.85, 175.65]]}, {\"category\": \"car\", \"corners_3d\": [[663.96, 207.29], [699.11, 207.33], [707.51, 210.41], [668.32, 210.36], [663.96, 175.37], [699.11, 175.36], [707.51, 174.77], [668.32, 174.78]]}, {\"category\": \"car\", \"corners_3d\": [[578.45, 188.42], [560.31, 188.41], [566.3, 187.86], [583.18, 187.87], [578.45, 171.84], [560.31, 171.85], [566.3, 172.46], [583.18, 172.45]]}]\n```" - } - ] - }, - { - "id": 1038, - "data_source": "Mono3DRefer", - "scene_name": "000639", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[463.01, 233.54], [444.4, 233.68], [451.88, 229.11], [469.07, 228.99], [463.01, 177.02], [444.4, 177.03], [451.88, 176.71], [469.07, 176.71]]}, {\"category\": \"cyclist\", \"corners_3d\": [[490.99, 226.73], [471.1, 226.97], [472.72, 222.9], [491.11, 222.69], [490.99, 176.69], [471.1, 176.71], [472.72, 176.42], [491.11, 176.4]]}]\n```" - } - ] - }, - { - "id": 1039, - "data_source": "Mono3DRefer", - "scene_name": "000639", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[957.54, 323.42], [813.39, 324.68], [735.38, 273.91], [832.03, 273.35], [957.54, 181.19], [813.39, 181.26], [735.38, 178.45], [832.03, 178.42]]}, {\"category\": \"car\", \"corners_3d\": [[646.15, 232.23], [702.52, 231.48], [743.31, 242.93], [676.38, 244.01], [646.15, 179.72], [702.52, 179.63], [743.31, 180.96], [676.38, 181.08]]}, {\"category\": \"car\", \"corners_3d\": [[387.76, 230.91], [442.53, 230.36], [429.04, 240.95], [363.91, 241.72], [387.76, 188.39], [442.53, 188.24], [429.04, 191.08], [363.91, 191.28]]}, {\"category\": \"car\", \"corners_3d\": [[603.99, 215.94], [644.94, 215.42], [674.2, 222.03], [627.22, 222.72], [603.99, 170.98], [644.94, 171.0], [674.2, 170.72], [627.22, 170.69]]}, {\"category\": \"car\", \"corners_3d\": [[558.55, 208.48], [589.67, 208.18], [601.97, 211.63], [567.89, 212.0], [558.55, 180.97], [589.67, 180.9], [601.97, 181.69], [567.89, 181.77]]}, {\"category\": \"car\", \"corners_3d\": [[839.52, 350.5], [1013.99, 350.54], [1792.19, 692.91], [1281.27, 692.62], [839.52, 155.27], [1013.99, 155.26], [1792.19, 121.36], [1281.27, 121.39]]}]\n```" - } - ] - }, - { - "id": 1040, - "data_source": "Mono3DRefer", - "scene_name": "000639", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[957.54, 323.42], [813.39, 324.68], [735.38, 273.91], [832.03, 273.35], [957.54, 181.19], [813.39, 181.26], [735.38, 178.45], [832.03, 178.42]]}, {\"category\": \"car\", \"corners_3d\": [[646.15, 232.23], [702.52, 231.48], [743.31, 242.93], [676.38, 244.01], [646.15, 179.72], [702.52, 179.63], [743.31, 180.96], [676.38, 181.08]]}, {\"category\": \"cyclist\", \"corners_3d\": [[463.01, 233.54], [444.4, 233.68], [451.88, 229.11], [469.07, 228.99], [463.01, 177.02], [444.4, 177.03], [451.88, 176.71], [469.07, 176.71]]}, {\"category\": \"car\", \"corners_3d\": [[387.76, 230.91], [442.53, 230.36], [429.04, 240.95], [363.91, 241.72], [387.76, 188.39], [442.53, 188.24], [429.04, 191.08], [363.91, 191.28]]}, {\"category\": \"car\", \"corners_3d\": [[603.99, 215.94], [644.94, 215.42], [674.2, 222.03], [627.22, 222.72], [603.99, 170.98], [644.94, 171.0], [674.2, 170.72], [627.22, 170.69]]}, {\"category\": \"car\", \"corners_3d\": [[558.55, 208.48], [589.67, 208.18], [601.97, 211.63], [567.89, 212.0], [558.55, 180.97], [589.67, 180.9], [601.97, 181.69], [567.89, 181.77]]}, {\"category\": \"cyclist\", \"corners_3d\": [[490.99, 226.73], [471.1, 226.97], [472.72, 222.9], [491.11, 222.69], [490.99, 176.69], [471.1, 176.71], [472.72, 176.42], [491.11, 176.4]]}, {\"category\": \"car\", \"corners_3d\": [[839.52, 350.5], [1013.99, 350.54], [1792.19, 692.91], [1281.27, 692.62], [839.52, 155.27], [1013.99, 155.26], [1792.19, 121.36], [1281.27, 121.39]]}]\n```" - } - ] - }, - { - "id": 1041, - "data_source": "Mono3DRefer", - "scene_name": "000640", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[395.86, 279.96], [499.99, 280.26], [432.36, 338.16], [267.79, 337.41], [395.86, 183.39], [499.99, 183.39], [432.36, 184.49], [267.79, 184.48]]}, {\"category\": \"car\", \"corners_3d\": [[493.87, 225.84], [547.54, 225.94], [532.85, 234.65], [468.74, 234.5], [493.87, 176.57], [547.54, 176.56], [532.85, 175.59], [468.74, 175.6]]}, {\"category\": \"car\", \"corners_3d\": [[597.7, 205.84], [638.0, 205.85], [643.15, 210.05], [595.9, 210.03], [597.7, 169.17], [638.0, 169.17], [643.15, 167.04], [595.9, 167.05]]}, {\"category\": \"car\", \"corners_3d\": [[560.3, 196.9], [591.69, 196.92], [588.25, 198.64], [553.36, 198.61], [560.3, 166.07], [591.69, 166.05], [588.25, 164.32], [553.36, 164.35]]}]\n```" - } - ] - }, - { - "id": 1042, - "data_source": "Mono3DRefer", - "scene_name": "000646", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[68.91, 296.39], [102.22, 290.83], [172.41, 293.77], [141.81, 299.64], [68.91, 157.65], [102.22, 158.74], [172.41, 158.16], [141.81, 157.01]]}]\n```" - } - ] - }, - { - "id": 1043, - "data_source": "Mono3DRefer", - "scene_name": "000648", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[543.14, 225.17], [495.27, 225.13], [512.0, 217.9], [553.25, 217.93], [543.14, 185.17], [495.27, 185.16], [512.0, 183.46], [553.25, 183.47]]}, {\"category\": \"car\", \"corners_3d\": [[101.98, 492.71], [-175.61, 497.6], [192.58, 339.38], [333.12, 338.09], [101.98, 242.89], [-175.61, 243.95], [192.58, 209.31], [333.12, 209.03]]}]\n```" - } - ] - }, - { - "id": 1044, - "data_source": "Mono3DRefer", - "scene_name": "000648", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[398.67, 305.55], [274.28, 305.8], [371.03, 266.52], [458.59, 266.4], [398.67, 170.59], [274.28, 170.59], [371.03, 171.26], [458.59, 171.26]]}]\n```" - } - ] - }, - { - "id": 1045, - "data_source": "Mono3DRefer", - "scene_name": "000648", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[543.14, 225.17], [495.27, 225.13], [512.0, 217.9], [553.25, 217.93], [543.14, 185.17], [495.27, 185.16], [512.0, 183.46], [553.25, 183.47]]}, {\"category\": \"car\", \"corners_3d\": [[101.98, 492.71], [-175.61, 497.6], [192.58, 339.38], [333.12, 338.09], [101.98, 242.89], [-175.61, 243.95], [192.58, 209.31], [333.12, 209.03]]}, {\"category\": \"van\", \"corners_3d\": [[398.67, 305.55], [274.28, 305.8], [371.03, 266.52], [458.59, 266.4], [398.67, 170.59], [274.28, 170.59], [371.03, 171.26], [458.59, 171.26]]}]\n```" - } - ] - }, - { - "id": 1046, - "data_source": "Mono3DRefer", - "scene_name": "000649", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[690.82, 184.5], [709.83, 184.51], [731.88, 188.55], [706.24, 188.53], [690.82, 157.94], [709.83, 157.92], [731.88, 152.74], [706.24, 152.77]]}]\n```" - } - ] - }, - { - "id": 1047, - "data_source": "Mono3DRefer", - "scene_name": "000649", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-12.58, 317.25], [-134.0, 316.77], [94.3, 273.39], [179.56, 273.63], [-12.58, 205.51], [-134.0, 205.4], [94.3, 195.59], [179.56, 195.65]]}, {\"category\": \"car\", \"corners_3d\": [[257.58, 257.06], [187.39, 256.9], [269.96, 240.99], [326.96, 241.09], [257.58, 177.69], [187.39, 177.68], [269.96, 176.77], [326.96, 176.77]]}, {\"category\": \"car\", \"corners_3d\": [[341.75, 238.0], [295.89, 237.53], [348.71, 228.83], [388.7, 229.18], [341.75, 189.55], [295.89, 189.43], [348.71, 187.2], [388.7, 187.29]]}, {\"category\": \"car\", \"corners_3d\": [[397.58, 228.06], [350.84, 227.92], [391.53, 220.11], [431.72, 220.21], [397.58, 179.5], [350.84, 179.49], [391.53, 178.55], [431.72, 178.56]]}, {\"category\": \"car\", \"corners_3d\": [[466.75, 214.2], [433.7, 214.11], [454.95, 209.96], [484.71, 210.04], [466.75, 183.34], [433.7, 183.31], [454.95, 182.26], [484.71, 182.28]]}, {\"category\": \"car\", \"corners_3d\": [[489.94, 208.22], [462.17, 208.14], [478.63, 205.09], [504.02, 205.15], [489.94, 182.18], [462.17, 182.15], [478.63, 181.35], [504.02, 181.37]]}]\n```" - } - ] - }, - { - "id": 1048, - "data_source": "Mono3DRefer", - "scene_name": "000649", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-12.58, 317.25], [-134.0, 316.77], [94.3, 273.39], [179.56, 273.63], [-12.58, 205.51], [-134.0, 205.4], [94.3, 195.59], [179.56, 195.65]]}, {\"category\": \"tram\", \"corners_3d\": [[690.82, 184.5], [709.83, 184.51], [731.88, 188.55], [706.24, 188.53], [690.82, 157.94], [709.83, 157.92], [731.88, 152.74], [706.24, 152.77]]}, {\"category\": \"car\", \"corners_3d\": [[257.58, 257.06], [187.39, 256.9], [269.96, 240.99], [326.96, 241.09], [257.58, 177.69], [187.39, 177.68], [269.96, 176.77], [326.96, 176.77]]}, {\"category\": \"car\", \"corners_3d\": [[341.75, 238.0], [295.89, 237.53], [348.71, 228.83], [388.7, 229.18], [341.75, 189.55], [295.89, 189.43], [348.71, 187.2], [388.7, 187.29]]}, {\"category\": \"car\", \"corners_3d\": [[397.58, 228.06], [350.84, 227.92], [391.53, 220.11], [431.72, 220.21], [397.58, 179.5], [350.84, 179.49], [391.53, 178.55], [431.72, 178.56]]}, {\"category\": \"car\", \"corners_3d\": [[466.75, 214.2], [433.7, 214.11], [454.95, 209.96], [484.71, 210.04], [466.75, 183.34], [433.7, 183.31], [454.95, 182.26], [484.71, 182.28]]}, {\"category\": \"car\", \"corners_3d\": [[489.94, 208.22], [462.17, 208.14], [478.63, 205.09], [504.02, 205.15], [489.94, 182.18], [462.17, 182.15], [478.63, 181.35], [504.02, 181.37]]}]\n```" - } - ] - }, - { - "id": 1049, - "data_source": "Mono3DRefer", - "scene_name": "000651", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[586.93, 217.33], [633.92, 217.33], [637.48, 223.99], [583.46, 223.99], [586.93, 179.52], [633.92, 179.52], [637.48, 180.51], [583.46, 180.51]]}, {\"category\": \"car\", \"corners_3d\": [[114.11, 282.4], [224.06, 269.71], [393.88, 291.32], [283.37, 310.87], [114.11, 183.93], [224.06, 182.64], [393.88, 184.83], [283.37, 186.8]]}, {\"category\": \"car\", \"corners_3d\": [[366.03, 227.83], [415.46, 224.46], [509.97, 231.11], [460.72, 235.45], [366.03, 167.34], [415.46, 167.68], [509.97, 167.01], [460.72, 166.58]]}, {\"category\": \"car\", \"corners_3d\": [[411.72, 219.63], [454.13, 217.59], [515.24, 222.89], [470.68, 225.45], [411.72, 165.78], [454.13, 166.09], [515.24, 165.29], [470.68, 164.9]]}, {\"category\": \"car\", \"corners_3d\": [[468.06, 205.85], [494.24, 204.72], [547.51, 207.07], [521.35, 208.38], [468.06, 168.49], [494.24, 168.64], [547.51, 168.33], [521.35, 168.15]]}, {\"category\": \"car\", \"corners_3d\": [[589.25, 202.7], [620.21, 202.71], [620.5, 205.77], [586.37, 205.75], [589.25, 174.52], [620.21, 174.52], [620.5, 174.69], [586.37, 174.69]]}]\n```" - } - ] - }, - { - "id": 1050, - "data_source": "Mono3DRefer", - "scene_name": "000651", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[189.97, 224.63], [113.12, 224.69], [232.34, 212.05], [290.35, 212.01], [189.97, 132.22], [113.12, 132.17], [232.34, 142.1], [290.35, 142.12]]}]\n```" - } - ] - }, - { - "id": 1051, - "data_source": "Mono3DRefer", - "scene_name": "000651", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[647.68, 200.7], [684.74, 200.68], [699.49, 205.29], [656.3, 205.32], [647.68, 158.21], [684.74, 158.22], [699.49, 155.79], [656.3, 155.78]]}, {\"category\": \"van\", \"corners_3d\": [[233.57, 258.48], [313.29, 250.35], [472.35, 265.28], [395.61, 277.08], [233.57, 158.83], [313.29, 160.16], [472.35, 157.72], [395.61, 155.78]]}]\n```" - } - ] - }, - { - "id": 1052, - "data_source": "Mono3DRefer", - "scene_name": "000651", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[586.93, 217.33], [633.92, 217.33], [637.48, 223.99], [583.46, 223.99], [586.93, 179.52], [633.92, 179.52], [637.48, 180.51], [583.46, 180.51]]}, {\"category\": \"car\", \"corners_3d\": [[114.11, 282.4], [224.06, 269.71], [393.88, 291.32], [283.37, 310.87], [114.11, 183.93], [224.06, 182.64], [393.88, 184.83], [283.37, 186.8]]}, {\"category\": \"car\", \"corners_3d\": [[366.03, 227.83], [415.46, 224.46], [509.97, 231.11], [460.72, 235.45], [366.03, 167.34], [415.46, 167.68], [509.97, 167.01], [460.72, 166.58]]}, {\"category\": \"car\", \"corners_3d\": [[411.72, 219.63], [454.13, 217.59], [515.24, 222.89], [470.68, 225.45], [411.72, 165.78], [454.13, 166.09], [515.24, 165.29], [470.68, 164.9]]}, {\"category\": \"car\", \"corners_3d\": [[468.06, 205.85], [494.24, 204.72], [547.51, 207.07], [521.35, 208.38], [468.06, 168.49], [494.24, 168.64], [547.51, 168.33], [521.35, 168.15]]}, {\"category\": \"car\", \"corners_3d\": [[589.25, 202.7], [620.21, 202.71], [620.5, 205.77], [586.37, 205.75], [589.25, 174.52], [620.21, 174.52], [620.5, 174.69], [586.37, 174.69]]}, {\"category\": \"truck\", \"corners_3d\": [[189.97, 224.63], [113.12, 224.69], [232.34, 212.05], [290.35, 212.01], [189.97, 132.22], [113.12, 132.17], [232.34, 142.1], [290.35, 142.12]]}]\n```" - } - ] - }, - { - "id": 1053, - "data_source": "Mono3DRefer", - "scene_name": "000651", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[586.93, 217.33], [633.92, 217.33], [637.48, 223.99], [583.46, 223.99], [586.93, 179.52], [633.92, 179.52], [637.48, 180.51], [583.46, 180.51]]}, {\"category\": \"car\", \"corners_3d\": [[114.11, 282.4], [224.06, 269.71], [393.88, 291.32], [283.37, 310.87], [114.11, 183.93], [224.06, 182.64], [393.88, 184.83], [283.37, 186.8]]}, {\"category\": \"car\", \"corners_3d\": [[366.03, 227.83], [415.46, 224.46], [509.97, 231.11], [460.72, 235.45], [366.03, 167.34], [415.46, 167.68], [509.97, 167.01], [460.72, 166.58]]}, {\"category\": \"car\", \"corners_3d\": [[411.72, 219.63], [454.13, 217.59], [515.24, 222.89], [470.68, 225.45], [411.72, 165.78], [454.13, 166.09], [515.24, 165.29], [470.68, 164.9]]}, {\"category\": \"car\", \"corners_3d\": [[468.06, 205.85], [494.24, 204.72], [547.51, 207.07], [521.35, 208.38], [468.06, 168.49], [494.24, 168.64], [547.51, 168.33], [521.35, 168.15]]}, {\"category\": \"van\", \"corners_3d\": [[647.68, 200.7], [684.74, 200.68], [699.49, 205.29], [656.3, 205.32], [647.68, 158.21], [684.74, 158.22], [699.49, 155.79], [656.3, 155.78]]}, {\"category\": \"car\", \"corners_3d\": [[589.25, 202.7], [620.21, 202.71], [620.5, 205.77], [586.37, 205.75], [589.25, 174.52], [620.21, 174.52], [620.5, 174.69], [586.37, 174.69]]}, {\"category\": \"van\", \"corners_3d\": [[233.57, 258.48], [313.29, 250.35], [472.35, 265.28], [395.61, 277.08], [233.57, 158.83], [313.29, 160.16], [472.35, 157.72], [395.61, 155.78]]}]\n```" - } - ] - }, - { - "id": 1054, - "data_source": "Mono3DRefer", - "scene_name": "000651", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[647.68, 200.7], [684.74, 200.68], [699.49, 205.29], [656.3, 205.32], [647.68, 158.21], [684.74, 158.22], [699.49, 155.79], [656.3, 155.78]]}, {\"category\": \"van\", \"corners_3d\": [[233.57, 258.48], [313.29, 250.35], [472.35, 265.28], [395.61, 277.08], [233.57, 158.83], [313.29, 160.16], [472.35, 157.72], [395.61, 155.78]]}, {\"category\": \"truck\", \"corners_3d\": [[189.97, 224.63], [113.12, 224.69], [232.34, 212.05], [290.35, 212.01], [189.97, 132.22], [113.12, 132.17], [232.34, 142.1], [290.35, 142.12]]}]\n```" - } - ] - }, - { - "id": 1055, - "data_source": "Mono3DRefer", - "scene_name": "000651", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[586.93, 217.33], [633.92, 217.33], [637.48, 223.99], [583.46, 223.99], [586.93, 179.52], [633.92, 179.52], [637.48, 180.51], [583.46, 180.51]]}, {\"category\": \"car\", \"corners_3d\": [[114.11, 282.4], [224.06, 269.71], [393.88, 291.32], [283.37, 310.87], [114.11, 183.93], [224.06, 182.64], [393.88, 184.83], [283.37, 186.8]]}, {\"category\": \"car\", \"corners_3d\": [[366.03, 227.83], [415.46, 224.46], [509.97, 231.11], [460.72, 235.45], [366.03, 167.34], [415.46, 167.68], [509.97, 167.01], [460.72, 166.58]]}, {\"category\": \"car\", \"corners_3d\": [[411.72, 219.63], [454.13, 217.59], [515.24, 222.89], [470.68, 225.45], [411.72, 165.78], [454.13, 166.09], [515.24, 165.29], [470.68, 164.9]]}, {\"category\": \"car\", \"corners_3d\": [[468.06, 205.85], [494.24, 204.72], [547.51, 207.07], [521.35, 208.38], [468.06, 168.49], [494.24, 168.64], [547.51, 168.33], [521.35, 168.15]]}, {\"category\": \"van\", \"corners_3d\": [[647.68, 200.7], [684.74, 200.68], [699.49, 205.29], [656.3, 205.32], [647.68, 158.21], [684.74, 158.22], [699.49, 155.79], [656.3, 155.78]]}, {\"category\": \"car\", \"corners_3d\": [[589.25, 202.7], [620.21, 202.71], [620.5, 205.77], [586.37, 205.75], [589.25, 174.52], [620.21, 174.52], [620.5, 174.69], [586.37, 174.69]]}, {\"category\": \"van\", \"corners_3d\": [[233.57, 258.48], [313.29, 250.35], [472.35, 265.28], [395.61, 277.08], [233.57, 158.83], [313.29, 160.16], [472.35, 157.72], [395.61, 155.78]]}, {\"category\": \"truck\", \"corners_3d\": [[189.97, 224.63], [113.12, 224.69], [232.34, 212.05], [290.35, 212.01], [189.97, 132.22], [113.12, 132.17], [232.34, 142.1], [290.35, 142.12]]}]\n```" - } - ] - }, - { - "id": 1056, - "data_source": "Mono3DRefer", - "scene_name": "000656", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[570.57, 189.3], [593.52, 189.31], [589.69, 190.0], [564.72, 189.99], [570.57, 168.04], [593.52, 168.02], [589.69, 166.83], [564.72, 166.85]]}, {\"category\": \"car\", \"corners_3d\": [[618.23, 183.96], [638.82, 183.96], [638.39, 184.16], [616.14, 184.15], [618.23, 165.24], [638.82, 165.22], [638.39, 163.89], [616.14, 163.93]]}]\n```" - } - ] - }, - { - "id": 1057, - "data_source": "Mono3DRefer", - "scene_name": "000657", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[588.28, 189.01], [607.86, 189.01], [607.71, 190.05], [586.87, 190.05], [588.28, 173.58], [607.86, 173.58], [607.71, 173.63], [586.87, 173.63]]}, {\"category\": \"car\", \"corners_3d\": [[564.34, 187.46], [581.11, 187.45], [580.72, 188.23], [563.04, 188.24], [564.34, 173.36], [581.11, 173.36], [580.72, 173.39], [563.04, 173.39]]}]\n```" - } - ] - }, - { - "id": 1058, - "data_source": "Mono3DRefer", - "scene_name": "000659", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[228.55, 212.35], [217.48, 212.26], [237.64, 210.63], [248.29, 210.7], [228.55, 178.07], [217.48, 178.06], [237.64, 177.85], [248.29, 177.86]]}]\n```" - } - ] - }, - { - "id": 1059, - "data_source": "Mono3DRefer", - "scene_name": "000659", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[216.03, 226.03], [171.99, 225.9], [222.86, 220.11], [262.19, 220.21], [216.03, 188.41], [171.99, 188.37], [222.86, 186.68], [262.19, 186.71]]}]\n```" - } - ] - }, - { - "id": 1060, - "data_source": "Mono3DRefer", - "scene_name": "000659", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[316.49, 204.59], [286.88, 204.56], [319.54, 201.48], [346.3, 201.5], [316.49, 170.67], [286.88, 170.67], [319.54, 170.88], [346.3, 170.88]]}]\n```" - } - ] - }, - { - "id": 1061, - "data_source": "Mono3DRefer", - "scene_name": "000659", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[228.55, 212.35], [217.48, 212.26], [237.64, 210.63], [248.29, 210.7], [228.55, 178.07], [217.48, 178.06], [237.64, 177.85], [248.29, 177.86]]}, {\"category\": \"car\", \"corners_3d\": [[216.03, 226.03], [171.99, 225.9], [222.86, 220.11], [262.19, 220.21], [216.03, 188.41], [171.99, 188.37], [222.86, 186.68], [262.19, 186.71]]}]\n```" - } - ] - }, - { - "id": 1062, - "data_source": "Mono3DRefer", - "scene_name": "000659", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[228.55, 212.35], [217.48, 212.26], [237.64, 210.63], [248.29, 210.7], [228.55, 178.07], [217.48, 178.06], [237.64, 177.85], [248.29, 177.86]]}, {\"category\": \"van\", \"corners_3d\": [[316.49, 204.59], [286.88, 204.56], [319.54, 201.48], [346.3, 201.5], [316.49, 170.67], [286.88, 170.67], [319.54, 170.88], [346.3, 170.88]]}]\n```" - } - ] - }, - { - "id": 1063, - "data_source": "Mono3DRefer", - "scene_name": "000659", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[216.03, 226.03], [171.99, 225.9], [222.86, 220.11], [262.19, 220.21], [216.03, 188.41], [171.99, 188.37], [222.86, 186.68], [262.19, 186.71]]}, {\"category\": \"van\", \"corners_3d\": [[316.49, 204.59], [286.88, 204.56], [319.54, 201.48], [346.3, 201.5], [316.49, 170.67], [286.88, 170.67], [319.54, 170.88], [346.3, 170.88]]}]\n```" - } - ] - }, - { - "id": 1064, - "data_source": "Mono3DRefer", - "scene_name": "000659", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[228.55, 212.35], [217.48, 212.26], [237.64, 210.63], [248.29, 210.7], [228.55, 178.07], [217.48, 178.06], [237.64, 177.85], [248.29, 177.86]]}, {\"category\": \"car\", \"corners_3d\": [[216.03, 226.03], [171.99, 225.9], [222.86, 220.11], [262.19, 220.21], [216.03, 188.41], [171.99, 188.37], [222.86, 186.68], [262.19, 186.71]]}, {\"category\": \"van\", \"corners_3d\": [[316.49, 204.59], [286.88, 204.56], [319.54, 201.48], [346.3, 201.5], [316.49, 170.67], [286.88, 170.67], [319.54, 170.88], [346.3, 170.88]]}]\n```" - } - ] - }, - { - "id": 1065, - "data_source": "Mono3DRefer", - "scene_name": "000661", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[726.77, 239.83], [785.0, 239.73], [826.18, 254.28], [755.34, 254.43], [726.77, 182.53], [785.0, 182.51], [826.18, 184.61], [755.34, 184.63]]}, {\"category\": \"car\", \"corners_3d\": [[895.71, 239.97], [956.77, 239.92], [1037.9, 255.3], [962.91, 255.38], [895.71, 180.71], [956.77, 180.71], [1037.9, 182.51], [962.91, 182.52]]}]\n```" - } - ] - }, - { - "id": 1066, - "data_source": "Mono3DRefer", - "scene_name": "000662", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[418.68, 229.96], [381.11, 230.86], [380.15, 225.94], [414.59, 225.18], [418.68, 196.18], [381.11, 196.55], [380.15, 194.54], [414.59, 194.23]]}, {\"category\": \"car\", \"corners_3d\": [[-425.49, 1196.11], [-1570.99, 1273.09], [128.19, 401.07], [353.63, 397.56], [-425.49, 336.4], [-1570.99, 348.7], [128.19, 209.33], [353.63, 208.77]]}]\n```" - } - ] - }, - { - "id": 1067, - "data_source": "Mono3DRefer", - "scene_name": "000663", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[920.51, 298.94], [841.99, 294.04], [845.35, 288.38], [920.19, 292.82], [920.51, 159.26], [841.99, 159.78], [845.35, 160.39], [920.19, 159.92]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[811.93, 283.77], [761.38, 281.56], [764.43, 276.87], [812.82, 278.89], [811.93, 178.96], [761.38, 178.83], [764.43, 178.58], [812.82, 178.69]]}]\n```" - } - ] - }, - { - "id": 1068, - "data_source": "Mono3DRefer", - "scene_name": "000664", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[363.44, 268.9], [454.45, 268.91], [398.45, 302.31], [272.68, 302.29], [363.44, 185.56], [454.45, 185.56], [398.45, 187.11], [272.68, 187.11]]}, {\"category\": \"car\", \"corners_3d\": [[542.07, 218.42], [591.04, 218.34], [593.56, 226.34], [533.95, 226.45], [542.07, 173.84], [591.04, 173.86], [593.56, 172.2], [533.95, 172.18]]}, {\"category\": \"car\", \"corners_3d\": [[527.74, 205.94], [563.05, 205.96], [556.38, 209.08], [516.58, 209.05], [527.74, 171.25], [563.05, 171.24], [556.38, 169.93], [516.58, 169.94]]}, {\"category\": \"car\", \"corners_3d\": [[444.18, 210.63], [409.0, 210.59], [430.69, 207.62], [462.3, 207.65], [444.18, 176.86], [409.0, 176.87], [430.69, 177.34], [462.3, 177.34]]}, {\"category\": \"car\", \"corners_3d\": [[590.34, 200.41], [619.63, 200.41], [620.86, 201.65], [589.64, 201.65], [590.34, 170.92], [619.63, 170.92], [620.86, 170.23], [589.64, 170.23]]}]\n```" - } - ] - }, - { - "id": 1069, - "data_source": "Mono3DRefer", - "scene_name": "000668", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[453.78, 270.14], [459.29, 258.39], [698.12, 254.2], [723.92, 264.75], [453.78, 163.93], [459.29, 165.01], [698.12, 165.39], [723.92, 164.43]]}, {\"category\": \"car\", \"corners_3d\": [[23.14, 203.48], [29.13, 204.92], [-124.1, 208.67], [-123.11, 206.88], [23.14, 164.73], [29.13, 164.35], [-124.1, 163.35], [-123.11, 163.83]]}, {\"category\": \"car\", \"corners_3d\": [[1214.06, 265.65], [1349.94, 274.5], [1380.04, 303.21], [1206.67, 289.01], [1214.06, 175.14], [1349.94, 175.36], [1380.04, 176.06], [1206.67, 175.71]]}, {\"category\": \"car\", \"corners_3d\": [[901.85, 234.53], [837.93, 230.73], [898.49, 223.27], [957.14, 226.13], [901.85, 177.25], [837.93, 176.98], [898.49, 176.44], [957.14, 176.65]]}, {\"category\": \"car\", \"corners_3d\": [[1032.0, 207.4], [983.68, 205.9], [1016.43, 203.07], [1061.8, 204.32], [1032.0, 165.83], [983.68, 166.13], [1016.43, 166.71], [1061.8, 166.45]]}]\n```" - } - ] - }, - { - "id": 1070, - "data_source": "Mono3DRefer", - "scene_name": "000671", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[786.39, 306.71], [748.86, 301.52], [791.92, 291.87], [828.04, 296.25], [786.39, 136.75], [748.86, 138.55], [791.92, 141.89], [828.04, 140.37]]}]\n```" - } - ] - }, - { - "id": 1071, - "data_source": "Mono3DRefer", - "scene_name": "000672", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[132.64, 565.31], [-237.99, 563.45], [230.05, 349.43], [398.17, 349.81], [132.64, 208.46], [-237.99, 208.29], [230.05, 188.87], [398.17, 188.91]]}, {\"category\": \"car\", \"corners_3d\": [[562.39, 216.01], [525.07, 216.01], [534.59, 211.11], [567.68, 211.11], [562.39, 184.09], [525.07, 184.09], [534.59, 182.81], [567.68, 182.81]]}]\n```" - } - ] - }, - { - "id": 1072, - "data_source": "Mono3DRefer", - "scene_name": "000673", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[365.91, 305.23], [243.11, 304.79], [352.25, 266.93], [439.98, 267.15], [365.91, 184.24], [243.11, 184.2], [352.25, 180.95], [439.98, 180.97]]}, {\"category\": \"car\", \"corners_3d\": [[660.92, 199.81], [691.85, 199.89], [694.66, 202.32], [660.94, 202.23], [660.92, 170.5], [691.85, 170.49], [694.66, 170.28], [660.94, 170.28]]}]\n```" - } - ] - }, - { - "id": 1073, - "data_source": "Mono3DRefer", - "scene_name": "000674", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[706.31, 242.25], [768.86, 241.64], [825.07, 258.37], [747.76, 259.31], [706.31, 176.69], [768.86, 176.66], [825.07, 177.59], [747.76, 177.64]]}, {\"category\": \"car\", \"corners_3d\": [[426.89, 244.77], [373.13, 244.83], [408.48, 233.7], [453.91, 233.66], [426.89, 190.07], [373.13, 190.09], [408.48, 187.42], [453.91, 187.41]]}, {\"category\": \"car\", \"corners_3d\": [[693.59, 223.35], [740.06, 223.29], [766.17, 232.41], [711.32, 232.5], [693.59, 178.63], [740.06, 178.63], [766.17, 179.67], [711.32, 179.68]]}, {\"category\": \"car\", \"corners_3d\": [[358.81, 271.71], [276.49, 271.49], [347.0, 251.44], [412.68, 251.58], [358.81, 192.2], [276.49, 192.16], [347.0, 188.24], [412.68, 188.26]]}, {\"category\": \"car\", \"corners_3d\": [[747.45, 271.99], [845.87, 271.45], [961.49, 314.54], [820.63, 315.64], [747.45, 177.82], [845.87, 177.79], [961.49, 179.95], [820.63, 180.01]]}]\n```" - } - ] - }, - { - "id": 1074, - "data_source": "Mono3DRefer", - "scene_name": "000674", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[749.57, 215.7], [769.53, 215.7], [774.89, 217.14], [754.26, 217.14], [749.57, 158.17], [769.53, 158.17], [774.89, 157.67], [754.26, 157.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[769.57, 215.04], [788.89, 215.04], [794.51, 216.36], [774.58, 216.36], [769.57, 157.39], [788.89, 157.39], [794.51, 156.91], [774.58, 156.91]]}]\n```" - } - ] - }, - { - "id": 1075, - "data_source": "Mono3DRefer", - "scene_name": "000674", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[706.31, 242.25], [768.86, 241.64], [825.07, 258.37], [747.76, 259.31], [706.31, 176.69], [768.86, 176.66], [825.07, 177.59], [747.76, 177.64]]}, {\"category\": \"car\", \"corners_3d\": [[426.89, 244.77], [373.13, 244.83], [408.48, 233.7], [453.91, 233.66], [426.89, 190.07], [373.13, 190.09], [408.48, 187.42], [453.91, 187.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[749.57, 215.7], [769.53, 215.7], [774.89, 217.14], [754.26, 217.14], [749.57, 158.17], [769.53, 158.17], [774.89, 157.67], [754.26, 157.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[769.57, 215.04], [788.89, 215.04], [794.51, 216.36], [774.58, 216.36], [769.57, 157.39], [788.89, 157.39], [794.51, 156.91], [774.58, 156.91]]}, {\"category\": \"car\", \"corners_3d\": [[693.59, 223.35], [740.06, 223.29], [766.17, 232.41], [711.32, 232.5], [693.59, 178.63], [740.06, 178.63], [766.17, 179.67], [711.32, 179.68]]}, {\"category\": \"car\", \"corners_3d\": [[358.81, 271.71], [276.49, 271.49], [347.0, 251.44], [412.68, 251.58], [358.81, 192.2], [276.49, 192.16], [347.0, 188.24], [412.68, 188.26]]}, {\"category\": \"car\", \"corners_3d\": [[747.45, 271.99], [845.87, 271.45], [961.49, 314.54], [820.63, 315.64], [747.45, 177.82], [845.87, 177.79], [961.49, 179.95], [820.63, 180.01]]}]\n```" - } - ] - }, - { - "id": 1076, - "data_source": "Mono3DRefer", - "scene_name": "000675", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[454.73, 270.36], [492.16, 268.8], [499.62, 274.7], [459.97, 276.46], [454.73, 176.51], [492.16, 176.45], [499.62, 176.68], [459.97, 176.74]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[426.32, 270.08], [460.47, 269.11], [461.38, 276.6], [424.55, 277.73], [426.32, 170.35], [460.47, 170.38], [461.38, 170.19], [424.55, 170.16]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[728.91, 236.07], [706.96, 236.63], [693.31, 233.7], [714.37, 233.19], [728.91, 162.75], [706.96, 162.67], [693.31, 163.13], [714.37, 163.21]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[762.9, 235.68], [738.86, 236.43], [724.67, 233.98], [747.95, 233.29], [762.9, 161.13], [738.86, 160.99], [724.67, 161.45], [747.95, 161.58]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[424.53, 237.19], [421.43, 238.35], [389.8, 238.41], [393.47, 237.24], [424.53, 168.55], [421.43, 168.47], [389.8, 168.47], [393.47, 168.55]]}]\n```" - } - ] - }, - { - "id": 1077, - "data_source": "Mono3DRefer", - "scene_name": "000678", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[385.15, 217.41], [354.53, 218.17], [349.39, 215.13], [378.07, 214.47], [385.15, 190.08], [354.53, 190.37], [349.39, 189.19], [378.07, 188.94]]}, {\"category\": \"car\", \"corners_3d\": [[443.59, 290.16], [331.23, 292.14], [387.02, 258.54], [467.44, 257.52], [443.59, 196.43], [331.23, 196.82], [387.02, 190.07], [467.44, 189.87]]}]\n```" - } - ] - }, - { - "id": 1078, - "data_source": "Mono3DRefer", - "scene_name": "000681", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[507.25, 206.75], [500.59, 208.16], [422.04, 207.85], [431.76, 206.46], [507.25, 183.34], [500.59, 183.78], [422.04, 183.68], [431.76, 183.26]]}]\n```" - } - ] - }, - { - "id": 1079, - "data_source": "Mono3DRefer", - "scene_name": "000682", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-0.05, 533.12], [-387.52, 538.93], [173.15, 328.39], [335.47, 327.33], [-0.05, 188.84], [-387.52, 189.1], [173.15, 179.75], [335.47, 179.71]]}, {\"category\": \"car\", \"corners_3d\": [[646.46, 208.03], [668.79, 208.03], [674.26, 211.32], [649.85, 211.32], [646.46, 188.11], [668.79, 188.11], [674.26, 189.53], [649.85, 189.53]]}]\n```" - } - ] - }, - { - "id": 1080, - "data_source": "Mono3DRefer", - "scene_name": "000688", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[98.26, 228.51], [53.36, 228.55], [123.27, 221.45], [162.41, 221.42], [98.26, 185.61], [53.36, 185.62], [123.27, 183.99], [162.41, 183.98]]}, {\"category\": \"car\", \"corners_3d\": [[183.9, 218.94], [143.6, 218.86], [199.87, 213.55], [235.6, 213.61], [183.9, 186.47], [143.6, 186.45], [199.87, 184.88], [235.6, 184.89]]}, {\"category\": \"car\", \"corners_3d\": [[264.6, 211.84], [234.19, 211.81], [272.69, 207.95], [300.11, 207.98], [264.6, 184.8], [234.19, 184.79], [272.69, 183.61], [300.11, 183.62]]}, {\"category\": \"car\", \"corners_3d\": [[312.28, 205.79], [289.56, 205.75], [314.67, 203.33], [335.73, 203.36], [312.28, 182.12], [289.56, 182.11], [314.67, 181.43], [335.73, 181.44]]}, {\"category\": \"car\", \"corners_3d\": [[347.62, 203.85], [324.11, 203.8], [344.19, 201.86], [366.26, 201.91], [347.62, 183.56], [324.11, 183.55], [344.19, 182.88], [366.26, 182.89]]}, {\"category\": \"car\", \"corners_3d\": [[396.49, 198.03], [379.75, 197.98], [394.91, 196.68], [410.82, 196.73], [396.49, 179.56], [379.75, 179.54], [394.91, 179.2], [410.82, 179.21]]}]\n```" - } - ] - }, - { - "id": 1081, - "data_source": "Mono3DRefer", - "scene_name": "000688", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[644.77, 199.98], [689.85, 200.14], [694.27, 208.41], [635.59, 208.15], [644.77, 145.41], [689.85, 145.25], [694.27, 136.88], [635.59, 137.14]]}]\n```" - } - ] - }, - { - "id": 1082, - "data_source": "Mono3DRefer", - "scene_name": "000688", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[644.77, 199.98], [689.85, 200.14], [694.27, 208.41], [635.59, 208.15], [644.77, 145.41], [689.85, 145.25], [694.27, 136.88], [635.59, 137.14]]}, {\"category\": \"car\", \"corners_3d\": [[98.26, 228.51], [53.36, 228.55], [123.27, 221.45], [162.41, 221.42], [98.26, 185.61], [53.36, 185.62], [123.27, 183.99], [162.41, 183.98]]}, {\"category\": \"car\", \"corners_3d\": [[183.9, 218.94], [143.6, 218.86], [199.87, 213.55], [235.6, 213.61], [183.9, 186.47], [143.6, 186.45], [199.87, 184.88], [235.6, 184.89]]}, {\"category\": \"car\", \"corners_3d\": [[264.6, 211.84], [234.19, 211.81], [272.69, 207.95], [300.11, 207.98], [264.6, 184.8], [234.19, 184.79], [272.69, 183.61], [300.11, 183.62]]}, {\"category\": \"car\", \"corners_3d\": [[312.28, 205.79], [289.56, 205.75], [314.67, 203.33], [335.73, 203.36], [312.28, 182.12], [289.56, 182.11], [314.67, 181.43], [335.73, 181.44]]}, {\"category\": \"car\", \"corners_3d\": [[347.62, 203.85], [324.11, 203.8], [344.19, 201.86], [366.26, 201.91], [347.62, 183.56], [324.11, 183.55], [344.19, 182.88], [366.26, 182.89]]}, {\"category\": \"car\", \"corners_3d\": [[396.49, 198.03], [379.75, 197.98], [394.91, 196.68], [410.82, 196.73], [396.49, 179.56], [379.75, 179.54], [394.91, 179.2], [410.82, 179.21]]}]\n```" - } - ] - }, - { - "id": 1083, - "data_source": "Mono3DRefer", - "scene_name": "000689", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[495.71, 209.34], [488.19, 211.01], [401.86, 210.69], [413.05, 209.05], [495.71, 183.76], [488.19, 184.26], [401.86, 184.16], [413.05, 183.67]]}]\n```" - } - ] - }, - { - "id": 1084, - "data_source": "Mono3DRefer", - "scene_name": "000697", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[52.77, 230.59], [16.13, 230.5], [89.56, 223.61], [121.91, 223.68], [52.77, 192.56], [16.13, 192.53], [89.56, 190.18], [121.91, 190.2]]}, {\"category\": \"car\", \"corners_3d\": [[145.43, 220.21], [120.73, 220.11], [170.64, 215.67], [193.12, 215.75], [145.43, 191.94], [120.73, 191.9], [170.64, 190.12], [193.12, 190.15]]}, {\"category\": \"car\", \"corners_3d\": [[196.97, 212.9], [168.9, 212.93], [206.48, 209.39], [232.04, 209.36], [196.97, 187.22], [168.9, 187.24], [206.48, 185.97], [232.04, 185.96]]}, {\"category\": \"car\", \"corners_3d\": [[247.18, 210.3], [223.38, 210.34], [250.04, 207.59], [272.07, 207.56], [247.18, 185.68], [223.38, 185.69], [250.04, 184.75], [272.07, 184.74]]}]\n```" - } - ] - }, - { - "id": 1085, - "data_source": "Mono3DRefer", - "scene_name": "000697", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[556.43, 204.35], [607.55, 204.28], [614.41, 215.58], [544.87, 215.7], [556.43, 142.06], [607.55, 142.12], [614.41, 131.07], [544.87, 130.95]]}]\n```" - } - ] - }, - { - "id": 1086, - "data_source": "Mono3DRefer", - "scene_name": "000697", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[556.43, 204.35], [607.55, 204.28], [614.41, 215.58], [544.87, 215.7], [556.43, 142.06], [607.55, 142.12], [614.41, 131.07], [544.87, 130.95]]}, {\"category\": \"car\", \"corners_3d\": [[52.77, 230.59], [16.13, 230.5], [89.56, 223.61], [121.91, 223.68], [52.77, 192.56], [16.13, 192.53], [89.56, 190.18], [121.91, 190.2]]}, {\"category\": \"car\", \"corners_3d\": [[145.43, 220.21], [120.73, 220.11], [170.64, 215.67], [193.12, 215.75], [145.43, 191.94], [120.73, 191.9], [170.64, 190.12], [193.12, 190.15]]}, {\"category\": \"car\", \"corners_3d\": [[196.97, 212.9], [168.9, 212.93], [206.48, 209.39], [232.04, 209.36], [196.97, 187.22], [168.9, 187.24], [206.48, 185.97], [232.04, 185.96]]}, {\"category\": \"car\", \"corners_3d\": [[247.18, 210.3], [223.38, 210.34], [250.04, 207.59], [272.07, 207.56], [247.18, 185.68], [223.38, 185.69], [250.04, 184.75], [272.07, 184.74]]}]\n```" - } - ] - }, - { - "id": 1087, - "data_source": "Mono3DRefer", - "scene_name": "000699", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[189.3, 232.98], [136.38, 232.8], [207.4, 224.31], [252.98, 224.44], [189.3, 185.19], [136.38, 185.15], [207.4, 183.41], [252.98, 183.44]]}]\n```" - } - ] - }, - { - "id": 1088, - "data_source": "Mono3DRefer", - "scene_name": "000699", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[751.29, 253.48], [818.89, 251.0], [940.47, 273.0], [858.05, 277.11], [751.29, 183.66], [818.89, 183.32], [940.47, 186.27], [858.05, 186.82]]}]\n```" - } - ] - }, - { - "id": 1089, - "data_source": "Mono3DRefer", - "scene_name": "000699", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[595.39, 189.11], [618.62, 189.1], [622.7, 191.99], [595.34, 192.0], [595.39, 163.93], [618.62, 163.93], [622.7, 162.34], [595.34, 162.34]]}]\n```" - } - ] - }, - { - "id": 1090, - "data_source": "Mono3DRefer", - "scene_name": "000699", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[751.29, 253.48], [818.89, 251.0], [940.47, 273.0], [858.05, 277.11], [751.29, 183.66], [818.89, 183.32], [940.47, 186.27], [858.05, 186.82]]}, {\"category\": \"van\", \"corners_3d\": [[189.3, 232.98], [136.38, 232.8], [207.4, 224.31], [252.98, 224.44], [189.3, 185.19], [136.38, 185.15], [207.4, 183.41], [252.98, 183.44]]}]\n```" - } - ] - }, - { - "id": 1091, - "data_source": "Mono3DRefer", - "scene_name": "000699", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[595.39, 189.11], [618.62, 189.1], [622.7, 191.99], [595.34, 192.0], [595.39, 163.93], [618.62, 163.93], [622.7, 162.34], [595.34, 162.34]]}, {\"category\": \"van\", \"corners_3d\": [[189.3, 232.98], [136.38, 232.8], [207.4, 224.31], [252.98, 224.44], [189.3, 185.19], [136.38, 185.15], [207.4, 183.41], [252.98, 183.44]]}]\n```" - } - ] - }, - { - "id": 1092, - "data_source": "Mono3DRefer", - "scene_name": "000699", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[595.39, 189.11], [618.62, 189.1], [622.7, 191.99], [595.34, 192.0], [595.39, 163.93], [618.62, 163.93], [622.7, 162.34], [595.34, 162.34]]}, {\"category\": \"car\", \"corners_3d\": [[751.29, 253.48], [818.89, 251.0], [940.47, 273.0], [858.05, 277.11], [751.29, 183.66], [818.89, 183.32], [940.47, 186.27], [858.05, 186.82]]}]\n```" - } - ] - }, - { - "id": 1093, - "data_source": "Mono3DRefer", - "scene_name": "000699", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[595.39, 189.11], [618.62, 189.1], [622.7, 191.99], [595.34, 192.0], [595.39, 163.93], [618.62, 163.93], [622.7, 162.34], [595.34, 162.34]]}, {\"category\": \"car\", \"corners_3d\": [[751.29, 253.48], [818.89, 251.0], [940.47, 273.0], [858.05, 277.11], [751.29, 183.66], [818.89, 183.32], [940.47, 186.27], [858.05, 186.82]]}, {\"category\": \"van\", \"corners_3d\": [[189.3, 232.98], [136.38, 232.8], [207.4, 224.31], [252.98, 224.44], [189.3, 185.19], [136.38, 185.15], [207.4, 183.41], [252.98, 183.44]]}]\n```" - } - ] - }, - { - "id": 1094, - "data_source": "Mono3DRefer", - "scene_name": "000700", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[582.29, 189.69], [599.46, 189.68], [600.07, 190.62], [581.95, 190.63], [582.29, 176.18], [599.46, 176.17], [600.07, 176.36], [581.95, 176.36]]}, {\"category\": \"car\", \"corners_3d\": [[358.13, 244.12], [292.44, 244.12], [345.03, 232.28], [399.81, 232.28], [358.13, 184.85], [292.44, 184.85], [345.03, 182.86], [399.81, 182.86]]}]\n```" - } - ] - }, - { - "id": 1095, - "data_source": "Mono3DRefer", - "scene_name": "000704", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[721.18, 185.1], [744.24, 185.27], [727.77, 186.35], [702.94, 186.16], [721.18, 154.59], [744.24, 154.35], [727.77, 152.73], [702.94, 153.01]]}]\n```" - } - ] - }, - { - "id": 1096, - "data_source": "Mono3DRefer", - "scene_name": "000704", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[55.37, 419.0], [284.6, 418.29], [-1512.15, 1803.66], [-3096.11, 1835.74], [55.37, 207.12], [284.6, 207.02], [-1512.15, 399.87], [-3096.11, 404.34]]}, {\"category\": \"car\", \"corners_3d\": [[658.89, 212.31], [704.08, 212.43], [713.0, 218.54], [660.83, 218.37], [658.89, 174.49], [704.08, 174.49], [713.0, 174.75], [660.83, 174.74]]}, {\"category\": \"car\", \"corners_3d\": [[751.43, 280.34], [867.54, 280.53], [1069.56, 367.47], [859.36, 366.87], [751.43, 169.86], [867.54, 169.86], [1069.56, 167.44], [859.36, 167.45]]}]\n```" - } - ] - }, - { - "id": 1097, - "data_source": "Mono3DRefer", - "scene_name": "000704", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[55.37, 419.0], [284.6, 418.29], [-1512.15, 1803.66], [-3096.11, 1835.74], [55.37, 207.12], [284.6, 207.02], [-1512.15, 399.87], [-3096.11, 404.34]]}, {\"category\": \"car\", \"corners_3d\": [[658.89, 212.31], [704.08, 212.43], [713.0, 218.54], [660.83, 218.37], [658.89, 174.49], [704.08, 174.49], [713.0, 174.75], [660.83, 174.74]]}, {\"category\": \"van\", \"corners_3d\": [[721.18, 185.1], [744.24, 185.27], [727.77, 186.35], [702.94, 186.16], [721.18, 154.59], [744.24, 154.35], [727.77, 152.73], [702.94, 153.01]]}, {\"category\": \"car\", \"corners_3d\": [[751.43, 280.34], [867.54, 280.53], [1069.56, 367.47], [859.36, 366.87], [751.43, 169.86], [867.54, 169.86], [1069.56, 167.44], [859.36, 167.45]]}]\n```" - } - ] - }, - { - "id": 1098, - "data_source": "Mono3DRefer", - "scene_name": "000706", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[693.64, 240.65], [751.14, 240.66], [785.21, 257.04], [713.81, 257.03], [693.64, 185.0], [751.14, 185.0], [785.21, 187.94], [713.81, 187.94]]}, {\"category\": \"car\", \"corners_3d\": [[673.66, 227.25], [722.38, 227.18], [745.47, 237.08], [687.89, 237.18], [673.66, 181.57], [722.38, 181.56], [745.47, 183.15], [687.89, 183.16]]}, {\"category\": \"car\", \"corners_3d\": [[558.07, 214.69], [525.38, 214.81], [529.56, 210.45], [558.86, 210.36], [558.07, 181.74], [525.38, 181.76], [529.56, 180.84], [558.86, 180.82]]}]\n```" - } - ] - }, - { - "id": 1099, - "data_source": "Mono3DRefer", - "scene_name": "000706", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[653.72, 215.21], [697.19, 215.21], [715.14, 223.94], [662.71, 223.94], [653.72, 162.97], [697.19, 162.97], [715.14, 160.93], [662.71, 160.93]]}]\n```" - } - ] - }, - { - "id": 1100, - "data_source": "Mono3DRefer", - "scene_name": "000706", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[693.64, 240.65], [751.14, 240.66], [785.21, 257.04], [713.81, 257.03], [693.64, 185.0], [751.14, 185.0], [785.21, 187.94], [713.81, 187.94]]}, {\"category\": \"car\", \"corners_3d\": [[673.66, 227.25], [722.38, 227.18], [745.47, 237.08], [687.89, 237.18], [673.66, 181.57], [722.38, 181.56], [745.47, 183.15], [687.89, 183.16]]}, {\"category\": \"van\", \"corners_3d\": [[653.72, 215.21], [697.19, 215.21], [715.14, 223.94], [662.71, 223.94], [653.72, 162.97], [697.19, 162.97], [715.14, 160.93], [662.71, 160.93]]}, {\"category\": \"car\", \"corners_3d\": [[558.07, 214.69], [525.38, 214.81], [529.56, 210.45], [558.86, 210.36], [558.07, 181.74], [525.38, 181.76], [529.56, 180.84], [558.86, 180.82]]}]\n```" - } - ] - }, - { - "id": 1101, - "data_source": "Mono3DRefer", - "scene_name": "000707", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[460.68, 281.47], [361.39, 281.18], [428.14, 253.59], [502.22, 253.75], [460.68, 193.64], [361.39, 193.58], [428.14, 188.3], [502.22, 188.33]]}, {\"category\": \"car\", \"corners_3d\": [[557.96, 211.79], [524.15, 211.73], [533.9, 208.17], [564.63, 208.22], [557.96, 183.03], [524.15, 183.02], [533.9, 182.09], [564.63, 182.1]]}]\n```" - } - ] - }, - { - "id": 1102, - "data_source": "Mono3DRefer", - "scene_name": "000708", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[519.04, 215.45], [479.52, 215.34], [497.84, 210.64], [533.02, 210.73], [519.04, 178.97], [479.52, 178.95], [497.84, 178.28], [533.02, 178.29]]}, {\"category\": \"car\", \"corners_3d\": [[694.54, 214.17], [734.46, 214.2], [750.7, 219.93], [705.24, 219.9], [694.54, 177.3], [734.46, 177.3], [750.7, 177.92], [705.24, 177.92]]}, {\"category\": \"car\", \"corners_3d\": [[-149.96, 518.85], [-495.72, 523.8], [56.11, 344.99], [223.09, 343.79], [-149.96, 237.67], [-495.72, 238.6], [56.11, 205.1], [223.09, 204.88]]}, {\"category\": \"car\", \"corners_3d\": [[272.8, 319.05], [133.29, 318.79], [294.62, 270.04], [387.66, 270.15], [272.8, 203.74], [133.29, 203.69], [294.62, 193.39], [387.66, 193.41]]}, {\"category\": \"car\", \"corners_3d\": [[853.9, 327.37], [1004.05, 326.75], [1462.26, 499.43], [1146.12, 502.25], [853.9, 188.55], [1004.05, 188.49], [1462.26, 206.04], [1146.12, 206.32]]}, {\"category\": \"car\", \"corners_3d\": [[498.61, 221.9], [452.54, 221.87], [471.1, 216.31], [511.95, 216.33], [498.61, 180.06], [452.54, 180.05], [471.1, 179.24], [511.95, 179.24]]}, {\"category\": \"car\", \"corners_3d\": [[763.21, 269.4], [859.8, 269.41], [957.96, 307.37], [823.38, 307.35], [763.21, 181.56], [859.8, 181.56], [957.96, 184.98], [823.38, 184.98]]}, {\"category\": \"car\", \"corners_3d\": [[727.46, 238.52], [792.11, 238.58], [834.44, 254.51], [754.1, 254.41], [727.46, 178.05], [792.11, 178.05], [834.44, 179.31], [754.1, 179.3]]}, {\"category\": \"car\", \"corners_3d\": [[700.83, 223.61], [751.81, 223.58], [780.17, 233.24], [719.49, 233.29], [700.83, 178.32], [751.81, 178.31], [780.17, 179.35], [719.49, 179.36]]}]\n```" - } - ] - }, - { - "id": 1103, - "data_source": "Mono3DRefer", - "scene_name": "000710", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[458.44, 257.3], [381.66, 258.89], [403.13, 239.64], [462.68, 238.68], [458.44, 191.87], [381.66, 192.22], [403.13, 187.89], [462.68, 187.67]]}, {\"category\": \"car\", \"corners_3d\": [[532.54, 199.77], [551.34, 199.61], [557.09, 201.0], [537.34, 201.18], [532.54, 181.37], [551.34, 181.32], [557.09, 181.77], [537.34, 181.82]]}]\n```" - } - ] - }, - { - "id": 1104, - "data_source": "Mono3DRefer", - "scene_name": "000714", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[1015.26, 203.16], [957.07, 202.31], [950.61, 196.81], [997.53, 197.37], [1015.26, 119.91], [957.07, 121.4], [950.61, 131.01], [997.53, 130.03]]}]\n```" - } - ] - }, - { - "id": 1105, - "data_source": "Mono3DRefer", - "scene_name": "000714", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[596.45, 250.2], [662.92, 253.25], [577.69, 280.39], [494.16, 275.02], [596.45, 189.62], [662.92, 190.28], [577.69, 196.17], [494.16, 195.0]]}, {\"category\": \"car\", \"corners_3d\": [[490.35, 247.68], [545.33, 250.43], [430.33, 273.48], [364.99, 268.89], [490.35, 170.57], [545.33, 170.49], [430.33, 169.79], [364.99, 169.93]]}, {\"category\": \"car\", \"corners_3d\": [[779.58, 219.54], [836.02, 220.81], [826.44, 230.84], [758.87, 229.0], [779.58, 172.84], [836.02, 172.84], [826.44, 172.84], [758.87, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[968.79, 206.25], [1017.91, 207.08], [1031.73, 212.61], [974.49, 211.49], [968.79, 165.69], [1017.91, 165.51], [1031.73, 164.32], [974.49, 164.56]]}, {\"category\": \"car\", \"corners_3d\": [[331.23, 244.33], [374.69, 246.44], [232.87, 266.06], [183.35, 262.7], [331.23, 192.74], [374.69, 193.33], [232.87, 198.78], [183.35, 197.85]]}, {\"category\": \"car\", \"corners_3d\": [[210.68, 244.82], [248.46, 246.55], [141.22, 259.19], [100.07, 256.83], [210.68, 200.62], [248.46, 201.29], [141.22, 206.17], [100.07, 205.26]]}]\n```" - } - ] - }, - { - "id": 1106, - "data_source": "Mono3DRefer", - "scene_name": "000714", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[904.02, 254.24], [1003.59, 258.39], [1037.04, 294.27], [896.26, 286.07], [904.02, 157.73], [1003.59, 156.95], [1037.04, 150.28], [896.26, 151.81]]}, {\"category\": \"van\", \"corners_3d\": [[1093.52, 200.81], [1042.08, 200.15], [1023.95, 196.29], [1067.59, 196.77], [1093.52, 140.73], [1042.08, 141.49], [1023.95, 145.92], [1067.59, 145.37]]}]\n```" - } - ] - }, - { - "id": 1107, - "data_source": "Mono3DRefer", - "scene_name": "000714", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[596.45, 250.2], [662.92, 253.25], [577.69, 280.39], [494.16, 275.02], [596.45, 189.62], [662.92, 190.28], [577.69, 196.17], [494.16, 195.0]]}, {\"category\": \"car\", \"corners_3d\": [[490.35, 247.68], [545.33, 250.43], [430.33, 273.48], [364.99, 268.89], [490.35, 170.57], [545.33, 170.49], [430.33, 169.79], [364.99, 169.93]]}, {\"category\": \"car\", \"corners_3d\": [[779.58, 219.54], [836.02, 220.81], [826.44, 230.84], [758.87, 229.0], [779.58, 172.84], [836.02, 172.84], [826.44, 172.84], [758.87, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[968.79, 206.25], [1017.91, 207.08], [1031.73, 212.61], [974.49, 211.49], [968.79, 165.69], [1017.91, 165.51], [1031.73, 164.32], [974.49, 164.56]]}, {\"category\": \"truck\", \"corners_3d\": [[1015.26, 203.16], [957.07, 202.31], [950.61, 196.81], [997.53, 197.37], [1015.26, 119.91], [957.07, 121.4], [950.61, 131.01], [997.53, 130.03]]}, {\"category\": \"car\", \"corners_3d\": [[331.23, 244.33], [374.69, 246.44], [232.87, 266.06], [183.35, 262.7], [331.23, 192.74], [374.69, 193.33], [232.87, 198.78], [183.35, 197.85]]}, {\"category\": \"car\", \"corners_3d\": [[210.68, 244.82], [248.46, 246.55], [141.22, 259.19], [100.07, 256.83], [210.68, 200.62], [248.46, 201.29], [141.22, 206.17], [100.07, 205.26]]}]\n```" - } - ] - }, - { - "id": 1108, - "data_source": "Mono3DRefer", - "scene_name": "000714", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[904.02, 254.24], [1003.59, 258.39], [1037.04, 294.27], [896.26, 286.07], [904.02, 157.73], [1003.59, 156.95], [1037.04, 150.28], [896.26, 151.81]]}, {\"category\": \"van\", \"corners_3d\": [[1093.52, 200.81], [1042.08, 200.15], [1023.95, 196.29], [1067.59, 196.77], [1093.52, 140.73], [1042.08, 141.49], [1023.95, 145.92], [1067.59, 145.37]]}, {\"category\": \"truck\", \"corners_3d\": [[1015.26, 203.16], [957.07, 202.31], [950.61, 196.81], [997.53, 197.37], [1015.26, 119.91], [957.07, 121.4], [950.61, 131.01], [997.53, 130.03]]}]\n```" - } - ] - }, - { - "id": 1109, - "data_source": "Mono3DRefer", - "scene_name": "000714", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[904.02, 254.24], [1003.59, 258.39], [1037.04, 294.27], [896.26, 286.07], [904.02, 157.73], [1003.59, 156.95], [1037.04, 150.28], [896.26, 151.81]]}, {\"category\": \"car\", \"corners_3d\": [[596.45, 250.2], [662.92, 253.25], [577.69, 280.39], [494.16, 275.02], [596.45, 189.62], [662.92, 190.28], [577.69, 196.17], [494.16, 195.0]]}, {\"category\": \"car\", \"corners_3d\": [[490.35, 247.68], [545.33, 250.43], [430.33, 273.48], [364.99, 268.89], [490.35, 170.57], [545.33, 170.49], [430.33, 169.79], [364.99, 169.93]]}, {\"category\": \"car\", \"corners_3d\": [[779.58, 219.54], [836.02, 220.81], [826.44, 230.84], [758.87, 229.0], [779.58, 172.84], [836.02, 172.84], [826.44, 172.84], [758.87, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[968.79, 206.25], [1017.91, 207.08], [1031.73, 212.61], [974.49, 211.49], [968.79, 165.69], [1017.91, 165.51], [1031.73, 164.32], [974.49, 164.56]]}, {\"category\": \"van\", \"corners_3d\": [[1093.52, 200.81], [1042.08, 200.15], [1023.95, 196.29], [1067.59, 196.77], [1093.52, 140.73], [1042.08, 141.49], [1023.95, 145.92], [1067.59, 145.37]]}, {\"category\": \"car\", \"corners_3d\": [[331.23, 244.33], [374.69, 246.44], [232.87, 266.06], [183.35, 262.7], [331.23, 192.74], [374.69, 193.33], [232.87, 198.78], [183.35, 197.85]]}, {\"category\": \"car\", \"corners_3d\": [[210.68, 244.82], [248.46, 246.55], [141.22, 259.19], [100.07, 256.83], [210.68, 200.62], [248.46, 201.29], [141.22, 206.17], [100.07, 205.26]]}]\n```" - } - ] - }, - { - "id": 1110, - "data_source": "Mono3DRefer", - "scene_name": "000714", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[904.02, 254.24], [1003.59, 258.39], [1037.04, 294.27], [896.26, 286.07], [904.02, 157.73], [1003.59, 156.95], [1037.04, 150.28], [896.26, 151.81]]}, {\"category\": \"car\", \"corners_3d\": [[596.45, 250.2], [662.92, 253.25], [577.69, 280.39], [494.16, 275.02], [596.45, 189.62], [662.92, 190.28], [577.69, 196.17], [494.16, 195.0]]}, {\"category\": \"car\", \"corners_3d\": [[490.35, 247.68], [545.33, 250.43], [430.33, 273.48], [364.99, 268.89], [490.35, 170.57], [545.33, 170.49], [430.33, 169.79], [364.99, 169.93]]}, {\"category\": \"car\", \"corners_3d\": [[779.58, 219.54], [836.02, 220.81], [826.44, 230.84], [758.87, 229.0], [779.58, 172.84], [836.02, 172.84], [826.44, 172.84], [758.87, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[968.79, 206.25], [1017.91, 207.08], [1031.73, 212.61], [974.49, 211.49], [968.79, 165.69], [1017.91, 165.51], [1031.73, 164.32], [974.49, 164.56]]}, {\"category\": \"van\", \"corners_3d\": [[1093.52, 200.81], [1042.08, 200.15], [1023.95, 196.29], [1067.59, 196.77], [1093.52, 140.73], [1042.08, 141.49], [1023.95, 145.92], [1067.59, 145.37]]}, {\"category\": \"truck\", \"corners_3d\": [[1015.26, 203.16], [957.07, 202.31], [950.61, 196.81], [997.53, 197.37], [1015.26, 119.91], [957.07, 121.4], [950.61, 131.01], [997.53, 130.03]]}, {\"category\": \"car\", \"corners_3d\": [[331.23, 244.33], [374.69, 246.44], [232.87, 266.06], [183.35, 262.7], [331.23, 192.74], [374.69, 193.33], [232.87, 198.78], [183.35, 197.85]]}, {\"category\": \"car\", \"corners_3d\": [[210.68, 244.82], [248.46, 246.55], [141.22, 259.19], [100.07, 256.83], [210.68, 200.62], [248.46, 201.29], [141.22, 206.17], [100.07, 205.26]]}]\n```" - } - ] - }, - { - "id": 1111, - "data_source": "Mono3DRefer", - "scene_name": "000715", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[531.42, 210.54], [492.31, 210.31], [517.05, 205.59], [551.33, 205.76], [531.42, 165.02], [492.31, 165.07], [517.05, 166.05], [551.33, 166.01]]}]\n```" - } - ] - }, - { - "id": 1112, - "data_source": "Mono3DRefer", - "scene_name": "000715", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[758.46, 189.31], [773.41, 189.33], [779.26, 190.25], [763.47, 190.22], [758.46, 173.28], [773.41, 173.28], [779.26, 173.31], [763.47, 173.31]]}]\n```" - } - ] - }, - { - "id": 1113, - "data_source": "Mono3DRefer", - "scene_name": "000715", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[758.46, 189.31], [773.41, 189.33], [779.26, 190.25], [763.47, 190.22], [758.46, 173.28], [773.41, 173.28], [779.26, 173.31], [763.47, 173.31]]}, {\"category\": \"van\", \"corners_3d\": [[531.42, 210.54], [492.31, 210.31], [517.05, 205.59], [551.33, 205.76], [531.42, 165.02], [492.31, 165.07], [517.05, 166.05], [551.33, 166.01]]}]\n```" - } - ] - }, - { - "id": 1114, - "data_source": "Mono3DRefer", - "scene_name": "000717", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[673.61, 242.96], [741.91, 243.17], [762.65, 256.41], [681.46, 256.13], [673.61, 174.58], [741.91, 174.59], [762.65, 174.91], [681.46, 174.9]]}, {\"category\": \"car\", \"corners_3d\": [[473.3, 270.87], [395.65, 269.58], [461.2, 248.47], [522.41, 249.26], [473.3, 191.91], [395.65, 191.66], [461.2, 187.56], [522.41, 187.71]]}]\n```" - } - ] - }, - { - "id": 1115, - "data_source": "Mono3DRefer", - "scene_name": "000722", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[818.4, 228.44], [838.51, 228.57], [855.71, 234.19], [833.54, 234.04], [818.4, 160.36], [838.51, 160.33], [855.71, 159.07], [833.54, 159.1]]}, {\"category\": \"cyclist\", \"corners_3d\": [[746.86, 204.3], [756.05, 204.34], [759.45, 205.8], [749.84, 205.75], [746.86, 170.59], [756.05, 170.58], [759.45, 170.48], [749.84, 170.48]]}]\n```" - } - ] - }, - { - "id": 1116, - "data_source": "Mono3DRefer", - "scene_name": "000722", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[843.14, 219.4], [865.3, 219.4], [873.26, 220.85], [850.41, 220.85], [843.14, 164.82], [865.3, 164.82], [873.26, 164.56], [850.41, 164.56]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[869.46, 218.72], [884.86, 218.73], [895.92, 220.57], [879.89, 220.57], [869.46, 164.18], [884.86, 164.18], [895.92, 163.83], [879.89, 163.83]]}]\n```" - } - ] - }, - { - "id": 1117, - "data_source": "Mono3DRefer", - "scene_name": "000722", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[818.4, 228.44], [838.51, 228.57], [855.71, 234.19], [833.54, 234.04], [818.4, 160.36], [838.51, 160.33], [855.71, 159.07], [833.54, 159.1]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[843.14, 219.4], [865.3, 219.4], [873.26, 220.85], [850.41, 220.85], [843.14, 164.82], [865.3, 164.82], [873.26, 164.56], [850.41, 164.56]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[869.46, 218.72], [884.86, 218.73], [895.92, 220.57], [879.89, 220.57], [869.46, 164.18], [884.86, 164.18], [895.92, 163.83], [879.89, 163.83]]}, {\"category\": \"cyclist\", \"corners_3d\": [[746.86, 204.3], [756.05, 204.34], [759.45, 205.8], [749.84, 205.75], [746.86, 170.59], [756.05, 170.58], [759.45, 170.48], [749.84, 170.48]]}]\n```" - } - ] - }, - { - "id": 1118, - "data_source": "Mono3DRefer", - "scene_name": "000728", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[504.13, 218.57], [468.0, 218.55], [483.58, 213.75], [515.92, 213.77], [504.13, 185.47], [468.0, 185.46], [483.58, 184.14], [515.92, 184.14]]}, {\"category\": \"car\", \"corners_3d\": [[399.03, 286.39], [291.03, 287.57], [354.14, 260.5], [436.35, 259.81], [399.03, 190.0], [291.03, 190.18], [354.14, 186.09], [436.35, 185.98]]}, {\"category\": \"car\", \"corners_3d\": [[64.37, 463.39], [-252.99, 478.68], [119.35, 330.34], [277.01, 326.18], [64.37, 211.3], [-252.99, 213.33], [119.35, 193.7], [277.01, 193.15]]}]\n```" - } - ] - }, - { - "id": 1119, - "data_source": "Mono3DRefer", - "scene_name": "000732", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[483.33, 202.25], [506.71, 201.33], [555.74, 203.23], [532.44, 204.28], [483.33, 168.32], [506.71, 168.47], [555.74, 168.17], [532.44, 168.01]]}, {\"category\": \"car\", \"corners_3d\": [[591.79, 202.85], [622.53, 202.86], [623.05, 205.9], [589.2, 205.89], [591.79, 174.87], [622.53, 174.87], [623.05, 175.08], [589.2, 175.08]]}, {\"category\": \"car\", \"corners_3d\": [[589.28, 216.24], [635.99, 216.24], [639.84, 222.7], [586.17, 222.69], [589.28, 178.65], [635.99, 178.65], [639.84, 179.51], [586.17, 179.51]]}, {\"category\": \"car\", \"corners_3d\": [[239.31, 258.18], [313.07, 250.55], [452.58, 263.3], [381.53, 273.81], [239.31, 184.28], [313.07, 183.26], [452.58, 184.97], [381.53, 186.38]]}, {\"category\": \"car\", \"corners_3d\": [[402.03, 220.01], [443.21, 217.52], [525.75, 222.5], [484.84, 225.59], [402.03, 168.13], [443.21, 168.38], [525.75, 167.88], [484.84, 167.57]]}, {\"category\": \"car\", \"corners_3d\": [[436.33, 212.13], [472.61, 210.61], [528.55, 214.46], [490.9, 216.33], [436.33, 164.83], [472.61, 165.14], [528.55, 164.36], [490.9, 163.98]]}]\n```" - } - ] - }, - { - "id": 1120, - "data_source": "Mono3DRefer", - "scene_name": "000732", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[247.13, 215.06], [180.57, 215.1], [272.64, 205.86], [324.59, 205.84], [247.13, 134.93], [180.57, 134.9], [272.64, 143.19], [324.59, 143.22]]}]\n```" - } - ] - }, - { - "id": 1121, - "data_source": "Mono3DRefer", - "scene_name": "000732", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[646.66, 198.33], [680.74, 198.3], [694.62, 202.12], [655.42, 202.17], [646.66, 159.2], [680.74, 159.22], [694.62, 157.17], [655.42, 157.14]]}, {\"category\": \"van\", \"corners_3d\": [[313.42, 239.92], [371.8, 234.77], [504.74, 244.09], [449.61, 250.99], [313.42, 160.87], [371.8, 161.79], [504.74, 160.13], [449.61, 158.9]]}, {\"category\": \"van\", \"corners_3d\": [[-111.8, 342.25], [119.29, 313.93], [313.17, 405.9], [-52.06, 494.81], [-111.8, 103.74], [119.29, 115.3], [313.17, 77.78], [-52.06, 41.51]]}]\n```" - } - ] - }, - { - "id": 1122, - "data_source": "Mono3DRefer", - "scene_name": "000732", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[483.33, 202.25], [506.71, 201.33], [555.74, 203.23], [532.44, 204.28], [483.33, 168.32], [506.71, 168.47], [555.74, 168.17], [532.44, 168.01]]}, {\"category\": \"car\", \"corners_3d\": [[591.79, 202.85], [622.53, 202.86], [623.05, 205.9], [589.2, 205.89], [591.79, 174.87], [622.53, 174.87], [623.05, 175.08], [589.2, 175.08]]}, {\"category\": \"truck\", \"corners_3d\": [[247.13, 215.06], [180.57, 215.1], [272.64, 205.86], [324.59, 205.84], [247.13, 134.93], [180.57, 134.9], [272.64, 143.19], [324.59, 143.22]]}, {\"category\": \"car\", \"corners_3d\": [[589.28, 216.24], [635.99, 216.24], [639.84, 222.7], [586.17, 222.69], [589.28, 178.65], [635.99, 178.65], [639.84, 179.51], [586.17, 179.51]]}, {\"category\": \"car\", \"corners_3d\": [[239.31, 258.18], [313.07, 250.55], [452.58, 263.3], [381.53, 273.81], [239.31, 184.28], [313.07, 183.26], [452.58, 184.97], [381.53, 186.38]]}, {\"category\": \"car\", \"corners_3d\": [[402.03, 220.01], [443.21, 217.52], [525.75, 222.5], [484.84, 225.59], [402.03, 168.13], [443.21, 168.38], [525.75, 167.88], [484.84, 167.57]]}, {\"category\": \"car\", \"corners_3d\": [[436.33, 212.13], [472.61, 210.61], [528.55, 214.46], [490.9, 216.33], [436.33, 164.83], [472.61, 165.14], [528.55, 164.36], [490.9, 163.98]]}]\n```" - } - ] - }, - { - "id": 1123, - "data_source": "Mono3DRefer", - "scene_name": "000732", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[483.33, 202.25], [506.71, 201.33], [555.74, 203.23], [532.44, 204.28], [483.33, 168.32], [506.71, 168.47], [555.74, 168.17], [532.44, 168.01]]}, {\"category\": \"van\", \"corners_3d\": [[646.66, 198.33], [680.74, 198.3], [694.62, 202.12], [655.42, 202.17], [646.66, 159.2], [680.74, 159.22], [694.62, 157.17], [655.42, 157.14]]}, {\"category\": \"car\", \"corners_3d\": [[591.79, 202.85], [622.53, 202.86], [623.05, 205.9], [589.2, 205.89], [591.79, 174.87], [622.53, 174.87], [623.05, 175.08], [589.2, 175.08]]}, {\"category\": \"van\", \"corners_3d\": [[313.42, 239.92], [371.8, 234.77], [504.74, 244.09], [449.61, 250.99], [313.42, 160.87], [371.8, 161.79], [504.74, 160.13], [449.61, 158.9]]}, {\"category\": \"car\", \"corners_3d\": [[589.28, 216.24], [635.99, 216.24], [639.84, 222.7], [586.17, 222.69], [589.28, 178.65], [635.99, 178.65], [639.84, 179.51], [586.17, 179.51]]}, {\"category\": \"van\", \"corners_3d\": [[-111.8, 342.25], [119.29, 313.93], [313.17, 405.9], [-52.06, 494.81], [-111.8, 103.74], [119.29, 115.3], [313.17, 77.78], [-52.06, 41.51]]}, {\"category\": \"car\", \"corners_3d\": [[239.31, 258.18], [313.07, 250.55], [452.58, 263.3], [381.53, 273.81], [239.31, 184.28], [313.07, 183.26], [452.58, 184.97], [381.53, 186.38]]}, {\"category\": \"car\", \"corners_3d\": [[402.03, 220.01], [443.21, 217.52], [525.75, 222.5], [484.84, 225.59], [402.03, 168.13], [443.21, 168.38], [525.75, 167.88], [484.84, 167.57]]}, {\"category\": \"car\", \"corners_3d\": [[436.33, 212.13], [472.61, 210.61], [528.55, 214.46], [490.9, 216.33], [436.33, 164.83], [472.61, 165.14], [528.55, 164.36], [490.9, 163.98]]}]\n```" - } - ] - }, - { - "id": 1124, - "data_source": "Mono3DRefer", - "scene_name": "000732", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[646.66, 198.33], [680.74, 198.3], [694.62, 202.12], [655.42, 202.17], [646.66, 159.2], [680.74, 159.22], [694.62, 157.17], [655.42, 157.14]]}, {\"category\": \"van\", \"corners_3d\": [[313.42, 239.92], [371.8, 234.77], [504.74, 244.09], [449.61, 250.99], [313.42, 160.87], [371.8, 161.79], [504.74, 160.13], [449.61, 158.9]]}, {\"category\": \"truck\", \"corners_3d\": [[247.13, 215.06], [180.57, 215.1], [272.64, 205.86], [324.59, 205.84], [247.13, 134.93], [180.57, 134.9], [272.64, 143.19], [324.59, 143.22]]}, {\"category\": \"van\", \"corners_3d\": [[-111.8, 342.25], [119.29, 313.93], [313.17, 405.9], [-52.06, 494.81], [-111.8, 103.74], [119.29, 115.3], [313.17, 77.78], [-52.06, 41.51]]}]\n```" - } - ] - }, - { - "id": 1125, - "data_source": "Mono3DRefer", - "scene_name": "000732", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[483.33, 202.25], [506.71, 201.33], [555.74, 203.23], [532.44, 204.28], [483.33, 168.32], [506.71, 168.47], [555.74, 168.17], [532.44, 168.01]]}, {\"category\": \"van\", \"corners_3d\": [[646.66, 198.33], [680.74, 198.3], [694.62, 202.12], [655.42, 202.17], [646.66, 159.2], [680.74, 159.22], [694.62, 157.17], [655.42, 157.14]]}, {\"category\": \"car\", \"corners_3d\": [[591.79, 202.85], [622.53, 202.86], [623.05, 205.9], [589.2, 205.89], [591.79, 174.87], [622.53, 174.87], [623.05, 175.08], [589.2, 175.08]]}, {\"category\": \"van\", \"corners_3d\": [[313.42, 239.92], [371.8, 234.77], [504.74, 244.09], [449.61, 250.99], [313.42, 160.87], [371.8, 161.79], [504.74, 160.13], [449.61, 158.9]]}, {\"category\": \"truck\", \"corners_3d\": [[247.13, 215.06], [180.57, 215.1], [272.64, 205.86], [324.59, 205.84], [247.13, 134.93], [180.57, 134.9], [272.64, 143.19], [324.59, 143.22]]}, {\"category\": \"car\", \"corners_3d\": [[589.28, 216.24], [635.99, 216.24], [639.84, 222.7], [586.17, 222.69], [589.28, 178.65], [635.99, 178.65], [639.84, 179.51], [586.17, 179.51]]}, {\"category\": \"van\", \"corners_3d\": [[-111.8, 342.25], [119.29, 313.93], [313.17, 405.9], [-52.06, 494.81], [-111.8, 103.74], [119.29, 115.3], [313.17, 77.78], [-52.06, 41.51]]}, {\"category\": \"car\", \"corners_3d\": [[239.31, 258.18], [313.07, 250.55], [452.58, 263.3], [381.53, 273.81], [239.31, 184.28], [313.07, 183.26], [452.58, 184.97], [381.53, 186.38]]}, {\"category\": \"car\", \"corners_3d\": [[402.03, 220.01], [443.21, 217.52], [525.75, 222.5], [484.84, 225.59], [402.03, 168.13], [443.21, 168.38], [525.75, 167.88], [484.84, 167.57]]}, {\"category\": \"car\", \"corners_3d\": [[436.33, 212.13], [472.61, 210.61], [528.55, 214.46], [490.9, 216.33], [436.33, 164.83], [472.61, 165.14], [528.55, 164.36], [490.9, 163.98]]}]\n```" - } - ] - }, - { - "id": 1126, - "data_source": "Mono3DRefer", - "scene_name": "000734", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[666.61, 240.29], [736.21, 239.9], [791.6, 261.82], [699.5, 262.51], [666.61, 176.28], [736.21, 176.26], [791.6, 177.37], [699.5, 177.4]]}, {\"category\": \"car\", \"corners_3d\": [[210.45, 319.98], [364.98, 318.17], [133.6, 484.78], [-209.18, 493.23], [210.45, 187.54], [364.98, 187.36], [133.6, 203.98], [-209.18, 204.83]]}, {\"category\": \"car\", \"corners_3d\": [[160.27, 252.58], [198.24, 245.97], [332.14, 246.13], [306.22, 252.77], [160.27, 191.85], [198.24, 190.27], [332.14, 190.31], [306.22, 191.89]]}, {\"category\": \"car\", \"corners_3d\": [[509.36, 212.08], [551.35, 211.94], [549.08, 217.7], [500.87, 217.87], [509.36, 173.8], [551.35, 173.8], [549.08, 173.94], [500.87, 173.94]]}, {\"category\": \"car\", \"corners_3d\": [[633.43, 216.17], [673.47, 215.93], [694.99, 222.75], [648.7, 223.08], [633.43, 176.81], [673.47, 176.78], [694.99, 177.41], [648.7, 177.44]]}, {\"category\": \"car\", \"corners_3d\": [[536.14, 203.74], [561.98, 203.78], [556.41, 206.25], [528.51, 206.21], [536.14, 179.49], [561.98, 179.5], [556.41, 180.03], [528.51, 180.02]]}, {\"category\": \"car\", \"corners_3d\": [[620.45, 199.54], [647.63, 199.56], [650.12, 202.44], [620.01, 202.41], [620.45, 171.14], [647.63, 171.14], [650.12, 170.96], [620.01, 170.96]]}]\n```" - } - ] - }, - { - "id": 1127, - "data_source": "Mono3DRefer", - "scene_name": "000734", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[128.1, 329.38], [158.86, 329.39], [139.99, 335.94], [107.95, 335.93], [128.1, 165.58], [158.86, 165.58], [139.99, 165.28], [107.95, 165.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[178.34, 329.88], [214.64, 329.88], [192.02, 338.86], [153.64, 338.86], [178.34, 168.28], [214.64, 168.28], [192.02, 168.02], [153.64, 168.02]]}]\n```" - } - ] - }, - { - "id": 1128, - "data_source": "Mono3DRefer", - "scene_name": "000734", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[666.61, 240.29], [736.21, 239.9], [791.6, 261.82], [699.5, 262.51], [666.61, 176.28], [736.21, 176.26], [791.6, 177.37], [699.5, 177.4]]}, {\"category\": \"car\", \"corners_3d\": [[210.45, 319.98], [364.98, 318.17], [133.6, 484.78], [-209.18, 493.23], [210.45, 187.54], [364.98, 187.36], [133.6, 203.98], [-209.18, 204.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[128.1, 329.38], [158.86, 329.39], [139.99, 335.94], [107.95, 335.93], [128.1, 165.58], [158.86, 165.58], [139.99, 165.28], [107.95, 165.28]]}, {\"category\": \"car\", \"corners_3d\": [[160.27, 252.58], [198.24, 245.97], [332.14, 246.13], [306.22, 252.77], [160.27, 191.85], [198.24, 190.27], [332.14, 190.31], [306.22, 191.89]]}, {\"category\": \"car\", \"corners_3d\": [[509.36, 212.08], [551.35, 211.94], [549.08, 217.7], [500.87, 217.87], [509.36, 173.8], [551.35, 173.8], [549.08, 173.94], [500.87, 173.94]]}, {\"category\": \"car\", \"corners_3d\": [[633.43, 216.17], [673.47, 215.93], [694.99, 222.75], [648.7, 223.08], [633.43, 176.81], [673.47, 176.78], [694.99, 177.41], [648.7, 177.44]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[178.34, 329.88], [214.64, 329.88], [192.02, 338.86], [153.64, 338.86], [178.34, 168.28], [214.64, 168.28], [192.02, 168.02], [153.64, 168.02]]}, {\"category\": \"car\", \"corners_3d\": [[536.14, 203.74], [561.98, 203.78], [556.41, 206.25], [528.51, 206.21], [536.14, 179.49], [561.98, 179.5], [556.41, 180.03], [528.51, 180.02]]}, {\"category\": \"car\", \"corners_3d\": [[620.45, 199.54], [647.63, 199.56], [650.12, 202.44], [620.01, 202.41], [620.45, 171.14], [647.63, 171.14], [650.12, 170.96], [620.01, 170.96]]}]\n```" - } - ] - }, - { - "id": 1129, - "data_source": "Mono3DRefer", - "scene_name": "000737", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[749.42, 287.53], [889.5, 287.55], [1403.06, 498.4], [1005.36, 498.25], [749.42, 106.68], [889.5, 106.67], [1403.06, -15.0], [1005.36, -14.91]]}, {\"category\": \"van\", \"corners_3d\": [[521.82, 213.8], [515.3, 216.62], [362.63, 216.52], [378.94, 213.72], [521.82, 143.24], [515.3, 141.2], [362.63, 141.27], [378.94, 143.3]]}]\n```" - } - ] - }, - { - "id": 1130, - "data_source": "Mono3DRefer", - "scene_name": "000737", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[270.67, 299.33], [192.68, 327.83], [-190.06, 327.0], [-42.32, 298.77], [270.67, 168.95], [192.68, 168.07], [-190.06, 168.09], [-42.32, 168.96]]}, {\"category\": \"car\", \"corners_3d\": [[-11.51, 288.04], [74.45, 271.08], [349.45, 269.22], [311.83, 285.49], [-11.51, 175.76], [74.45, 175.33], [349.45, 175.29], [311.83, 175.7]]}, {\"category\": \"car\", \"corners_3d\": [[448.71, 235.65], [429.35, 242.88], [234.26, 242.69], [273.67, 235.5], [448.71, 171.16], [429.35, 170.97], [234.26, 170.97], [273.67, 171.17]]}, {\"category\": \"car\", \"corners_3d\": [[511.51, 213.69], [507.4, 216.4], [378.61, 216.79], [390.82, 214.03], [511.51, 173.11], [507.4, 173.13], [378.61, 173.13], [390.82, 173.12]]}, {\"category\": \"car\", \"corners_3d\": [[662.02, 215.43], [706.25, 215.44], [724.36, 223.46], [671.8, 223.46], [662.02, 176.94], [706.25, 176.94], [724.36, 177.71], [671.8, 177.71]]}, {\"category\": \"car\", \"corners_3d\": [[802.73, 209.95], [810.71, 212.03], [728.29, 211.72], [724.71, 209.68], [802.73, 175.74], [810.71, 175.9], [728.29, 175.87], [724.71, 175.72]]}]\n```" - } - ] - }, - { - "id": 1131, - "data_source": "Mono3DRefer", - "scene_name": "000737", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[749.42, 287.53], [889.5, 287.55], [1403.06, 498.4], [1005.36, 498.25], [749.42, 106.68], [889.5, 106.67], [1403.06, -15.0], [1005.36, -14.91]]}, {\"category\": \"car\", \"corners_3d\": [[270.67, 299.33], [192.68, 327.83], [-190.06, 327.0], [-42.32, 298.77], [270.67, 168.95], [192.68, 168.07], [-190.06, 168.09], [-42.32, 168.96]]}, {\"category\": \"car\", \"corners_3d\": [[-11.51, 288.04], [74.45, 271.08], [349.45, 269.22], [311.83, 285.49], [-11.51, 175.76], [74.45, 175.33], [349.45, 175.29], [311.83, 175.7]]}, {\"category\": \"car\", \"corners_3d\": [[448.71, 235.65], [429.35, 242.88], [234.26, 242.69], [273.67, 235.5], [448.71, 171.16], [429.35, 170.97], [234.26, 170.97], [273.67, 171.17]]}, {\"category\": \"car\", \"corners_3d\": [[511.51, 213.69], [507.4, 216.4], [378.61, 216.79], [390.82, 214.03], [511.51, 173.11], [507.4, 173.13], [378.61, 173.13], [390.82, 173.12]]}, {\"category\": \"van\", \"corners_3d\": [[521.82, 213.8], [515.3, 216.62], [362.63, 216.52], [378.94, 213.72], [521.82, 143.24], [515.3, 141.2], [362.63, 141.27], [378.94, 143.3]]}, {\"category\": \"car\", \"corners_3d\": [[662.02, 215.43], [706.25, 215.44], [724.36, 223.46], [671.8, 223.46], [662.02, 176.94], [706.25, 176.94], [724.36, 177.71], [671.8, 177.71]]}, {\"category\": \"car\", \"corners_3d\": [[802.73, 209.95], [810.71, 212.03], [728.29, 211.72], [724.71, 209.68], [802.73, 175.74], [810.71, 175.9], [728.29, 175.87], [724.71, 175.72]]}]\n```" - } - ] - }, - { - "id": 1132, - "data_source": "Mono3DRefer", - "scene_name": "000738", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[644.77, 212.93], [678.28, 212.78], [690.9, 216.78], [654.07, 216.96], [644.77, 180.06], [678.28, 180.04], [690.9, 180.75], [654.07, 180.79]]}]\n```" - } - ] - }, - { - "id": 1133, - "data_source": "Mono3DRefer", - "scene_name": "000740", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[509.87, 310.64], [408.14, 299.48], [568.8, 264.74], [650.91, 270.47], [509.87, 181.85], [408.14, 181.12], [568.8, 178.85], [650.91, 179.23]]}, {\"category\": \"car\", \"corners_3d\": [[1510.15, 604.55], [1087.75, 615.42], [786.69, 352.96], [964.18, 351.13], [1510.15, 218.92], [1087.75, 220.08], [786.69, 192.07], [964.18, 191.88]]}]\n```" - } - ] - }, - { - "id": 1134, - "data_source": "Mono3DRefer", - "scene_name": "000744", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[789.68, 321.45], [770.65, 311.32], [857.08, 309.48], [882.41, 319.31], [789.68, 145.54], [770.65, 148.06], [857.08, 148.51], [882.41, 146.07]]}]\n```" - } - ] - }, - { - "id": 1135, - "data_source": "Mono3DRefer", - "scene_name": "000747", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[613.48, 199.86], [635.62, 199.9], [634.83, 201.9], [611.06, 201.86], [613.48, 178.96], [635.62, 178.97], [634.83, 179.42], [611.06, 179.41]]}, {\"category\": \"car\", \"corners_3d\": [[488.33, 215.01], [459.23, 214.71], [483.28, 211.03], [509.96, 211.28], [488.33, 184.98], [459.23, 184.89], [483.28, 183.83], [509.96, 183.9]]}, {\"category\": \"car\", \"corners_3d\": [[511.55, 210.77], [485.27, 210.56], [505.23, 207.32], [529.34, 207.5], [511.55, 177.65], [485.27, 177.63], [505.23, 177.22], [529.34, 177.24]]}, {\"category\": \"car\", \"corners_3d\": [[656.78, 199.04], [681.05, 199.1], [683.35, 201.29], [657.05, 201.22], [656.78, 176.18], [681.05, 176.18], [683.35, 176.46], [657.05, 176.45]]}, {\"category\": \"car\", \"corners_3d\": [[942.51, 191.5], [932.52, 190.78], [1000.25, 190.94], [1013.07, 191.67], [942.51, 164.77], [932.52, 165.08], [1000.25, 165.01], [1013.07, 164.7]]}]\n```" - } - ] - }, - { - "id": 1136, - "data_source": "Mono3DRefer", - "scene_name": "000749", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[638.31, 197.58], [660.07, 197.54], [666.22, 199.3], [642.91, 199.35], [638.31, 176.17], [660.07, 176.16], [666.22, 176.4], [642.91, 176.4]]}, {\"category\": \"car\", \"corners_3d\": [[689.89, 233.18], [744.06, 233.05], [775.44, 245.19], [710.4, 245.38], [689.89, 182.71], [744.06, 182.69], [775.44, 184.68], [710.4, 184.71]]}, {\"category\": \"car\", \"corners_3d\": [[854.68, 318.13], [995.06, 317.3], [1358.3, 446.34], [1095.12, 449.35], [854.68, 186.03], [995.06, 185.96], [1358.3, 197.66], [1095.12, 197.94]]}, {\"category\": \"car\", \"corners_3d\": [[705.92, 199.11], [730.08, 199.05], [743.15, 201.15], [717.08, 201.21], [705.92, 175.49], [730.08, 175.49], [743.15, 175.7], [717.08, 175.7]]}, {\"category\": \"car\", \"corners_3d\": [[759.66, 210.99], [795.75, 210.84], [831.84, 216.47], [790.55, 216.67], [759.66, 176.87], [795.75, 176.86], [831.84, 177.45], [790.55, 177.47]]}]\n```" - } - ] - }, - { - "id": 1137, - "data_source": "Mono3DRefer", - "scene_name": "000749", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[947.67, 262.28], [1086.23, 261.77], [1852.21, 398.6], [1508.54, 401.9], [947.67, 66.28], [1086.23, 66.88], [1852.21, -96.2], [1508.54, -100.14]]}]\n```" - } - ] - }, - { - "id": 1138, - "data_source": "Mono3DRefer", - "scene_name": "000749", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[638.31, 197.58], [660.07, 197.54], [666.22, 199.3], [642.91, 199.35], [638.31, 176.17], [660.07, 176.16], [666.22, 176.4], [642.91, 176.4]]}, {\"category\": \"car\", \"corners_3d\": [[689.89, 233.18], [744.06, 233.05], [775.44, 245.19], [710.4, 245.38], [689.89, 182.71], [744.06, 182.69], [775.44, 184.68], [710.4, 184.71]]}, {\"category\": \"car\", \"corners_3d\": [[854.68, 318.13], [995.06, 317.3], [1358.3, 446.34], [1095.12, 449.35], [854.68, 186.03], [995.06, 185.96], [1358.3, 197.66], [1095.12, 197.94]]}, {\"category\": \"car\", \"corners_3d\": [[705.92, 199.11], [730.08, 199.05], [743.15, 201.15], [717.08, 201.21], [705.92, 175.49], [730.08, 175.49], [743.15, 175.7], [717.08, 175.7]]}, {\"category\": \"car\", \"corners_3d\": [[759.66, 210.99], [795.75, 210.84], [831.84, 216.47], [790.55, 216.67], [759.66, 176.87], [795.75, 176.86], [831.84, 177.45], [790.55, 177.47]]}, {\"category\": \"truck\", \"corners_3d\": [[947.67, 262.28], [1086.23, 261.77], [1852.21, 398.6], [1508.54, 401.9], [947.67, 66.28], [1086.23, 66.88], [1852.21, -96.2], [1508.54, -100.14]]}]\n```" - } - ] - }, - { - "id": 1139, - "data_source": "Mono3DRefer", - "scene_name": "000753", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[261.93, 408.91], [13.72, 419.68], [212.41, 320.32], [358.08, 316.4], [261.93, 201.35], [13.72, 202.65], [212.41, 190.65], [358.08, 190.18]]}, {\"category\": \"car\", \"corners_3d\": [[611.63, 202.22], [607.94, 200.91], [693.79, 200.54], [701.37, 201.81], [611.63, 168.57], [607.94, 168.76], [693.79, 168.81], [701.37, 168.63]]}, {\"category\": \"car\", \"corners_3d\": [[481.78, 204.62], [453.04, 204.75], [459.51, 202.04], [485.8, 201.93], [481.78, 179.27], [453.04, 179.29], [459.51, 178.75], [485.8, 178.72]]}, {\"category\": \"car\", \"corners_3d\": [[742.84, 177.04], [759.59, 177.03], [772.32, 177.26], [754.7, 177.27], [742.84, 158.17], [759.59, 158.21], [772.32, 157.42], [754.7, 157.38]]}]\n```" - } - ] - }, - { - "id": 1140, - "data_source": "Mono3DRefer", - "scene_name": "000760", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[421.95, 296.07], [373.97, 319.46], [38.39, 313.12], [136.19, 291.56], [421.95, 199.98], [373.97, 205.13], [38.39, 203.74], [136.19, 198.99]]}, {\"category\": \"car\", \"corners_3d\": [[1432.24, 347.35], [1778.48, 424.87], [1032.71, 410.61], [920.54, 340.39], [1432.24, 170.42], [1778.48, 169.34], [1032.71, 169.54], [920.54, 170.52]]}, {\"category\": \"car\", \"corners_3d\": [[1145.77, 290.81], [1265.08, 316.04], [863.15, 318.5], [813.97, 292.48], [1145.77, 157.5], [1265.08, 154.22], [863.15, 153.9], [813.97, 157.29]]}, {\"category\": \"car\", \"corners_3d\": [[338.09, 238.37], [365.06, 233.25], [507.48, 234.51], [492.25, 239.86], [338.09, 184.25], [365.06, 183.36], [507.48, 183.58], [492.25, 184.51]]}, {\"category\": \"car\", \"corners_3d\": [[404.27, 223.43], [413.09, 220.22], [534.12, 219.51], [533.5, 222.62], [404.27, 182.84], [413.09, 182.21], [534.12, 182.07], [533.5, 182.68]]}, {\"category\": \"car\", \"corners_3d\": [[923.87, 239.61], [956.79, 246.77], [799.71, 246.61], [782.03, 239.48], [923.87, 171.51], [956.79, 171.37], [799.71, 171.37], [782.03, 171.51]]}, {\"category\": \"car\", \"corners_3d\": [[997.12, 247.57], [1049.57, 257.87], [815.78, 257.6], [791.74, 247.36], [997.12, 170.8], [1049.57, 170.52], [815.78, 170.53], [791.74, 170.81]]}]\n```" - } - ] - }, - { - "id": 1141, - "data_source": "Mono3DRefer", - "scene_name": "000760", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[542.92, 209.61], [556.2, 209.27], [566.46, 210.44], [552.85, 210.79], [542.92, 169.19], [556.2, 169.22], [566.46, 169.11], [552.85, 169.07]]}]\n```" - } - ] - }, - { - "id": 1142, - "data_source": "Mono3DRefer", - "scene_name": "000760", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[421.95, 296.07], [373.97, 319.46], [38.39, 313.12], [136.19, 291.56], [421.95, 199.98], [373.97, 205.13], [38.39, 203.74], [136.19, 198.99]]}, {\"category\": \"car\", \"corners_3d\": [[1432.24, 347.35], [1778.48, 424.87], [1032.71, 410.61], [920.54, 340.39], [1432.24, 170.42], [1778.48, 169.34], [1032.71, 169.54], [920.54, 170.52]]}, {\"category\": \"car\", \"corners_3d\": [[1145.77, 290.81], [1265.08, 316.04], [863.15, 318.5], [813.97, 292.48], [1145.77, 157.5], [1265.08, 154.22], [863.15, 153.9], [813.97, 157.29]]}, {\"category\": \"car\", \"corners_3d\": [[338.09, 238.37], [365.06, 233.25], [507.48, 234.51], [492.25, 239.86], [338.09, 184.25], [365.06, 183.36], [507.48, 183.58], [492.25, 184.51]]}, {\"category\": \"car\", \"corners_3d\": [[404.27, 223.43], [413.09, 220.22], [534.12, 219.51], [533.5, 222.62], [404.27, 182.84], [413.09, 182.21], [534.12, 182.07], [533.5, 182.68]]}, {\"category\": \"car\", \"corners_3d\": [[923.87, 239.61], [956.79, 246.77], [799.71, 246.61], [782.03, 239.48], [923.87, 171.51], [956.79, 171.37], [799.71, 171.37], [782.03, 171.51]]}, {\"category\": \"car\", \"corners_3d\": [[997.12, 247.57], [1049.57, 257.87], [815.78, 257.6], [791.74, 247.36], [997.12, 170.8], [1049.57, 170.52], [815.78, 170.53], [791.74, 170.81]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[542.92, 209.61], [556.2, 209.27], [566.46, 210.44], [552.85, 210.79], [542.92, 169.19], [556.2, 169.22], [566.46, 169.11], [552.85, 169.07]]}]\n```" - } - ] - }, - { - "id": 1143, - "data_source": "Mono3DRefer", - "scene_name": "000763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[791.81, 199.26], [793.9, 198.59], [838.81, 199.13], [837.86, 199.83], [791.81, 162.83], [793.9, 163.08], [838.81, 162.88], [837.86, 162.61]]}]\n```" - } - ] - }, - { - "id": 1144, - "data_source": "Mono3DRefer", - "scene_name": "000763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[540.73, 201.03], [539.27, 201.57], [528.76, 201.57], [530.43, 201.03], [540.73, 169.35], [539.27, 169.28], [528.76, 169.28], [530.43, 169.35]]}]\n```" - } - ] - }, - { - "id": 1145, - "data_source": "Mono3DRefer", - "scene_name": "000763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[876.99, 183.52], [874.83, 183.21], [933.47, 183.41], [937.48, 183.74], [876.99, 161.77], [874.83, 162.09], [933.47, 161.88], [937.48, 161.54]]}, {\"category\": \"car\", \"corners_3d\": [[412.93, 197.0], [397.69, 197.65], [362.58, 196.93], [378.29, 196.31], [412.93, 177.01], [397.69, 177.12], [362.58, 177.0], [378.29, 176.89]]}, {\"category\": \"car\", \"corners_3d\": [[711.39, 216.79], [758.71, 216.93], [782.15, 226.12], [724.91, 225.9], [711.39, 174.61], [758.71, 174.62], [782.15, 174.99], [724.91, 174.98]]}, {\"category\": \"car\", \"corners_3d\": [[731.35, 185.18], [728.31, 184.82], [780.76, 184.84], [785.36, 185.2], [731.35, 166.28], [728.31, 166.47], [780.76, 166.46], [785.36, 166.27]]}]\n```" - } - ] - }, - { - "id": 1146, - "data_source": "Mono3DRefer", - "scene_name": "000763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1135.51, 181.87], [1124.61, 181.58], [1190.48, 181.76], [1203.9, 182.07], [1135.51, 154.8], [1124.61, 155.39], [1190.48, 155.03], [1203.9, 154.41]]}]\n```" - } - ] - }, - { - "id": 1147, - "data_source": "Mono3DRefer", - "scene_name": "000763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[540.73, 201.03], [539.27, 201.57], [528.76, 201.57], [530.43, 201.03], [540.73, 169.35], [539.27, 169.28], [528.76, 169.28], [530.43, 169.35]]}, {\"category\": \"cyclist\", \"corners_3d\": [[791.81, 199.26], [793.9, 198.59], [838.81, 199.13], [837.86, 199.83], [791.81, 162.83], [793.9, 163.08], [838.81, 162.88], [837.86, 162.61]]}]\n```" - } - ] - }, - { - "id": 1148, - "data_source": "Mono3DRefer", - "scene_name": "000763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[791.81, 199.26], [793.9, 198.59], [838.81, 199.13], [837.86, 199.83], [791.81, 162.83], [793.9, 163.08], [838.81, 162.88], [837.86, 162.61]]}, {\"category\": \"car\", \"corners_3d\": [[876.99, 183.52], [874.83, 183.21], [933.47, 183.41], [937.48, 183.74], [876.99, 161.77], [874.83, 162.09], [933.47, 161.88], [937.48, 161.54]]}, {\"category\": \"car\", \"corners_3d\": [[412.93, 197.0], [397.69, 197.65], [362.58, 196.93], [378.29, 196.31], [412.93, 177.01], [397.69, 177.12], [362.58, 177.0], [378.29, 176.89]]}, {\"category\": \"car\", \"corners_3d\": [[711.39, 216.79], [758.71, 216.93], [782.15, 226.12], [724.91, 225.9], [711.39, 174.61], [758.71, 174.62], [782.15, 174.99], [724.91, 174.98]]}, {\"category\": \"car\", \"corners_3d\": [[731.35, 185.18], [728.31, 184.82], [780.76, 184.84], [785.36, 185.2], [731.35, 166.28], [728.31, 166.47], [780.76, 166.46], [785.36, 166.27]]}]\n```" - } - ] - }, - { - "id": 1149, - "data_source": "Mono3DRefer", - "scene_name": "000763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[791.81, 199.26], [793.9, 198.59], [838.81, 199.13], [837.86, 199.83], [791.81, 162.83], [793.9, 163.08], [838.81, 162.88], [837.86, 162.61]]}, {\"category\": \"van\", \"corners_3d\": [[1135.51, 181.87], [1124.61, 181.58], [1190.48, 181.76], [1203.9, 182.07], [1135.51, 154.8], [1124.61, 155.39], [1190.48, 155.03], [1203.9, 154.41]]}]\n```" - } - ] - }, - { - "id": 1150, - "data_source": "Mono3DRefer", - "scene_name": "000763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[540.73, 201.03], [539.27, 201.57], [528.76, 201.57], [530.43, 201.03], [540.73, 169.35], [539.27, 169.28], [528.76, 169.28], [530.43, 169.35]]}, {\"category\": \"car\", \"corners_3d\": [[876.99, 183.52], [874.83, 183.21], [933.47, 183.41], [937.48, 183.74], [876.99, 161.77], [874.83, 162.09], [933.47, 161.88], [937.48, 161.54]]}, {\"category\": \"car\", \"corners_3d\": [[412.93, 197.0], [397.69, 197.65], [362.58, 196.93], [378.29, 196.31], [412.93, 177.01], [397.69, 177.12], [362.58, 177.0], [378.29, 176.89]]}, {\"category\": \"car\", \"corners_3d\": [[711.39, 216.79], [758.71, 216.93], [782.15, 226.12], [724.91, 225.9], [711.39, 174.61], [758.71, 174.62], [782.15, 174.99], [724.91, 174.98]]}, {\"category\": \"car\", \"corners_3d\": [[731.35, 185.18], [728.31, 184.82], [780.76, 184.84], [785.36, 185.2], [731.35, 166.28], [728.31, 166.47], [780.76, 166.46], [785.36, 166.27]]}]\n```" - } - ] - }, - { - "id": 1151, - "data_source": "Mono3DRefer", - "scene_name": "000763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[540.73, 201.03], [539.27, 201.57], [528.76, 201.57], [530.43, 201.03], [540.73, 169.35], [539.27, 169.28], [528.76, 169.28], [530.43, 169.35]]}, {\"category\": \"van\", \"corners_3d\": [[1135.51, 181.87], [1124.61, 181.58], [1190.48, 181.76], [1203.9, 182.07], [1135.51, 154.8], [1124.61, 155.39], [1190.48, 155.03], [1203.9, 154.41]]}]\n```" - } - ] - }, - { - "id": 1152, - "data_source": "Mono3DRefer", - "scene_name": "000763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[876.99, 183.52], [874.83, 183.21], [933.47, 183.41], [937.48, 183.74], [876.99, 161.77], [874.83, 162.09], [933.47, 161.88], [937.48, 161.54]]}, {\"category\": \"van\", \"corners_3d\": [[1135.51, 181.87], [1124.61, 181.58], [1190.48, 181.76], [1203.9, 182.07], [1135.51, 154.8], [1124.61, 155.39], [1190.48, 155.03], [1203.9, 154.41]]}, {\"category\": \"car\", \"corners_3d\": [[412.93, 197.0], [397.69, 197.65], [362.58, 196.93], [378.29, 196.31], [412.93, 177.01], [397.69, 177.12], [362.58, 177.0], [378.29, 176.89]]}, {\"category\": \"car\", \"corners_3d\": [[711.39, 216.79], [758.71, 216.93], [782.15, 226.12], [724.91, 225.9], [711.39, 174.61], [758.71, 174.62], [782.15, 174.99], [724.91, 174.98]]}, {\"category\": \"car\", \"corners_3d\": [[731.35, 185.18], [728.31, 184.82], [780.76, 184.84], [785.36, 185.2], [731.35, 166.28], [728.31, 166.47], [780.76, 166.46], [785.36, 166.27]]}]\n```" - } - ] - }, - { - "id": 1153, - "data_source": "Mono3DRefer", - "scene_name": "000763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[540.73, 201.03], [539.27, 201.57], [528.76, 201.57], [530.43, 201.03], [540.73, 169.35], [539.27, 169.28], [528.76, 169.28], [530.43, 169.35]]}, {\"category\": \"cyclist\", \"corners_3d\": [[791.81, 199.26], [793.9, 198.59], [838.81, 199.13], [837.86, 199.83], [791.81, 162.83], [793.9, 163.08], [838.81, 162.88], [837.86, 162.61]]}, {\"category\": \"car\", \"corners_3d\": [[876.99, 183.52], [874.83, 183.21], [933.47, 183.41], [937.48, 183.74], [876.99, 161.77], [874.83, 162.09], [933.47, 161.88], [937.48, 161.54]]}, {\"category\": \"car\", \"corners_3d\": [[412.93, 197.0], [397.69, 197.65], [362.58, 196.93], [378.29, 196.31], [412.93, 177.01], [397.69, 177.12], [362.58, 177.0], [378.29, 176.89]]}, {\"category\": \"car\", \"corners_3d\": [[711.39, 216.79], [758.71, 216.93], [782.15, 226.12], [724.91, 225.9], [711.39, 174.61], [758.71, 174.62], [782.15, 174.99], [724.91, 174.98]]}, {\"category\": \"car\", \"corners_3d\": [[731.35, 185.18], [728.31, 184.82], [780.76, 184.84], [785.36, 185.2], [731.35, 166.28], [728.31, 166.47], [780.76, 166.46], [785.36, 166.27]]}]\n```" - } - ] - }, - { - "id": 1154, - "data_source": "Mono3DRefer", - "scene_name": "000763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[540.73, 201.03], [539.27, 201.57], [528.76, 201.57], [530.43, 201.03], [540.73, 169.35], [539.27, 169.28], [528.76, 169.28], [530.43, 169.35]]}, {\"category\": \"cyclist\", \"corners_3d\": [[791.81, 199.26], [793.9, 198.59], [838.81, 199.13], [837.86, 199.83], [791.81, 162.83], [793.9, 163.08], [838.81, 162.88], [837.86, 162.61]]}, {\"category\": \"van\", \"corners_3d\": [[1135.51, 181.87], [1124.61, 181.58], [1190.48, 181.76], [1203.9, 182.07], [1135.51, 154.8], [1124.61, 155.39], [1190.48, 155.03], [1203.9, 154.41]]}]\n```" - } - ] - }, - { - "id": 1155, - "data_source": "Mono3DRefer", - "scene_name": "000763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[791.81, 199.26], [793.9, 198.59], [838.81, 199.13], [837.86, 199.83], [791.81, 162.83], [793.9, 163.08], [838.81, 162.88], [837.86, 162.61]]}, {\"category\": \"car\", \"corners_3d\": [[876.99, 183.52], [874.83, 183.21], [933.47, 183.41], [937.48, 183.74], [876.99, 161.77], [874.83, 162.09], [933.47, 161.88], [937.48, 161.54]]}, {\"category\": \"van\", \"corners_3d\": [[1135.51, 181.87], [1124.61, 181.58], [1190.48, 181.76], [1203.9, 182.07], [1135.51, 154.8], [1124.61, 155.39], [1190.48, 155.03], [1203.9, 154.41]]}, {\"category\": \"car\", \"corners_3d\": [[412.93, 197.0], [397.69, 197.65], [362.58, 196.93], [378.29, 196.31], [412.93, 177.01], [397.69, 177.12], [362.58, 177.0], [378.29, 176.89]]}, {\"category\": \"car\", \"corners_3d\": [[711.39, 216.79], [758.71, 216.93], [782.15, 226.12], [724.91, 225.9], [711.39, 174.61], [758.71, 174.62], [782.15, 174.99], [724.91, 174.98]]}, {\"category\": \"car\", \"corners_3d\": [[731.35, 185.18], [728.31, 184.82], [780.76, 184.84], [785.36, 185.2], [731.35, 166.28], [728.31, 166.47], [780.76, 166.46], [785.36, 166.27]]}]\n```" - } - ] - }, - { - "id": 1156, - "data_source": "Mono3DRefer", - "scene_name": "000763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[540.73, 201.03], [539.27, 201.57], [528.76, 201.57], [530.43, 201.03], [540.73, 169.35], [539.27, 169.28], [528.76, 169.28], [530.43, 169.35]]}, {\"category\": \"car\", \"corners_3d\": [[876.99, 183.52], [874.83, 183.21], [933.47, 183.41], [937.48, 183.74], [876.99, 161.77], [874.83, 162.09], [933.47, 161.88], [937.48, 161.54]]}, {\"category\": \"van\", \"corners_3d\": [[1135.51, 181.87], [1124.61, 181.58], [1190.48, 181.76], [1203.9, 182.07], [1135.51, 154.8], [1124.61, 155.39], [1190.48, 155.03], [1203.9, 154.41]]}, {\"category\": \"car\", \"corners_3d\": [[412.93, 197.0], [397.69, 197.65], [362.58, 196.93], [378.29, 196.31], [412.93, 177.01], [397.69, 177.12], [362.58, 177.0], [378.29, 176.89]]}, {\"category\": \"car\", \"corners_3d\": [[711.39, 216.79], [758.71, 216.93], [782.15, 226.12], [724.91, 225.9], [711.39, 174.61], [758.71, 174.62], [782.15, 174.99], [724.91, 174.98]]}, {\"category\": \"car\", \"corners_3d\": [[731.35, 185.18], [728.31, 184.82], [780.76, 184.84], [785.36, 185.2], [731.35, 166.28], [728.31, 166.47], [780.76, 166.46], [785.36, 166.27]]}]\n```" - } - ] - }, - { - "id": 1157, - "data_source": "Mono3DRefer", - "scene_name": "000763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[540.73, 201.03], [539.27, 201.57], [528.76, 201.57], [530.43, 201.03], [540.73, 169.35], [539.27, 169.28], [528.76, 169.28], [530.43, 169.35]]}, {\"category\": \"cyclist\", \"corners_3d\": [[791.81, 199.26], [793.9, 198.59], [838.81, 199.13], [837.86, 199.83], [791.81, 162.83], [793.9, 163.08], [838.81, 162.88], [837.86, 162.61]]}, {\"category\": \"car\", \"corners_3d\": [[876.99, 183.52], [874.83, 183.21], [933.47, 183.41], [937.48, 183.74], [876.99, 161.77], [874.83, 162.09], [933.47, 161.88], [937.48, 161.54]]}, {\"category\": \"van\", \"corners_3d\": [[1135.51, 181.87], [1124.61, 181.58], [1190.48, 181.76], [1203.9, 182.07], [1135.51, 154.8], [1124.61, 155.39], [1190.48, 155.03], [1203.9, 154.41]]}, {\"category\": \"car\", \"corners_3d\": [[412.93, 197.0], [397.69, 197.65], [362.58, 196.93], [378.29, 196.31], [412.93, 177.01], [397.69, 177.12], [362.58, 177.0], [378.29, 176.89]]}, {\"category\": \"car\", \"corners_3d\": [[711.39, 216.79], [758.71, 216.93], [782.15, 226.12], [724.91, 225.9], [711.39, 174.61], [758.71, 174.62], [782.15, 174.99], [724.91, 174.98]]}, {\"category\": \"car\", \"corners_3d\": [[731.35, 185.18], [728.31, 184.82], [780.76, 184.84], [785.36, 185.2], [731.35, 166.28], [728.31, 166.47], [780.76, 166.46], [785.36, 166.27]]}]\n```" - } - ] - }, - { - "id": 1158, - "data_source": "Mono3DRefer", - "scene_name": "000769", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[394.87, 228.08], [369.3, 228.04], [375.29, 226.74], [400.26, 226.78], [394.87, 177.94], [369.3, 177.94], [375.29, 177.82], [400.26, 177.82]]}]\n```" - } - ] - }, - { - "id": 1159, - "data_source": "Mono3DRefer", - "scene_name": "000770", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[711.7, 200.58], [703.74, 200.26], [725.45, 199.9], [733.61, 200.21], [711.7, 154.6], [703.74, 155.02], [725.45, 155.48], [733.61, 155.08]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[722.22, 201.42], [717.8, 200.98], [741.39, 200.89], [746.3, 201.33], [722.22, 152.27], [717.8, 152.86], [741.39, 152.98], [746.3, 152.4]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[730.14, 264.7], [763.5, 267.77], [732.14, 273.53], [697.84, 270.05], [730.14, 163.3], [763.5, 162.67], [732.14, 161.49], [697.84, 162.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[894.17, 232.39], [919.67, 234.43], [907.45, 235.04], [882.15, 232.96], [894.17, 149.06], [919.67, 147.82], [907.45, 147.45], [882.15, 148.71]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[696.19, 206.34], [676.33, 206.06], [682.26, 205.34], [701.62, 205.6], [696.19, 158.32], [676.33, 158.55], [682.26, 159.17], [701.62, 158.95]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[676.1, 204.17], [690.64, 204.41], [686.47, 204.77], [671.75, 204.52], [676.1, 162.74], [690.64, 162.56], [686.47, 162.3], [671.75, 162.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[692.81, 199.92], [706.58, 200.12], [702.54, 200.39], [688.62, 200.19], [692.81, 158.23], [706.58, 158.0], [702.54, 157.69], [688.62, 157.93]]}]\n```" - } - ] - }, - { - "id": 1160, - "data_source": "Mono3DRefer", - "scene_name": "000770", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[834.5, 254.98], [854.21, 255.05], [890.51, 266.9], [867.64, 266.81], [834.5, 151.72], [854.21, 151.7], [890.51, 147.12], [867.64, 147.15]]}, {\"category\": \"cyclist\", \"corners_3d\": [[517.47, 217.57], [533.57, 217.84], [509.41, 220.95], [492.17, 220.63], [517.47, 166.25], [533.57, 166.14], [509.41, 164.95], [492.17, 165.07]]}]\n```" - } - ] - }, - { - "id": 1161, - "data_source": "Mono3DRefer", - "scene_name": "000770", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[711.7, 200.58], [703.74, 200.26], [725.45, 199.9], [733.61, 200.21], [711.7, 154.6], [703.74, 155.02], [725.45, 155.48], [733.61, 155.08]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[722.22, 201.42], [717.8, 200.98], [741.39, 200.89], [746.3, 201.33], [722.22, 152.27], [717.8, 152.86], [741.39, 152.98], [746.3, 152.4]]}, {\"category\": \"cyclist\", \"corners_3d\": [[834.5, 254.98], [854.21, 255.05], [890.51, 266.9], [867.64, 266.81], [834.5, 151.72], [854.21, 151.7], [890.51, 147.12], [867.64, 147.15]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[730.14, 264.7], [763.5, 267.77], [732.14, 273.53], [697.84, 270.05], [730.14, 163.3], [763.5, 162.67], [732.14, 161.49], [697.84, 162.2]]}, {\"category\": \"cyclist\", \"corners_3d\": [[517.47, 217.57], [533.57, 217.84], [509.41, 220.95], [492.17, 220.63], [517.47, 166.25], [533.57, 166.14], [509.41, 164.95], [492.17, 165.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[894.17, 232.39], [919.67, 234.43], [907.45, 235.04], [882.15, 232.96], [894.17, 149.06], [919.67, 147.82], [907.45, 147.45], [882.15, 148.71]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[696.19, 206.34], [676.33, 206.06], [682.26, 205.34], [701.62, 205.6], [696.19, 158.32], [676.33, 158.55], [682.26, 159.17], [701.62, 158.95]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[676.1, 204.17], [690.64, 204.41], [686.47, 204.77], [671.75, 204.52], [676.1, 162.74], [690.64, 162.56], [686.47, 162.3], [671.75, 162.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[692.81, 199.92], [706.58, 200.12], [702.54, 200.39], [688.62, 200.19], [692.81, 158.23], [706.58, 158.0], [702.54, 157.69], [688.62, 157.93]]}]\n```" - } - ] - }, - { - "id": 1162, - "data_source": "Mono3DRefer", - "scene_name": "000776", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[835.82, 201.83], [801.07, 201.42], [805.17, 198.42], [836.27, 198.74], [835.82, 161.48], [801.07, 161.64], [805.17, 162.82], [836.27, 162.69]]}]\n```" - } - ] - }, - { - "id": 1163, - "data_source": "Mono3DRefer", - "scene_name": "000776", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[644.42, 276.55], [565.78, 272.51], [638.05, 249.01], [699.8, 251.35], [644.42, 192.67], [565.78, 191.9], [638.05, 187.41], [699.8, 187.85]]}, {\"category\": \"car\", \"corners_3d\": [[847.5, 232.1], [788.78, 230.35], [805.25, 222.24], [855.89, 223.52], [847.5, 183.77], [788.78, 183.45], [805.25, 181.95], [855.89, 182.19]]}, {\"category\": \"car\", \"corners_3d\": [[718.06, 245.28], [653.72, 242.85], [696.76, 230.47], [750.62, 232.11], [718.06, 188.13], [653.72, 187.62], [696.76, 185.01], [750.62, 185.35]]}, {\"category\": \"car\", \"corners_3d\": [[768.59, 226.27], [728.86, 225.17], [750.15, 218.86], [785.4, 219.71], [768.59, 187.21], [728.86, 186.91], [750.15, 185.21], [785.4, 185.44]]}, {\"category\": \"car\", \"corners_3d\": [[821.74, 207.8], [789.66, 207.36], [792.55, 205.0], [822.46, 205.39], [821.74, 180.59], [789.66, 180.5], [792.55, 179.97], [822.46, 180.06]]}]\n```" - } - ] - }, - { - "id": 1164, - "data_source": "Mono3DRefer", - "scene_name": "000776", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[644.42, 276.55], [565.78, 272.51], [638.05, 249.01], [699.8, 251.35], [644.42, 192.67], [565.78, 191.9], [638.05, 187.41], [699.8, 187.85]]}, {\"category\": \"car\", \"corners_3d\": [[847.5, 232.1], [788.78, 230.35], [805.25, 222.24], [855.89, 223.52], [847.5, 183.77], [788.78, 183.45], [805.25, 181.95], [855.89, 182.19]]}, {\"category\": \"car\", \"corners_3d\": [[718.06, 245.28], [653.72, 242.85], [696.76, 230.47], [750.62, 232.11], [718.06, 188.13], [653.72, 187.62], [696.76, 185.01], [750.62, 185.35]]}, {\"category\": \"car\", \"corners_3d\": [[768.59, 226.27], [728.86, 225.17], [750.15, 218.86], [785.4, 219.71], [768.59, 187.21], [728.86, 186.91], [750.15, 185.21], [785.4, 185.44]]}, {\"category\": \"van\", \"corners_3d\": [[835.82, 201.83], [801.07, 201.42], [805.17, 198.42], [836.27, 198.74], [835.82, 161.48], [801.07, 161.64], [805.17, 162.82], [836.27, 162.69]]}, {\"category\": \"car\", \"corners_3d\": [[821.74, 207.8], [789.66, 207.36], [792.55, 205.0], [822.46, 205.39], [821.74, 180.59], [789.66, 180.5], [792.55, 179.97], [822.46, 180.06]]}]\n```" - } - ] - }, - { - "id": 1165, - "data_source": "Mono3DRefer", - "scene_name": "000780", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[425.91, 244.14], [359.87, 243.95], [407.53, 231.44], [462.03, 231.57], [425.91, 175.74], [359.87, 175.73], [407.53, 175.23], [462.03, 175.23]]}, {\"category\": \"car\", \"corners_3d\": [[790.01, 306.45], [921.04, 305.73], [1127.05, 388.05], [915.94, 389.93], [790.01, 171.1], [921.04, 171.11], [1127.05, 170.03], [915.94, 170.0]]}, {\"category\": \"car\", \"corners_3d\": [[709.33, 253.74], [792.86, 253.75], [863.66, 285.09], [747.76, 285.08], [709.33, 162.92], [792.86, 162.92], [863.66, 159.07], [747.76, 159.07]]}, {\"category\": \"car\", \"corners_3d\": [[679.95, 229.7], [742.35, 229.7], [781.23, 246.41], [700.48, 246.41], [679.95, 162.13], [742.35, 162.13], [781.23, 158.98], [700.48, 158.98]]}, {\"category\": \"car\", \"corners_3d\": [[671.74, 218.71], [719.27, 218.81], [735.11, 227.12], [678.96, 227.0], [671.74, 171.67], [719.27, 171.67], [735.11, 171.46], [678.96, 171.46]]}, {\"category\": \"car\", \"corners_3d\": [[555.34, 195.5], [531.34, 195.49], [537.72, 193.79], [559.92, 193.8], [555.34, 167.03], [531.34, 167.04], [537.72, 167.47], [559.92, 167.47]]}, {\"category\": \"car\", \"corners_3d\": [[480.07, 221.35], [437.46, 221.27], [462.72, 214.94], [499.79, 215.0], [480.07, 177.25], [437.46, 177.25], [462.72, 176.67], [499.79, 176.68]]}, {\"category\": \"car\", \"corners_3d\": [[-1339.34, 982.26], [-2004.25, 958.6], [-113.19, 394.67], [86.28, 396.52], [-1339.34, 308.35], [-2004.25, 304.39], [-113.19, 209.99], [86.28, 210.3]]}, {\"category\": \"car\", \"corners_3d\": [[153.1, 365.91], [-11.31, 365.95], [208.83, 297.42], [314.87, 297.41], [153.1, 201.62], [-11.31, 201.63], [208.83, 191.41], [314.87, 191.41]]}, {\"category\": \"car\", \"corners_3d\": [[346.85, 284.23], [248.33, 284.24], [344.13, 254.66], [416.47, 254.65], [346.85, 187.6], [248.33, 187.6], [344.13, 183.69], [416.47, 183.69]]}]\n```" - } - ] - }, - { - "id": 1166, - "data_source": "Mono3DRefer", - "scene_name": "000780", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[279.72, 351.45], [227.03, 351.46], [302.13, 316.34], [344.45, 316.34], [279.72, 164.56], [227.03, 164.56], [302.13, 166.19], [344.45, 166.19]]}]\n```" - } - ] - }, - { - "id": 1167, - "data_source": "Mono3DRefer", - "scene_name": "000780", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1076.1, 306.57], [1151.05, 306.58], [1188.82, 315.92], [1108.63, 315.91], [1076.1, 127.87], [1151.05, 127.87], [1188.82, 124.73], [1108.63, 124.73]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[663.92, 197.78], [659.02, 198.0], [642.91, 197.54], [647.86, 197.33], [663.92, 160.77], [659.02, 160.67], [642.91, 160.89], [647.86, 160.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[671.23, 197.61], [665.81, 197.87], [651.71, 197.5], [657.18, 197.25], [671.23, 161.99], [665.81, 161.88], [651.71, 162.04], [657.18, 162.15]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[611.88, 190.73], [612.19, 190.9], [599.39, 190.91], [599.21, 190.74], [611.88, 166.98], [612.19, 166.92], [599.39, 166.92], [599.21, 166.98]]}]\n```" - } - ] - }, - { - "id": 1168, - "data_source": "Mono3DRefer", - "scene_name": "000780", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[425.91, 244.14], [359.87, 243.95], [407.53, 231.44], [462.03, 231.57], [425.91, 175.74], [359.87, 175.73], [407.53, 175.23], [462.03, 175.23]]}, {\"category\": \"cyclist\", \"corners_3d\": [[279.72, 351.45], [227.03, 351.46], [302.13, 316.34], [344.45, 316.34], [279.72, 164.56], [227.03, 164.56], [302.13, 166.19], [344.45, 166.19]]}, {\"category\": \"car\", \"corners_3d\": [[790.01, 306.45], [921.04, 305.73], [1127.05, 388.05], [915.94, 389.93], [790.01, 171.1], [921.04, 171.11], [1127.05, 170.03], [915.94, 170.0]]}, {\"category\": \"car\", \"corners_3d\": [[709.33, 253.74], [792.86, 253.75], [863.66, 285.09], [747.76, 285.08], [709.33, 162.92], [792.86, 162.92], [863.66, 159.07], [747.76, 159.07]]}, {\"category\": \"car\", \"corners_3d\": [[679.95, 229.7], [742.35, 229.7], [781.23, 246.41], [700.48, 246.41], [679.95, 162.13], [742.35, 162.13], [781.23, 158.98], [700.48, 158.98]]}, {\"category\": \"car\", \"corners_3d\": [[671.74, 218.71], [719.27, 218.81], [735.11, 227.12], [678.96, 227.0], [671.74, 171.67], [719.27, 171.67], [735.11, 171.46], [678.96, 171.46]]}, {\"category\": \"car\", \"corners_3d\": [[555.34, 195.5], [531.34, 195.49], [537.72, 193.79], [559.92, 193.8], [555.34, 167.03], [531.34, 167.04], [537.72, 167.47], [559.92, 167.47]]}, {\"category\": \"car\", \"corners_3d\": [[480.07, 221.35], [437.46, 221.27], [462.72, 214.94], [499.79, 215.0], [480.07, 177.25], [437.46, 177.25], [462.72, 176.67], [499.79, 176.68]]}, {\"category\": \"car\", \"corners_3d\": [[-1339.34, 982.26], [-2004.25, 958.6], [-113.19, 394.67], [86.28, 396.52], [-1339.34, 308.35], [-2004.25, 304.39], [-113.19, 209.99], [86.28, 210.3]]}, {\"category\": \"car\", \"corners_3d\": [[153.1, 365.91], [-11.31, 365.95], [208.83, 297.42], [314.87, 297.41], [153.1, 201.62], [-11.31, 201.63], [208.83, 191.41], [314.87, 191.41]]}, {\"category\": \"car\", \"corners_3d\": [[346.85, 284.23], [248.33, 284.24], [344.13, 254.66], [416.47, 254.65], [346.85, 187.6], [248.33, 187.6], [344.13, 183.69], [416.47, 183.69]]}]\n```" - } - ] - }, - { - "id": 1169, - "data_source": "Mono3DRefer", - "scene_name": "000780", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[425.91, 244.14], [359.87, 243.95], [407.53, 231.44], [462.03, 231.57], [425.91, 175.74], [359.87, 175.73], [407.53, 175.23], [462.03, 175.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1076.1, 306.57], [1151.05, 306.58], [1188.82, 315.92], [1108.63, 315.91], [1076.1, 127.87], [1151.05, 127.87], [1188.82, 124.73], [1108.63, 124.73]]}, {\"category\": \"car\", \"corners_3d\": [[790.01, 306.45], [921.04, 305.73], [1127.05, 388.05], [915.94, 389.93], [790.01, 171.1], [921.04, 171.11], [1127.05, 170.03], [915.94, 170.0]]}, {\"category\": \"car\", \"corners_3d\": [[709.33, 253.74], [792.86, 253.75], [863.66, 285.09], [747.76, 285.08], [709.33, 162.92], [792.86, 162.92], [863.66, 159.07], [747.76, 159.07]]}, {\"category\": \"car\", \"corners_3d\": [[679.95, 229.7], [742.35, 229.7], [781.23, 246.41], [700.48, 246.41], [679.95, 162.13], [742.35, 162.13], [781.23, 158.98], [700.48, 158.98]]}, {\"category\": \"car\", \"corners_3d\": [[671.74, 218.71], [719.27, 218.81], [735.11, 227.12], [678.96, 227.0], [671.74, 171.67], [719.27, 171.67], [735.11, 171.46], [678.96, 171.46]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[663.92, 197.78], [659.02, 198.0], [642.91, 197.54], [647.86, 197.33], [663.92, 160.77], [659.02, 160.67], [642.91, 160.89], [647.86, 160.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[671.23, 197.61], [665.81, 197.87], [651.71, 197.5], [657.18, 197.25], [671.23, 161.99], [665.81, 161.88], [651.71, 162.04], [657.18, 162.15]]}, {\"category\": \"car\", \"corners_3d\": [[555.34, 195.5], [531.34, 195.49], [537.72, 193.79], [559.92, 193.8], [555.34, 167.03], [531.34, 167.04], [537.72, 167.47], [559.92, 167.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[611.88, 190.73], [612.19, 190.9], [599.39, 190.91], [599.21, 190.74], [611.88, 166.98], [612.19, 166.92], [599.39, 166.92], [599.21, 166.98]]}, {\"category\": \"car\", \"corners_3d\": [[480.07, 221.35], [437.46, 221.27], [462.72, 214.94], [499.79, 215.0], [480.07, 177.25], [437.46, 177.25], [462.72, 176.67], [499.79, 176.68]]}, {\"category\": \"car\", \"corners_3d\": [[-1339.34, 982.26], [-2004.25, 958.6], [-113.19, 394.67], [86.28, 396.52], [-1339.34, 308.35], [-2004.25, 304.39], [-113.19, 209.99], [86.28, 210.3]]}, {\"category\": \"car\", \"corners_3d\": [[153.1, 365.91], [-11.31, 365.95], [208.83, 297.42], [314.87, 297.41], [153.1, 201.62], [-11.31, 201.63], [208.83, 191.41], [314.87, 191.41]]}, {\"category\": \"car\", \"corners_3d\": [[346.85, 284.23], [248.33, 284.24], [344.13, 254.66], [416.47, 254.65], [346.85, 187.6], [248.33, 187.6], [344.13, 183.69], [416.47, 183.69]]}]\n```" - } - ] - }, - { - "id": 1170, - "data_source": "Mono3DRefer", - "scene_name": "000780", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1076.1, 306.57], [1151.05, 306.58], [1188.82, 315.92], [1108.63, 315.91], [1076.1, 127.87], [1151.05, 127.87], [1188.82, 124.73], [1108.63, 124.73]]}, {\"category\": \"cyclist\", \"corners_3d\": [[279.72, 351.45], [227.03, 351.46], [302.13, 316.34], [344.45, 316.34], [279.72, 164.56], [227.03, 164.56], [302.13, 166.19], [344.45, 166.19]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[663.92, 197.78], [659.02, 198.0], [642.91, 197.54], [647.86, 197.33], [663.92, 160.77], [659.02, 160.67], [642.91, 160.89], [647.86, 160.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[671.23, 197.61], [665.81, 197.87], [651.71, 197.5], [657.18, 197.25], [671.23, 161.99], [665.81, 161.88], [651.71, 162.04], [657.18, 162.15]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[611.88, 190.73], [612.19, 190.9], [599.39, 190.91], [599.21, 190.74], [611.88, 166.98], [612.19, 166.92], [599.39, 166.92], [599.21, 166.98]]}]\n```" - } - ] - }, - { - "id": 1171, - "data_source": "Mono3DRefer", - "scene_name": "000780", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[425.91, 244.14], [359.87, 243.95], [407.53, 231.44], [462.03, 231.57], [425.91, 175.74], [359.87, 175.73], [407.53, 175.23], [462.03, 175.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1076.1, 306.57], [1151.05, 306.58], [1188.82, 315.92], [1108.63, 315.91], [1076.1, 127.87], [1151.05, 127.87], [1188.82, 124.73], [1108.63, 124.73]]}, {\"category\": \"cyclist\", \"corners_3d\": [[279.72, 351.45], [227.03, 351.46], [302.13, 316.34], [344.45, 316.34], [279.72, 164.56], [227.03, 164.56], [302.13, 166.19], [344.45, 166.19]]}, {\"category\": \"car\", \"corners_3d\": [[790.01, 306.45], [921.04, 305.73], [1127.05, 388.05], [915.94, 389.93], [790.01, 171.1], [921.04, 171.11], [1127.05, 170.03], [915.94, 170.0]]}, {\"category\": \"car\", \"corners_3d\": [[709.33, 253.74], [792.86, 253.75], [863.66, 285.09], [747.76, 285.08], [709.33, 162.92], [792.86, 162.92], [863.66, 159.07], [747.76, 159.07]]}, {\"category\": \"car\", \"corners_3d\": [[679.95, 229.7], [742.35, 229.7], [781.23, 246.41], [700.48, 246.41], [679.95, 162.13], [742.35, 162.13], [781.23, 158.98], [700.48, 158.98]]}, {\"category\": \"car\", \"corners_3d\": [[671.74, 218.71], [719.27, 218.81], [735.11, 227.12], [678.96, 227.0], [671.74, 171.67], [719.27, 171.67], [735.11, 171.46], [678.96, 171.46]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[663.92, 197.78], [659.02, 198.0], [642.91, 197.54], [647.86, 197.33], [663.92, 160.77], [659.02, 160.67], [642.91, 160.89], [647.86, 160.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[671.23, 197.61], [665.81, 197.87], [651.71, 197.5], [657.18, 197.25], [671.23, 161.99], [665.81, 161.88], [651.71, 162.04], [657.18, 162.15]]}, {\"category\": \"car\", \"corners_3d\": [[555.34, 195.5], [531.34, 195.49], [537.72, 193.79], [559.92, 193.8], [555.34, 167.03], [531.34, 167.04], [537.72, 167.47], [559.92, 167.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[611.88, 190.73], [612.19, 190.9], [599.39, 190.91], [599.21, 190.74], [611.88, 166.98], [612.19, 166.92], [599.39, 166.92], [599.21, 166.98]]}, {\"category\": \"car\", \"corners_3d\": [[480.07, 221.35], [437.46, 221.27], [462.72, 214.94], [499.79, 215.0], [480.07, 177.25], [437.46, 177.25], [462.72, 176.67], [499.79, 176.68]]}, {\"category\": \"car\", \"corners_3d\": [[-1339.34, 982.26], [-2004.25, 958.6], [-113.19, 394.67], [86.28, 396.52], [-1339.34, 308.35], [-2004.25, 304.39], [-113.19, 209.99], [86.28, 210.3]]}, {\"category\": \"car\", \"corners_3d\": [[153.1, 365.91], [-11.31, 365.95], [208.83, 297.42], [314.87, 297.41], [153.1, 201.62], [-11.31, 201.63], [208.83, 191.41], [314.87, 191.41]]}, {\"category\": \"car\", \"corners_3d\": [[346.85, 284.23], [248.33, 284.24], [344.13, 254.66], [416.47, 254.65], [346.85, 187.6], [248.33, 187.6], [344.13, 183.69], [416.47, 183.69]]}]\n```" - } - ] - }, - { - "id": 1172, - "data_source": "Mono3DRefer", - "scene_name": "000781", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[435.66, 224.11], [420.79, 227.62], [292.28, 227.03], [315.15, 223.59], [435.66, 186.63], [420.79, 187.57], [292.28, 187.41], [315.15, 186.49]]}]\n```" - } - ] - }, - { - "id": 1173, - "data_source": "Mono3DRefer", - "scene_name": "000783", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[659.44, 199.31], [692.21, 199.38], [695.93, 201.93], [660.01, 201.84], [659.44, 168.24], [692.21, 168.23], [695.93, 167.79], [660.01, 167.8]]}, {\"category\": \"car\", \"corners_3d\": [[281.79, 351.96], [113.89, 351.99], [288.86, 288.68], [397.41, 288.67], [281.79, 188.26], [113.89, 188.26], [288.86, 182.82], [397.41, 182.82]]}]\n```" - } - ] - }, - { - "id": 1174, - "data_source": "Mono3DRefer", - "scene_name": "000786", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[550.74, 217.07], [513.23, 217.07], [524.14, 212.03], [557.37, 212.03], [550.74, 184.12], [513.23, 184.12], [524.14, 182.84], [557.37, 182.84]]}, {\"category\": \"car\", \"corners_3d\": [[574.66, 199.27], [553.39, 199.26], [557.13, 197.69], [577.13, 197.7], [574.66, 181.24], [553.39, 181.24], [557.13, 180.74], [577.13, 180.74]]}]\n```" - } - ] - }, - { - "id": 1175, - "data_source": "Mono3DRefer", - "scene_name": "000788", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[813.55, 233.37], [819.59, 235.1], [782.71, 235.13], [777.7, 233.4], [813.55, 166.03], [819.59, 165.84], [782.71, 165.83], [777.7, 166.03]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[757.23, 213.14], [770.37, 213.11], [775.56, 214.22], [762.06, 214.25], [757.23, 170.84], [770.37, 170.84], [775.56, 170.79], [762.06, 170.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[744.59, 213.06], [759.7, 213.01], [765.0, 214.11], [749.48, 214.16], [744.59, 166.56], [759.7, 166.57], [765.0, 166.4], [749.48, 166.39]]}]\n```" - } - ] - }, - { - "id": 1176, - "data_source": "Mono3DRefer", - "scene_name": "000788", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[125.91, 364.29], [-45.69, 365.67], [183.73, 295.79], [292.38, 295.23], [125.91, 166.43], [-45.69, 166.39], [183.73, 168.73], [292.38, 168.75]]}]\n```" - } - ] - }, - { - "id": 1177, - "data_source": "Mono3DRefer", - "scene_name": "000788", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[125.91, 364.29], [-45.69, 365.67], [183.73, 295.79], [292.38, 295.23], [125.91, 166.43], [-45.69, 166.39], [183.73, 168.73], [292.38, 168.75]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[813.55, 233.37], [819.59, 235.1], [782.71, 235.13], [777.7, 233.4], [813.55, 166.03], [819.59, 165.84], [782.71, 165.83], [777.7, 166.03]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[757.23, 213.14], [770.37, 213.11], [775.56, 214.22], [762.06, 214.25], [757.23, 170.84], [770.37, 170.84], [775.56, 170.79], [762.06, 170.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[744.59, 213.06], [759.7, 213.01], [765.0, 214.11], [749.48, 214.16], [744.59, 166.56], [759.7, 166.57], [765.0, 166.4], [749.48, 166.39]]}]\n```" - } - ] - }, - { - "id": 1178, - "data_source": "Mono3DRefer", - "scene_name": "000791", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[457.34, 245.41], [520.06, 246.75], [442.54, 270.99], [361.61, 268.64], [457.34, 188.58], [520.06, 188.87], [442.54, 194.13], [361.61, 193.62]]}, {\"category\": \"car\", \"corners_3d\": [[358.02, 244.39], [412.61, 245.91], [297.77, 267.88], [230.29, 265.32], [358.02, 169.92], [412.61, 169.86], [297.77, 168.96], [230.29, 169.06]]}, {\"category\": \"car\", \"corners_3d\": [[630.72, 217.51], [680.05, 217.97], [672.37, 226.82], [613.58, 226.16], [630.72, 174.94], [680.05, 174.96], [672.37, 175.38], [613.58, 175.35]]}, {\"category\": \"car\", \"corners_3d\": [[1013.81, 240.68], [1096.64, 242.1], [1222.22, 267.66], [1106.17, 265.01], [1013.81, 175.43], [1096.64, 175.49], [1222.22, 176.46], [1106.17, 176.36]]}, {\"category\": \"car\", \"corners_3d\": [[797.69, 201.8], [837.5, 202.17], [845.06, 206.53], [799.3, 206.05], [797.69, 165.87], [837.5, 165.78], [845.06, 164.73], [799.3, 164.85]]}, {\"category\": \"car\", \"corners_3d\": [[1126.54, 236.35], [1214.23, 237.9], [1372.2, 261.52], [1248.61, 258.66], [1126.54, 165.01], [1214.23, 164.82], [1372.2, 161.9], [1248.61, 162.25]]}, {\"category\": \"car\", \"corners_3d\": [[188.4, 244.95], [236.02, 246.11], [85.76, 266.96], [27.92, 265.05], [188.4, 192.91], [236.02, 193.23], [85.76, 199.03], [27.92, 198.5]]}, {\"category\": \"car\", \"corners_3d\": [[44.86, 246.28], [89.13, 247.24], [-34.86, 261.38], [-85.51, 260.02], [44.86, 199.76], [89.13, 200.11], [-34.86, 205.29], [-85.51, 204.79]]}]\n```" - } - ] - }, - { - "id": 1179, - "data_source": "Mono3DRefer", - "scene_name": "000791", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[839.47, 199.91], [791.98, 199.59], [784.0, 194.77], [822.78, 194.98], [839.47, 126.72], [791.98, 127.27], [784.0, 135.48], [822.78, 135.12]]}]\n```" - } - ] - }, - { - "id": 1180, - "data_source": "Mono3DRefer", - "scene_name": "000791", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[727.44, 244.51], [805.08, 246.18], [816.99, 272.19], [712.29, 269.15], [727.44, 162.0], [805.08, 161.75], [816.99, 157.81], [712.29, 158.27]]}, {\"category\": \"van\", \"corners_3d\": [[901.21, 198.55], [860.88, 198.3], [844.75, 194.9], [879.5, 195.08], [901.21, 146.58], [860.88, 146.83], [844.75, 150.31], [879.5, 150.13]]}]\n```" - } - ] - }, - { - "id": 1181, - "data_source": "Mono3DRefer", - "scene_name": "000791", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[457.34, 245.41], [520.06, 246.75], [442.54, 270.99], [361.61, 268.64], [457.34, 188.58], [520.06, 188.87], [442.54, 194.13], [361.61, 193.62]]}, {\"category\": \"car\", \"corners_3d\": [[358.02, 244.39], [412.61, 245.91], [297.77, 267.88], [230.29, 265.32], [358.02, 169.92], [412.61, 169.86], [297.77, 168.96], [230.29, 169.06]]}, {\"category\": \"car\", \"corners_3d\": [[630.72, 217.51], [680.05, 217.97], [672.37, 226.82], [613.58, 226.16], [630.72, 174.94], [680.05, 174.96], [672.37, 175.38], [613.58, 175.35]]}, {\"category\": \"car\", \"corners_3d\": [[1013.81, 240.68], [1096.64, 242.1], [1222.22, 267.66], [1106.17, 265.01], [1013.81, 175.43], [1096.64, 175.49], [1222.22, 176.46], [1106.17, 176.36]]}, {\"category\": \"car\", \"corners_3d\": [[797.69, 201.8], [837.5, 202.17], [845.06, 206.53], [799.3, 206.05], [797.69, 165.87], [837.5, 165.78], [845.06, 164.73], [799.3, 164.85]]}, {\"category\": \"car\", \"corners_3d\": [[1126.54, 236.35], [1214.23, 237.9], [1372.2, 261.52], [1248.61, 258.66], [1126.54, 165.01], [1214.23, 164.82], [1372.2, 161.9], [1248.61, 162.25]]}, {\"category\": \"truck\", \"corners_3d\": [[839.47, 199.91], [791.98, 199.59], [784.0, 194.77], [822.78, 194.98], [839.47, 126.72], [791.98, 127.27], [784.0, 135.48], [822.78, 135.12]]}, {\"category\": \"car\", \"corners_3d\": [[188.4, 244.95], [236.02, 246.11], [85.76, 266.96], [27.92, 265.05], [188.4, 192.91], [236.02, 193.23], [85.76, 199.03], [27.92, 198.5]]}, {\"category\": \"car\", \"corners_3d\": [[44.86, 246.28], [89.13, 247.24], [-34.86, 261.38], [-85.51, 260.02], [44.86, 199.76], [89.13, 200.11], [-34.86, 205.29], [-85.51, 204.79]]}]\n```" - } - ] - }, - { - "id": 1182, - "data_source": "Mono3DRefer", - "scene_name": "000791", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[727.44, 244.51], [805.08, 246.18], [816.99, 272.19], [712.29, 269.15], [727.44, 162.0], [805.08, 161.75], [816.99, 157.81], [712.29, 158.27]]}, {\"category\": \"car\", \"corners_3d\": [[457.34, 245.41], [520.06, 246.75], [442.54, 270.99], [361.61, 268.64], [457.34, 188.58], [520.06, 188.87], [442.54, 194.13], [361.61, 193.62]]}, {\"category\": \"car\", \"corners_3d\": [[358.02, 244.39], [412.61, 245.91], [297.77, 267.88], [230.29, 265.32], [358.02, 169.92], [412.61, 169.86], [297.77, 168.96], [230.29, 169.06]]}, {\"category\": \"car\", \"corners_3d\": [[630.72, 217.51], [680.05, 217.97], [672.37, 226.82], [613.58, 226.16], [630.72, 174.94], [680.05, 174.96], [672.37, 175.38], [613.58, 175.35]]}, {\"category\": \"car\", \"corners_3d\": [[1013.81, 240.68], [1096.64, 242.1], [1222.22, 267.66], [1106.17, 265.01], [1013.81, 175.43], [1096.64, 175.49], [1222.22, 176.46], [1106.17, 176.36]]}, {\"category\": \"car\", \"corners_3d\": [[797.69, 201.8], [837.5, 202.17], [845.06, 206.53], [799.3, 206.05], [797.69, 165.87], [837.5, 165.78], [845.06, 164.73], [799.3, 164.85]]}, {\"category\": \"van\", \"corners_3d\": [[901.21, 198.55], [860.88, 198.3], [844.75, 194.9], [879.5, 195.08], [901.21, 146.58], [860.88, 146.83], [844.75, 150.31], [879.5, 150.13]]}, {\"category\": \"car\", \"corners_3d\": [[1126.54, 236.35], [1214.23, 237.9], [1372.2, 261.52], [1248.61, 258.66], [1126.54, 165.01], [1214.23, 164.82], [1372.2, 161.9], [1248.61, 162.25]]}, {\"category\": \"car\", \"corners_3d\": [[188.4, 244.95], [236.02, 246.11], [85.76, 266.96], [27.92, 265.05], [188.4, 192.91], [236.02, 193.23], [85.76, 199.03], [27.92, 198.5]]}, {\"category\": \"car\", \"corners_3d\": [[44.86, 246.28], [89.13, 247.24], [-34.86, 261.38], [-85.51, 260.02], [44.86, 199.76], [89.13, 200.11], [-34.86, 205.29], [-85.51, 204.79]]}]\n```" - } - ] - }, - { - "id": 1183, - "data_source": "Mono3DRefer", - "scene_name": "000791", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[727.44, 244.51], [805.08, 246.18], [816.99, 272.19], [712.29, 269.15], [727.44, 162.0], [805.08, 161.75], [816.99, 157.81], [712.29, 158.27]]}, {\"category\": \"van\", \"corners_3d\": [[901.21, 198.55], [860.88, 198.3], [844.75, 194.9], [879.5, 195.08], [901.21, 146.58], [860.88, 146.83], [844.75, 150.31], [879.5, 150.13]]}, {\"category\": \"truck\", \"corners_3d\": [[839.47, 199.91], [791.98, 199.59], [784.0, 194.77], [822.78, 194.98], [839.47, 126.72], [791.98, 127.27], [784.0, 135.48], [822.78, 135.12]]}]\n```" - } - ] - }, - { - "id": 1184, - "data_source": "Mono3DRefer", - "scene_name": "000791", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[727.44, 244.51], [805.08, 246.18], [816.99, 272.19], [712.29, 269.15], [727.44, 162.0], [805.08, 161.75], [816.99, 157.81], [712.29, 158.27]]}, {\"category\": \"car\", \"corners_3d\": [[457.34, 245.41], [520.06, 246.75], [442.54, 270.99], [361.61, 268.64], [457.34, 188.58], [520.06, 188.87], [442.54, 194.13], [361.61, 193.62]]}, {\"category\": \"car\", \"corners_3d\": [[358.02, 244.39], [412.61, 245.91], [297.77, 267.88], [230.29, 265.32], [358.02, 169.92], [412.61, 169.86], [297.77, 168.96], [230.29, 169.06]]}, {\"category\": \"car\", \"corners_3d\": [[630.72, 217.51], [680.05, 217.97], [672.37, 226.82], [613.58, 226.16], [630.72, 174.94], [680.05, 174.96], [672.37, 175.38], [613.58, 175.35]]}, {\"category\": \"car\", \"corners_3d\": [[1013.81, 240.68], [1096.64, 242.1], [1222.22, 267.66], [1106.17, 265.01], [1013.81, 175.43], [1096.64, 175.49], [1222.22, 176.46], [1106.17, 176.36]]}, {\"category\": \"car\", \"corners_3d\": [[797.69, 201.8], [837.5, 202.17], [845.06, 206.53], [799.3, 206.05], [797.69, 165.87], [837.5, 165.78], [845.06, 164.73], [799.3, 164.85]]}, {\"category\": \"van\", \"corners_3d\": [[901.21, 198.55], [860.88, 198.3], [844.75, 194.9], [879.5, 195.08], [901.21, 146.58], [860.88, 146.83], [844.75, 150.31], [879.5, 150.13]]}, {\"category\": \"car\", \"corners_3d\": [[1126.54, 236.35], [1214.23, 237.9], [1372.2, 261.52], [1248.61, 258.66], [1126.54, 165.01], [1214.23, 164.82], [1372.2, 161.9], [1248.61, 162.25]]}, {\"category\": \"truck\", \"corners_3d\": [[839.47, 199.91], [791.98, 199.59], [784.0, 194.77], [822.78, 194.98], [839.47, 126.72], [791.98, 127.27], [784.0, 135.48], [822.78, 135.12]]}, {\"category\": \"car\", \"corners_3d\": [[188.4, 244.95], [236.02, 246.11], [85.76, 266.96], [27.92, 265.05], [188.4, 192.91], [236.02, 193.23], [85.76, 199.03], [27.92, 198.5]]}, {\"category\": \"car\", \"corners_3d\": [[44.86, 246.28], [89.13, 247.24], [-34.86, 261.38], [-85.51, 260.02], [44.86, 199.76], [89.13, 200.11], [-34.86, 205.29], [-85.51, 204.79]]}]\n```" - } - ] - }, - { - "id": 1185, - "data_source": "Mono3DRefer", - "scene_name": "000793", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[331.15, 267.14], [216.99, 267.42], [314.98, 244.57], [398.62, 244.42], [331.15, 156.52], [216.99, 156.44], [314.98, 163.11], [398.62, 163.15]]}]\n```" - } - ] - }, - { - "id": 1186, - "data_source": "Mono3DRefer", - "scene_name": "000793", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[495.46, 226.64], [550.1, 226.78], [530.94, 237.67], [463.29, 237.46], [495.46, 175.94], [550.1, 175.93], [530.94, 174.58], [463.29, 174.61]]}, {\"category\": \"car\", \"corners_3d\": [[532.47, 208.75], [568.06, 208.75], [564.07, 212.05], [524.17, 212.05], [532.47, 176.16], [568.06, 176.16], [564.07, 175.51], [524.17, 175.51]]}, {\"category\": \"car\", \"corners_3d\": [[605.75, 201.71], [639.58, 201.75], [640.97, 204.6], [602.38, 204.55], [605.75, 170.96], [639.58, 170.94], [640.97, 169.45], [602.38, 169.48]]}, {\"category\": \"car\", \"corners_3d\": [[572.3, 193.39], [597.51, 193.41], [594.68, 194.45], [567.27, 194.43], [572.3, 168.64], [597.51, 168.62], [594.68, 167.49], [567.27, 167.51]]}]\n```" - } - ] - }, - { - "id": 1187, - "data_source": "Mono3DRefer", - "scene_name": "000793", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[495.46, 226.64], [550.1, 226.78], [530.94, 237.67], [463.29, 237.46], [495.46, 175.94], [550.1, 175.93], [530.94, 174.58], [463.29, 174.61]]}, {\"category\": \"car\", \"corners_3d\": [[532.47, 208.75], [568.06, 208.75], [564.07, 212.05], [524.17, 212.05], [532.47, 176.16], [568.06, 176.16], [564.07, 175.51], [524.17, 175.51]]}, {\"category\": \"car\", \"corners_3d\": [[605.75, 201.71], [639.58, 201.75], [640.97, 204.6], [602.38, 204.55], [605.75, 170.96], [639.58, 170.94], [640.97, 169.45], [602.38, 169.48]]}, {\"category\": \"van\", \"corners_3d\": [[331.15, 267.14], [216.99, 267.42], [314.98, 244.57], [398.62, 244.42], [331.15, 156.52], [216.99, 156.44], [314.98, 163.11], [398.62, 163.15]]}, {\"category\": \"car\", \"corners_3d\": [[572.3, 193.39], [597.51, 193.41], [594.68, 194.45], [567.27, 194.43], [572.3, 168.64], [597.51, 168.62], [594.68, 167.49], [567.27, 167.51]]}]\n```" - } - ] - }, - { - "id": 1188, - "data_source": "Mono3DRefer", - "scene_name": "000794", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[541.82, 198.0], [520.75, 197.86], [537.57, 196.0], [557.16, 196.13], [541.82, 177.96], [520.75, 177.93], [537.57, 177.56], [557.16, 177.58]]}, {\"category\": \"car\", \"corners_3d\": [[717.78, 194.64], [738.48, 194.74], [739.43, 196.49], [717.07, 196.38], [717.78, 176.31], [738.48, 176.33], [739.43, 176.6], [717.07, 176.59]]}]\n```" - } - ] - }, - { - "id": 1189, - "data_source": "Mono3DRefer", - "scene_name": "000797", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[719.27, 232.8], [721.7, 234.05], [694.82, 234.08], [692.93, 232.83], [719.27, 183.39], [721.7, 183.61], [694.82, 183.62], [692.93, 183.4]]}]\n```" - } - ] - }, - { - "id": 1190, - "data_source": "Mono3DRefer", - "scene_name": "000797", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[378.32, 288.57], [297.93, 288.84], [360.99, 264.18], [424.21, 264.01], [378.32, 194.16], [297.93, 194.2], [360.99, 189.67], [424.21, 189.63]]}]\n```" - } - ] - }, - { - "id": 1191, - "data_source": "Mono3DRefer", - "scene_name": "000797", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[378.32, 288.57], [297.93, 288.84], [360.99, 264.18], [424.21, 264.01], [378.32, 194.16], [297.93, 194.2], [360.99, 189.67], [424.21, 189.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[719.27, 232.8], [721.7, 234.05], [694.82, 234.08], [692.93, 232.83], [719.27, 183.39], [721.7, 183.61], [694.82, 183.62], [692.93, 183.4]]}]\n```" - } - ] - }, - { - "id": 1192, - "data_source": "Mono3DRefer", - "scene_name": "000802", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[481.64, 243.78], [420.84, 243.36], [459.12, 232.06], [510.31, 232.35], [481.64, 185.4], [420.84, 185.33], [459.12, 183.33], [510.31, 183.38]]}, {\"category\": \"car\", \"corners_3d\": [[1188.3, 271.67], [1186.7, 262.31], [1589.32, 286.25], [1646.36, 301.73], [1188.3, 178.46], [1186.7, 177.93], [1589.32, 179.29], [1646.36, 180.17]]}, {\"category\": \"car\", \"corners_3d\": [[520.28, 230.03], [469.82, 229.92], [492.45, 221.89], [535.84, 221.97], [520.28, 183.06], [469.82, 183.04], [492.45, 181.6], [535.84, 181.62]]}, {\"category\": \"car\", \"corners_3d\": [[693.98, 204.95], [724.01, 205.05], [731.56, 209.39], [697.47, 209.26], [693.98, 175.2], [724.01, 175.21], [731.56, 175.53], [697.47, 175.52]]}, {\"category\": \"car\", \"corners_3d\": [[539.35, 220.91], [502.29, 220.71], [520.17, 215.51], [553.26, 215.67], [539.35, 187.17], [502.29, 187.11], [520.17, 185.56], [553.26, 185.61]]}, {\"category\": \"car\", \"corners_3d\": [[556.03, 215.16], [522.23, 214.98], [537.7, 210.69], [568.1, 210.83], [556.03, 184.42], [522.23, 184.37], [537.7, 183.2], [568.1, 183.24]]}, {\"category\": \"car\", \"corners_3d\": [[276.91, 225.32], [255.73, 229.16], [131.48, 229.73], [161.42, 225.81], [276.91, 185.77], [255.73, 186.71], [131.48, 186.85], [161.42, 185.89]]}, {\"category\": \"car\", \"corners_3d\": [[878.49, 332.52], [1042.91, 333.98], [1471.97, 505.14], [1128.16, 498.98], [878.49, 204.39], [1042.91, 204.68], [1471.97, 238.48], [1128.16, 237.27]]}, {\"category\": \"car\", \"corners_3d\": [[775.04, 266.35], [862.28, 266.81], [959.75, 307.83], [834.02, 306.89], [775.04, 186.65], [862.28, 186.72], [959.75, 192.77], [834.02, 192.63]]}]\n```" - } - ] - }, - { - "id": 1193, - "data_source": "Mono3DRefer", - "scene_name": "000807", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[227.78, 226.81], [166.67, 226.86], [254.03, 216.02], [302.81, 215.99], [227.78, 142.19], [166.67, 142.16], [254.03, 148.32], [302.81, 148.34]]}, {\"category\": \"truck\", \"corners_3d\": [[311.7, 213.28], [265.07, 213.28], [320.11, 206.81], [359.27, 206.81], [311.7, 146.85], [265.07, 146.85], [320.11, 151.01], [359.27, 151.01]]}]\n```" - } - ] - }, - { - "id": 1194, - "data_source": "Mono3DRefer", - "scene_name": "000807", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[379.06, 214.86], [345.3, 214.86], [372.33, 210.56], [402.63, 210.55], [379.06, 187.03], [345.3, 187.03], [372.33, 185.58], [402.63, 185.58]]}]\n```" - } - ] - }, - { - "id": 1195, - "data_source": "Mono3DRefer", - "scene_name": "000807", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[379.06, 214.86], [345.3, 214.86], [372.33, 210.56], [402.63, 210.55], [379.06, 187.03], [345.3, 187.03], [372.33, 185.58], [402.63, 185.58]]}, {\"category\": \"truck\", \"corners_3d\": [[227.78, 226.81], [166.67, 226.86], [254.03, 216.02], [302.81, 215.99], [227.78, 142.19], [166.67, 142.16], [254.03, 148.32], [302.81, 148.34]]}, {\"category\": \"truck\", \"corners_3d\": [[311.7, 213.28], [265.07, 213.28], [320.11, 206.81], [359.27, 206.81], [311.7, 146.85], [265.07, 146.85], [320.11, 151.01], [359.27, 151.01]]}]\n```" - } - ] - }, - { - "id": 1196, - "data_source": "Mono3DRefer", - "scene_name": "000808", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[646.34, 204.37], [675.17, 204.42], [678.94, 207.72], [647.11, 207.66], [646.34, 171.61], [675.17, 171.61], [678.94, 171.48], [647.11, 171.48]]}]\n```" - } - ] - }, - { - "id": 1197, - "data_source": "Mono3DRefer", - "scene_name": "000808", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.47, 216.2], [627.23, 216.25], [627.56, 221.62], [586.25, 221.57], [590.47, 182.75], [627.23, 182.76], [627.56, 183.99], [586.25, 183.97]]}, {\"category\": \"car\", \"corners_3d\": [[603.41, 228.09], [657.07, 228.1], [665.27, 237.75], [602.23, 237.75], [603.41, 184.91], [657.07, 184.91], [665.27, 187.02], [602.23, 187.02]]}, {\"category\": \"car\", \"corners_3d\": [[989.38, 209.69], [968.41, 207.81], [1061.43, 207.62], [1087.27, 209.48], [989.38, 175.88], [968.41, 175.72], [1061.43, 175.71], [1087.27, 175.86]]}, {\"category\": \"car\", \"corners_3d\": [[430.98, 213.63], [402.42, 213.62], [419.9, 210.28], [446.12, 210.3], [430.98, 187.16], [402.42, 187.16], [419.9, 185.99], [446.12, 185.99]]}]\n```" - } - ] - }, - { - "id": 1198, - "data_source": "Mono3DRefer", - "scene_name": "000808", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[743.29, 208.39], [741.21, 207.56], [781.13, 207.67], [784.17, 208.51], [743.29, 174.65], [741.21, 174.61], [781.13, 174.62], [784.17, 174.66]]}]\n```" - } - ] - }, - { - "id": 1199, - "data_source": "Mono3DRefer", - "scene_name": "000808", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[886.25, 225.42], [868.65, 225.44], [858.52, 223.5], [875.47, 223.48], [886.25, 170.35], [868.65, 170.35], [858.52, 170.45], [875.47, 170.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[552.43, 213.91], [553.06, 214.6], [540.5, 214.7], [540.08, 214.0], [552.43, 178.52], [553.06, 178.62], [540.5, 178.63], [540.08, 178.54]]}]\n```" - } - ] - }, - { - "id": 1200, - "data_source": "Mono3DRefer", - "scene_name": "000808", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.47, 216.2], [627.23, 216.25], [627.56, 221.62], [586.25, 221.57], [590.47, 182.75], [627.23, 182.76], [627.56, 183.99], [586.25, 183.97]]}, {\"category\": \"car\", \"corners_3d\": [[603.41, 228.09], [657.07, 228.1], [665.27, 237.75], [602.23, 237.75], [603.41, 184.91], [657.07, 184.91], [665.27, 187.02], [602.23, 187.02]]}, {\"category\": \"car\", \"corners_3d\": [[989.38, 209.69], [968.41, 207.81], [1061.43, 207.62], [1087.27, 209.48], [989.38, 175.88], [968.41, 175.72], [1061.43, 175.71], [1087.27, 175.86]]}, {\"category\": \"car\", \"corners_3d\": [[430.98, 213.63], [402.42, 213.62], [419.9, 210.28], [446.12, 210.3], [430.98, 187.16], [402.42, 187.16], [419.9, 185.99], [446.12, 185.99]]}, {\"category\": \"van\", \"corners_3d\": [[646.34, 204.37], [675.17, 204.42], [678.94, 207.72], [647.11, 207.66], [646.34, 171.61], [675.17, 171.61], [678.94, 171.48], [647.11, 171.48]]}]\n```" - } - ] - }, - { - "id": 1201, - "data_source": "Mono3DRefer", - "scene_name": "000808", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[743.29, 208.39], [741.21, 207.56], [781.13, 207.67], [784.17, 208.51], [743.29, 174.65], [741.21, 174.61], [781.13, 174.62], [784.17, 174.66]]}, {\"category\": \"van\", \"corners_3d\": [[646.34, 204.37], [675.17, 204.42], [678.94, 207.72], [647.11, 207.66], [646.34, 171.61], [675.17, 171.61], [678.94, 171.48], [647.11, 171.48]]}]\n```" - } - ] - }, - { - "id": 1202, - "data_source": "Mono3DRefer", - "scene_name": "000808", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[886.25, 225.42], [868.65, 225.44], [858.52, 223.5], [875.47, 223.48], [886.25, 170.35], [868.65, 170.35], [858.52, 170.45], [875.47, 170.45]]}, {\"category\": \"van\", \"corners_3d\": [[646.34, 204.37], [675.17, 204.42], [678.94, 207.72], [647.11, 207.66], [646.34, 171.61], [675.17, 171.61], [678.94, 171.48], [647.11, 171.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[552.43, 213.91], [553.06, 214.6], [540.5, 214.7], [540.08, 214.0], [552.43, 178.52], [553.06, 178.62], [540.5, 178.63], [540.08, 178.54]]}]\n```" - } - ] - }, - { - "id": 1203, - "data_source": "Mono3DRefer", - "scene_name": "000808", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.47, 216.2], [627.23, 216.25], [627.56, 221.62], [586.25, 221.57], [590.47, 182.75], [627.23, 182.76], [627.56, 183.99], [586.25, 183.97]]}, {\"category\": \"car\", \"corners_3d\": [[603.41, 228.09], [657.07, 228.1], [665.27, 237.75], [602.23, 237.75], [603.41, 184.91], [657.07, 184.91], [665.27, 187.02], [602.23, 187.02]]}, {\"category\": \"car\", \"corners_3d\": [[989.38, 209.69], [968.41, 207.81], [1061.43, 207.62], [1087.27, 209.48], [989.38, 175.88], [968.41, 175.72], [1061.43, 175.71], [1087.27, 175.86]]}, {\"category\": \"car\", \"corners_3d\": [[430.98, 213.63], [402.42, 213.62], [419.9, 210.28], [446.12, 210.3], [430.98, 187.16], [402.42, 187.16], [419.9, 185.99], [446.12, 185.99]]}, {\"category\": \"cyclist\", \"corners_3d\": [[743.29, 208.39], [741.21, 207.56], [781.13, 207.67], [784.17, 208.51], [743.29, 174.65], [741.21, 174.61], [781.13, 174.62], [784.17, 174.66]]}]\n```" - } - ] - }, - { - "id": 1204, - "data_source": "Mono3DRefer", - "scene_name": "000808", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.47, 216.2], [627.23, 216.25], [627.56, 221.62], [586.25, 221.57], [590.47, 182.75], [627.23, 182.76], [627.56, 183.99], [586.25, 183.97]]}, {\"category\": \"car\", \"corners_3d\": [[603.41, 228.09], [657.07, 228.1], [665.27, 237.75], [602.23, 237.75], [603.41, 184.91], [657.07, 184.91], [665.27, 187.02], [602.23, 187.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[886.25, 225.42], [868.65, 225.44], [858.52, 223.5], [875.47, 223.48], [886.25, 170.35], [868.65, 170.35], [858.52, 170.45], [875.47, 170.45]]}, {\"category\": \"car\", \"corners_3d\": [[989.38, 209.69], [968.41, 207.81], [1061.43, 207.62], [1087.27, 209.48], [989.38, 175.88], [968.41, 175.72], [1061.43, 175.71], [1087.27, 175.86]]}, {\"category\": \"car\", \"corners_3d\": [[430.98, 213.63], [402.42, 213.62], [419.9, 210.28], [446.12, 210.3], [430.98, 187.16], [402.42, 187.16], [419.9, 185.99], [446.12, 185.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[552.43, 213.91], [553.06, 214.6], [540.5, 214.7], [540.08, 214.0], [552.43, 178.52], [553.06, 178.62], [540.5, 178.63], [540.08, 178.54]]}]\n```" - } - ] - }, - { - "id": 1205, - "data_source": "Mono3DRefer", - "scene_name": "000808", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[886.25, 225.42], [868.65, 225.44], [858.52, 223.5], [875.47, 223.48], [886.25, 170.35], [868.65, 170.35], [858.52, 170.45], [875.47, 170.45]]}, {\"category\": \"cyclist\", \"corners_3d\": [[743.29, 208.39], [741.21, 207.56], [781.13, 207.67], [784.17, 208.51], [743.29, 174.65], [741.21, 174.61], [781.13, 174.62], [784.17, 174.66]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[552.43, 213.91], [553.06, 214.6], [540.5, 214.7], [540.08, 214.0], [552.43, 178.52], [553.06, 178.62], [540.5, 178.63], [540.08, 178.54]]}]\n```" - } - ] - }, - { - "id": 1206, - "data_source": "Mono3DRefer", - "scene_name": "000808", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.47, 216.2], [627.23, 216.25], [627.56, 221.62], [586.25, 221.57], [590.47, 182.75], [627.23, 182.76], [627.56, 183.99], [586.25, 183.97]]}, {\"category\": \"car\", \"corners_3d\": [[603.41, 228.09], [657.07, 228.1], [665.27, 237.75], [602.23, 237.75], [603.41, 184.91], [657.07, 184.91], [665.27, 187.02], [602.23, 187.02]]}, {\"category\": \"car\", \"corners_3d\": [[989.38, 209.69], [968.41, 207.81], [1061.43, 207.62], [1087.27, 209.48], [989.38, 175.88], [968.41, 175.72], [1061.43, 175.71], [1087.27, 175.86]]}, {\"category\": \"car\", \"corners_3d\": [[430.98, 213.63], [402.42, 213.62], [419.9, 210.28], [446.12, 210.3], [430.98, 187.16], [402.42, 187.16], [419.9, 185.99], [446.12, 185.99]]}, {\"category\": \"cyclist\", \"corners_3d\": [[743.29, 208.39], [741.21, 207.56], [781.13, 207.67], [784.17, 208.51], [743.29, 174.65], [741.21, 174.61], [781.13, 174.62], [784.17, 174.66]]}, {\"category\": \"van\", \"corners_3d\": [[646.34, 204.37], [675.17, 204.42], [678.94, 207.72], [647.11, 207.66], [646.34, 171.61], [675.17, 171.61], [678.94, 171.48], [647.11, 171.48]]}]\n```" - } - ] - }, - { - "id": 1207, - "data_source": "Mono3DRefer", - "scene_name": "000808", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.47, 216.2], [627.23, 216.25], [627.56, 221.62], [586.25, 221.57], [590.47, 182.75], [627.23, 182.76], [627.56, 183.99], [586.25, 183.97]]}, {\"category\": \"car\", \"corners_3d\": [[603.41, 228.09], [657.07, 228.1], [665.27, 237.75], [602.23, 237.75], [603.41, 184.91], [657.07, 184.91], [665.27, 187.02], [602.23, 187.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[886.25, 225.42], [868.65, 225.44], [858.52, 223.5], [875.47, 223.48], [886.25, 170.35], [868.65, 170.35], [858.52, 170.45], [875.47, 170.45]]}, {\"category\": \"car\", \"corners_3d\": [[989.38, 209.69], [968.41, 207.81], [1061.43, 207.62], [1087.27, 209.48], [989.38, 175.88], [968.41, 175.72], [1061.43, 175.71], [1087.27, 175.86]]}, {\"category\": \"car\", \"corners_3d\": [[430.98, 213.63], [402.42, 213.62], [419.9, 210.28], [446.12, 210.3], [430.98, 187.16], [402.42, 187.16], [419.9, 185.99], [446.12, 185.99]]}, {\"category\": \"van\", \"corners_3d\": [[646.34, 204.37], [675.17, 204.42], [678.94, 207.72], [647.11, 207.66], [646.34, 171.61], [675.17, 171.61], [678.94, 171.48], [647.11, 171.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[552.43, 213.91], [553.06, 214.6], [540.5, 214.7], [540.08, 214.0], [552.43, 178.52], [553.06, 178.62], [540.5, 178.63], [540.08, 178.54]]}]\n```" - } - ] - }, - { - "id": 1208, - "data_source": "Mono3DRefer", - "scene_name": "000808", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[886.25, 225.42], [868.65, 225.44], [858.52, 223.5], [875.47, 223.48], [886.25, 170.35], [868.65, 170.35], [858.52, 170.45], [875.47, 170.45]]}, {\"category\": \"cyclist\", \"corners_3d\": [[743.29, 208.39], [741.21, 207.56], [781.13, 207.67], [784.17, 208.51], [743.29, 174.65], [741.21, 174.61], [781.13, 174.62], [784.17, 174.66]]}, {\"category\": \"van\", \"corners_3d\": [[646.34, 204.37], [675.17, 204.42], [678.94, 207.72], [647.11, 207.66], [646.34, 171.61], [675.17, 171.61], [678.94, 171.48], [647.11, 171.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[552.43, 213.91], [553.06, 214.6], [540.5, 214.7], [540.08, 214.0], [552.43, 178.52], [553.06, 178.62], [540.5, 178.63], [540.08, 178.54]]}]\n```" - } - ] - }, - { - "id": 1209, - "data_source": "Mono3DRefer", - "scene_name": "000808", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.47, 216.2], [627.23, 216.25], [627.56, 221.62], [586.25, 221.57], [590.47, 182.75], [627.23, 182.76], [627.56, 183.99], [586.25, 183.97]]}, {\"category\": \"car\", \"corners_3d\": [[603.41, 228.09], [657.07, 228.1], [665.27, 237.75], [602.23, 237.75], [603.41, 184.91], [657.07, 184.91], [665.27, 187.02], [602.23, 187.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[886.25, 225.42], [868.65, 225.44], [858.52, 223.5], [875.47, 223.48], [886.25, 170.35], [868.65, 170.35], [858.52, 170.45], [875.47, 170.45]]}, {\"category\": \"car\", \"corners_3d\": [[989.38, 209.69], [968.41, 207.81], [1061.43, 207.62], [1087.27, 209.48], [989.38, 175.88], [968.41, 175.72], [1061.43, 175.71], [1087.27, 175.86]]}, {\"category\": \"car\", \"corners_3d\": [[430.98, 213.63], [402.42, 213.62], [419.9, 210.28], [446.12, 210.3], [430.98, 187.16], [402.42, 187.16], [419.9, 185.99], [446.12, 185.99]]}, {\"category\": \"cyclist\", \"corners_3d\": [[743.29, 208.39], [741.21, 207.56], [781.13, 207.67], [784.17, 208.51], [743.29, 174.65], [741.21, 174.61], [781.13, 174.62], [784.17, 174.66]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[552.43, 213.91], [553.06, 214.6], [540.5, 214.7], [540.08, 214.0], [552.43, 178.52], [553.06, 178.62], [540.5, 178.63], [540.08, 178.54]]}]\n```" - } - ] - }, - { - "id": 1210, - "data_source": "Mono3DRefer", - "scene_name": "000808", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.47, 216.2], [627.23, 216.25], [627.56, 221.62], [586.25, 221.57], [590.47, 182.75], [627.23, 182.76], [627.56, 183.99], [586.25, 183.97]]}, {\"category\": \"car\", \"corners_3d\": [[603.41, 228.09], [657.07, 228.1], [665.27, 237.75], [602.23, 237.75], [603.41, 184.91], [657.07, 184.91], [665.27, 187.02], [602.23, 187.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[886.25, 225.42], [868.65, 225.44], [858.52, 223.5], [875.47, 223.48], [886.25, 170.35], [868.65, 170.35], [858.52, 170.45], [875.47, 170.45]]}, {\"category\": \"car\", \"corners_3d\": [[989.38, 209.69], [968.41, 207.81], [1061.43, 207.62], [1087.27, 209.48], [989.38, 175.88], [968.41, 175.72], [1061.43, 175.71], [1087.27, 175.86]]}, {\"category\": \"car\", \"corners_3d\": [[430.98, 213.63], [402.42, 213.62], [419.9, 210.28], [446.12, 210.3], [430.98, 187.16], [402.42, 187.16], [419.9, 185.99], [446.12, 185.99]]}, {\"category\": \"cyclist\", \"corners_3d\": [[743.29, 208.39], [741.21, 207.56], [781.13, 207.67], [784.17, 208.51], [743.29, 174.65], [741.21, 174.61], [781.13, 174.62], [784.17, 174.66]]}, {\"category\": \"van\", \"corners_3d\": [[646.34, 204.37], [675.17, 204.42], [678.94, 207.72], [647.11, 207.66], [646.34, 171.61], [675.17, 171.61], [678.94, 171.48], [647.11, 171.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[552.43, 213.91], [553.06, 214.6], [540.5, 214.7], [540.08, 214.0], [552.43, 178.52], [553.06, 178.62], [540.5, 178.63], [540.08, 178.54]]}]\n```" - } - ] - }, - { - "id": 1211, - "data_source": "Mono3DRefer", - "scene_name": "000810", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[668.84, 243.48], [725.78, 243.31], [764.12, 262.99], [691.35, 263.26], [668.84, 192.4], [725.78, 192.36], [764.12, 197.8], [691.35, 197.88]]}]\n```" - } - ] - }, - { - "id": 1212, - "data_source": "Mono3DRefer", - "scene_name": "000811", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[703.08, 211.17], [722.75, 211.23], [728.33, 214.02], [707.21, 213.96], [703.08, 162.95], [722.75, 162.93], [728.33, 162.21], [707.21, 162.22]]}, {\"category\": \"cyclist\", \"corners_3d\": [[680.88, 200.33], [689.59, 200.35], [690.8, 201.51], [681.73, 201.48], [680.88, 170.35], [689.59, 170.35], [690.8, 170.24], [681.73, 170.24]]}]\n```" - } - ] - }, - { - "id": 1213, - "data_source": "Mono3DRefer", - "scene_name": "000811", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[896.49, 249.77], [936.44, 249.73], [958.64, 254.85], [916.05, 254.9], [896.49, 155.8], [936.44, 155.81], [958.64, 154.67], [916.05, 154.66]]}]\n```" - } - ] - }, - { - "id": 1214, - "data_source": "Mono3DRefer", - "scene_name": "000811", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[703.08, 211.17], [722.75, 211.23], [728.33, 214.02], [707.21, 213.96], [703.08, 162.95], [722.75, 162.93], [728.33, 162.21], [707.21, 162.22]]}, {\"category\": \"cyclist\", \"corners_3d\": [[680.88, 200.33], [689.59, 200.35], [690.8, 201.51], [681.73, 201.48], [680.88, 170.35], [689.59, 170.35], [690.8, 170.24], [681.73, 170.24]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[896.49, 249.77], [936.44, 249.73], [958.64, 254.85], [916.05, 254.9], [896.49, 155.8], [936.44, 155.81], [958.64, 154.67], [916.05, 154.66]]}]\n```" - } - ] - }, - { - "id": 1215, - "data_source": "Mono3DRefer", - "scene_name": "000813", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[524.62, 200.25], [501.9, 200.08], [521.52, 197.91], [542.52, 198.06], [524.62, 178.54], [501.9, 178.5], [521.52, 178.05], [542.52, 178.08]]}, {\"category\": \"car\", \"corners_3d\": [[716.79, 195.81], [738.35, 195.92], [739.33, 197.85], [715.97, 197.72], [716.79, 176.72], [738.35, 176.74], [739.33, 177.06], [715.97, 177.04]]}]\n```" - } - ] - }, - { - "id": 1216, - "data_source": "Mono3DRefer", - "scene_name": "000816", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1006.59, 326.71], [1106.92, 326.53], [1190.72, 350.21], [1074.12, 350.46], [1006.59, 148.06], [1106.92, 148.1], [1190.72, 142.64], [1074.12, 142.58]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[775.84, 241.54], [821.8, 241.77], [828.27, 243.96], [780.61, 243.71], [775.84, 156.78], [821.8, 156.68], [828.27, 155.78], [780.61, 155.88]]}]\n```" - } - ] - }, - { - "id": 1217, - "data_source": "Mono3DRefer", - "scene_name": "000816", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[642.33, 208.72], [696.44, 208.76], [718.58, 216.2], [649.69, 216.13], [642.33, 141.65], [696.44, 141.59], [718.58, 130.69], [649.69, 130.79]]}]\n```" - } - ] - }, - { - "id": 1218, - "data_source": "Mono3DRefer", - "scene_name": "000816", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.47, 228.9], [538.4, 228.73], [554.66, 220.22], [608.01, 220.33], [603.47, 170.95], [538.4, 170.99], [554.66, 172.89], [608.01, 172.86]]}]\n```" - } - ] - }, - { - "id": 1219, - "data_source": "Mono3DRefer", - "scene_name": "000816", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1006.59, 326.71], [1106.92, 326.53], [1190.72, 350.21], [1074.12, 350.46], [1006.59, 148.06], [1106.92, 148.1], [1190.72, 142.64], [1074.12, 142.58]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[775.84, 241.54], [821.8, 241.77], [828.27, 243.96], [780.61, 243.71], [775.84, 156.78], [821.8, 156.68], [828.27, 155.78], [780.61, 155.88]]}, {\"category\": \"van\", \"corners_3d\": [[642.33, 208.72], [696.44, 208.76], [718.58, 216.2], [649.69, 216.13], [642.33, 141.65], [696.44, 141.59], [718.58, 130.69], [649.69, 130.79]]}]\n```" - } - ] - }, - { - "id": 1220, - "data_source": "Mono3DRefer", - "scene_name": "000816", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1006.59, 326.71], [1106.92, 326.53], [1190.72, 350.21], [1074.12, 350.46], [1006.59, 148.06], [1106.92, 148.1], [1190.72, 142.64], [1074.12, 142.58]]}, {\"category\": \"car\", \"corners_3d\": [[603.47, 228.9], [538.4, 228.73], [554.66, 220.22], [608.01, 220.33], [603.47, 170.95], [538.4, 170.99], [554.66, 172.89], [608.01, 172.86]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[775.84, 241.54], [821.8, 241.77], [828.27, 243.96], [780.61, 243.71], [775.84, 156.78], [821.8, 156.68], [828.27, 155.78], [780.61, 155.88]]}]\n```" - } - ] - }, - { - "id": 1221, - "data_source": "Mono3DRefer", - "scene_name": "000816", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.47, 228.9], [538.4, 228.73], [554.66, 220.22], [608.01, 220.33], [603.47, 170.95], [538.4, 170.99], [554.66, 172.89], [608.01, 172.86]]}, {\"category\": \"van\", \"corners_3d\": [[642.33, 208.72], [696.44, 208.76], [718.58, 216.2], [649.69, 216.13], [642.33, 141.65], [696.44, 141.59], [718.58, 130.69], [649.69, 130.79]]}]\n```" - } - ] - }, - { - "id": 1222, - "data_source": "Mono3DRefer", - "scene_name": "000816", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1006.59, 326.71], [1106.92, 326.53], [1190.72, 350.21], [1074.12, 350.46], [1006.59, 148.06], [1106.92, 148.1], [1190.72, 142.64], [1074.12, 142.58]]}, {\"category\": \"car\", \"corners_3d\": [[603.47, 228.9], [538.4, 228.73], [554.66, 220.22], [608.01, 220.33], [603.47, 170.95], [538.4, 170.99], [554.66, 172.89], [608.01, 172.86]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[775.84, 241.54], [821.8, 241.77], [828.27, 243.96], [780.61, 243.71], [775.84, 156.78], [821.8, 156.68], [828.27, 155.78], [780.61, 155.88]]}, {\"category\": \"van\", \"corners_3d\": [[642.33, 208.72], [696.44, 208.76], [718.58, 216.2], [649.69, 216.13], [642.33, 141.65], [696.44, 141.59], [718.58, 130.69], [649.69, 130.79]]}]\n```" - } - ] - }, - { - "id": 1223, - "data_source": "Mono3DRefer", - "scene_name": "000817", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[22.95, 515.87], [-314.71, 516.0], [179.18, 332.52], [336.24, 332.5], [22.95, 203.62], [-314.71, 203.64], [179.18, 187.18], [336.24, 187.17]]}, {\"category\": \"car\", \"corners_3d\": [[39.07, 331.83], [-121.39, 332.55], [137.54, 274.79], [239.39, 274.5], [39.07, 198.26], [-121.39, 198.38], [137.54, 189.15], [239.39, 189.1]]}, {\"category\": \"car\", \"corners_3d\": [[383.36, 232.4], [430.54, 232.72], [393.87, 241.97], [339.67, 241.54], [383.36, 190.16], [430.54, 190.26], [393.87, 192.94], [339.67, 192.82]]}, {\"category\": \"car\", \"corners_3d\": [[337.45, 231.79], [379.58, 232.27], [313.2, 244.4], [263.19, 243.72], [337.45, 190.01], [379.58, 190.15], [313.2, 193.69], [263.19, 193.48]]}]\n```" - } - ] - }, - { - "id": 1224, - "data_source": "Mono3DRefer", - "scene_name": "000818", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[571.61, 193.85], [543.89, 193.87], [547.58, 192.09], [572.95, 192.07], [571.61, 164.71], [543.89, 164.7], [547.58, 165.4], [572.95, 165.41]]}, {\"category\": \"car\", \"corners_3d\": [[183.41, 524.95], [-157.96, 528.38], [245.71, 337.66], [402.97, 336.92], [183.41, 239.19], [-157.96, 239.84], [245.71, 203.9], [402.97, 203.77]]}]\n```" - } - ] - }, - { - "id": 1225, - "data_source": "Mono3DRefer", - "scene_name": "000821", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[751.78, 286.73], [805.34, 307.87], [425.49, 323.11], [431.96, 297.39], [751.78, 187.24], [805.34, 189.91], [425.49, 191.83], [431.96, 188.58]]}, {\"category\": \"car\", \"corners_3d\": [[1297.95, 305.36], [1448.91, 330.21], [1151.11, 343.12], [1040.1, 314.39], [1297.95, 189.92], [1448.91, 193.13], [1151.11, 194.79], [1040.1, 191.09]]}]\n```" - } - ] - }, - { - "id": 1226, - "data_source": "Mono3DRefer", - "scene_name": "000822", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[556.18, 199.63], [584.65, 199.67], [577.98, 201.71], [546.34, 201.65], [556.18, 173.22], [584.65, 173.2], [577.98, 172.27], [546.34, 172.3]]}, {\"category\": \"car\", \"corners_3d\": [[615.45, 191.49], [639.01, 191.51], [638.54, 192.45], [612.77, 192.43], [615.45, 170.08], [639.01, 170.06], [638.54, 168.98], [612.77, 169.01]]}, {\"category\": \"car\", \"corners_3d\": [[592.25, 187.19], [610.73, 187.2], [609.53, 187.55], [589.9, 187.55], [592.25, 169.06], [610.73, 169.05], [609.53, 168.26], [589.9, 168.28]]}]\n```" - } - ] - }, - { - "id": 1227, - "data_source": "Mono3DRefer", - "scene_name": "000826", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[1439.16, 553.49], [1235.99, 517.93], [1195.07, 459.24], [1356.25, 483.03], [1439.16, 246.86], [1235.99, 240.61], [1195.07, 230.3], [1356.25, 234.48]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[831.17, 327.27], [896.03, 327.28], [927.22, 342.69], [855.49, 342.67], [831.17, 209.02], [896.03, 209.03], [927.22, 211.93], [855.49, 211.93]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[883.54, 326.95], [907.86, 319.49], [976.07, 329.0], [953.94, 337.56], [883.54, 205.13], [907.86, 203.92], [976.07, 205.47], [953.94, 206.86]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[872.58, 315.4], [823.34, 318.38], [787.23, 306.89], [833.15, 304.39], [872.58, 212.65], [823.34, 213.34], [787.23, 210.67], [833.15, 210.09]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[917.03, 315.39], [855.13, 311.86], [855.04, 302.69], [912.47, 305.72], [917.03, 201.43], [855.13, 200.91], [855.04, 199.54], [912.47, 199.99]]}]\n```" - } - ] - }, - { - "id": 1228, - "data_source": "Mono3DRefer", - "scene_name": "000826", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[544.57, 228.53], [532.58, 226.87], [563.58, 225.85], [576.4, 227.44], [544.57, 172.81], [532.58, 173.12], [563.58, 173.31], [576.4, 173.01]]}]\n```" - } - ] - }, - { - "id": 1229, - "data_source": "Mono3DRefer", - "scene_name": "000826", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[486.14, 207.28], [499.41, 209.36], [342.14, 211.6], [340.51, 209.19], [486.14, 131.93], [499.41, 127.93], [342.14, 123.61], [340.51, 128.26]]}, {\"category\": \"tram\", \"corners_3d\": [[548.76, 202.9], [566.4, 204.57], [500.44, 205.37], [487.15, 203.6], [548.76, 130.51], [566.4, 126.54], [500.44, 124.62], [487.15, 128.86]]}, {\"category\": \"tram\", \"corners_3d\": [[764.65, 194.87], [794.81, 195.78], [568.1, 197.73], [550.4, 196.56], [764.65, 130.08], [794.81, 126.57], [568.1, 119.06], [550.4, 123.55]]}]\n```" - } - ] - }, - { - "id": 1230, - "data_source": "Mono3DRefer", - "scene_name": "000826", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[544.57, 228.53], [532.58, 226.87], [563.58, 225.85], [576.4, 227.44], [544.57, 172.81], [532.58, 173.12], [563.58, 173.31], [576.4, 173.01]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1439.16, 553.49], [1235.99, 517.93], [1195.07, 459.24], [1356.25, 483.03], [1439.16, 246.86], [1235.99, 240.61], [1195.07, 230.3], [1356.25, 234.48]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[831.17, 327.27], [896.03, 327.28], [927.22, 342.69], [855.49, 342.67], [831.17, 209.02], [896.03, 209.03], [927.22, 211.93], [855.49, 211.93]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[883.54, 326.95], [907.86, 319.49], [976.07, 329.0], [953.94, 337.56], [883.54, 205.13], [907.86, 203.92], [976.07, 205.47], [953.94, 206.86]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[872.58, 315.4], [823.34, 318.38], [787.23, 306.89], [833.15, 304.39], [872.58, 212.65], [823.34, 213.34], [787.23, 210.67], [833.15, 210.09]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[917.03, 315.39], [855.13, 311.86], [855.04, 302.69], [912.47, 305.72], [917.03, 201.43], [855.13, 200.91], [855.04, 199.54], [912.47, 199.99]]}]\n```" - } - ] - }, - { - "id": 1231, - "data_source": "Mono3DRefer", - "scene_name": "000826", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[486.14, 207.28], [499.41, 209.36], [342.14, 211.6], [340.51, 209.19], [486.14, 131.93], [499.41, 127.93], [342.14, 123.61], [340.51, 128.26]]}, {\"category\": \"tram\", \"corners_3d\": [[548.76, 202.9], [566.4, 204.57], [500.44, 205.37], [487.15, 203.6], [548.76, 130.51], [566.4, 126.54], [500.44, 124.62], [487.15, 128.86]]}, {\"category\": \"tram\", \"corners_3d\": [[764.65, 194.87], [794.81, 195.78], [568.1, 197.73], [550.4, 196.56], [764.65, 130.08], [794.81, 126.57], [568.1, 119.06], [550.4, 123.55]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1439.16, 553.49], [1235.99, 517.93], [1195.07, 459.24], [1356.25, 483.03], [1439.16, 246.86], [1235.99, 240.61], [1195.07, 230.3], [1356.25, 234.48]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[831.17, 327.27], [896.03, 327.28], [927.22, 342.69], [855.49, 342.67], [831.17, 209.02], [896.03, 209.03], [927.22, 211.93], [855.49, 211.93]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[883.54, 326.95], [907.86, 319.49], [976.07, 329.0], [953.94, 337.56], [883.54, 205.13], [907.86, 203.92], [976.07, 205.47], [953.94, 206.86]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[872.58, 315.4], [823.34, 318.38], [787.23, 306.89], [833.15, 304.39], [872.58, 212.65], [823.34, 213.34], [787.23, 210.67], [833.15, 210.09]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[917.03, 315.39], [855.13, 311.86], [855.04, 302.69], [912.47, 305.72], [917.03, 201.43], [855.13, 200.91], [855.04, 199.54], [912.47, 199.99]]}]\n```" - } - ] - }, - { - "id": 1232, - "data_source": "Mono3DRefer", - "scene_name": "000826", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[544.57, 228.53], [532.58, 226.87], [563.58, 225.85], [576.4, 227.44], [544.57, 172.81], [532.58, 173.12], [563.58, 173.31], [576.4, 173.01]]}, {\"category\": \"tram\", \"corners_3d\": [[486.14, 207.28], [499.41, 209.36], [342.14, 211.6], [340.51, 209.19], [486.14, 131.93], [499.41, 127.93], [342.14, 123.61], [340.51, 128.26]]}, {\"category\": \"tram\", \"corners_3d\": [[548.76, 202.9], [566.4, 204.57], [500.44, 205.37], [487.15, 203.6], [548.76, 130.51], [566.4, 126.54], [500.44, 124.62], [487.15, 128.86]]}, {\"category\": \"tram\", \"corners_3d\": [[764.65, 194.87], [794.81, 195.78], [568.1, 197.73], [550.4, 196.56], [764.65, 130.08], [794.81, 126.57], [568.1, 119.06], [550.4, 123.55]]}]\n```" - } - ] - }, - { - "id": 1233, - "data_source": "Mono3DRefer", - "scene_name": "000826", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, pedestrian, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[544.57, 228.53], [532.58, 226.87], [563.58, 225.85], [576.4, 227.44], [544.57, 172.81], [532.58, 173.12], [563.58, 173.31], [576.4, 173.01]]}, {\"category\": \"tram\", \"corners_3d\": [[486.14, 207.28], [499.41, 209.36], [342.14, 211.6], [340.51, 209.19], [486.14, 131.93], [499.41, 127.93], [342.14, 123.61], [340.51, 128.26]]}, {\"category\": \"tram\", \"corners_3d\": [[548.76, 202.9], [566.4, 204.57], [500.44, 205.37], [487.15, 203.6], [548.76, 130.51], [566.4, 126.54], [500.44, 124.62], [487.15, 128.86]]}, {\"category\": \"tram\", \"corners_3d\": [[764.65, 194.87], [794.81, 195.78], [568.1, 197.73], [550.4, 196.56], [764.65, 130.08], [794.81, 126.57], [568.1, 119.06], [550.4, 123.55]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1439.16, 553.49], [1235.99, 517.93], [1195.07, 459.24], [1356.25, 483.03], [1439.16, 246.86], [1235.99, 240.61], [1195.07, 230.3], [1356.25, 234.48]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[831.17, 327.27], [896.03, 327.28], [927.22, 342.69], [855.49, 342.67], [831.17, 209.02], [896.03, 209.03], [927.22, 211.93], [855.49, 211.93]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[883.54, 326.95], [907.86, 319.49], [976.07, 329.0], [953.94, 337.56], [883.54, 205.13], [907.86, 203.92], [976.07, 205.47], [953.94, 206.86]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[872.58, 315.4], [823.34, 318.38], [787.23, 306.89], [833.15, 304.39], [872.58, 212.65], [823.34, 213.34], [787.23, 210.67], [833.15, 210.09]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[917.03, 315.39], [855.13, 311.86], [855.04, 302.69], [912.47, 305.72], [917.03, 201.43], [855.13, 200.91], [855.04, 199.54], [912.47, 199.99]]}]\n```" - } - ] - }, - { - "id": 1234, - "data_source": "Mono3DRefer", - "scene_name": "000829", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[509.61, 244.09], [443.03, 244.02], [473.0, 231.71], [528.07, 231.75], [509.61, 184.78], [443.03, 184.77], [473.0, 182.71], [528.07, 182.72]]}, {\"category\": \"car\", \"corners_3d\": [[550.09, 221.72], [505.74, 221.66], [520.18, 215.67], [559.1, 215.72], [550.09, 180.26], [505.74, 180.25], [520.18, 179.34], [559.1, 179.35]]}]\n```" - } - ] - }, - { - "id": 1235, - "data_source": "Mono3DRefer", - "scene_name": "000830", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[587.17, 193.97], [588.81, 194.23], [581.95, 194.27], [580.39, 194.01], [587.17, 172.85], [588.81, 172.85], [581.95, 172.85], [580.39, 172.85]]}]\n```" - } - ] - }, - { - "id": 1236, - "data_source": "Mono3DRefer", - "scene_name": "000830", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[398.69, 216.18], [362.6, 215.86], [405.44, 210.57], [437.34, 210.81], [398.69, 181.27], [362.6, 181.21], [405.44, 180.18], [437.34, 180.23]]}, {\"category\": \"car\", \"corners_3d\": [[704.32, 199.34], [730.66, 199.46], [734.19, 202.26], [705.07, 202.11], [704.32, 175.96], [730.66, 175.97], [734.19, 176.3], [705.07, 176.28]]}, {\"category\": \"car\", \"corners_3d\": [[828.39, 185.64], [828.67, 185.38], [870.71, 185.59], [871.33, 185.86], [828.39, 171.44], [828.67, 171.47], [870.71, 171.45], [871.33, 171.42]]}, {\"category\": \"car\", \"corners_3d\": [[886.0, 183.84], [885.4, 183.61], [930.11, 183.81], [931.74, 184.06], [886.0, 167.51], [885.4, 167.62], [930.11, 167.52], [931.74, 167.4]]}]\n```" - } - ] - }, - { - "id": 1237, - "data_source": "Mono3DRefer", - "scene_name": "000830", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[398.69, 216.18], [362.6, 215.86], [405.44, 210.57], [437.34, 210.81], [398.69, 181.27], [362.6, 181.21], [405.44, 180.18], [437.34, 180.23]]}, {\"category\": \"car\", \"corners_3d\": [[704.32, 199.34], [730.66, 199.46], [734.19, 202.26], [705.07, 202.11], [704.32, 175.96], [730.66, 175.97], [734.19, 176.3], [705.07, 176.28]]}, {\"category\": \"car\", \"corners_3d\": [[828.39, 185.64], [828.67, 185.38], [870.71, 185.59], [871.33, 185.86], [828.39, 171.44], [828.67, 171.47], [870.71, 171.45], [871.33, 171.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[587.17, 193.97], [588.81, 194.23], [581.95, 194.27], [580.39, 194.01], [587.17, 172.85], [588.81, 172.85], [581.95, 172.85], [580.39, 172.85]]}, {\"category\": \"car\", \"corners_3d\": [[886.0, 183.84], [885.4, 183.61], [930.11, 183.81], [931.74, 184.06], [886.0, 167.51], [885.4, 167.62], [930.11, 167.52], [931.74, 167.4]]}]\n```" - } - ] - }, - { - "id": 1238, - "data_source": "Mono3DRefer", - "scene_name": "000832", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[676.85, 193.44], [697.26, 193.4], [707.71, 195.05], [685.69, 195.09], [676.85, 174.9], [697.26, 174.89], [707.71, 175.06], [685.69, 175.06]]}, {\"category\": \"car\", \"corners_3d\": [[800.72, 200.28], [816.39, 199.8], [866.76, 201.77], [850.89, 202.32], [800.72, 174.91], [816.39, 174.87], [866.76, 175.02], [850.89, 175.06]]}]\n```" - } - ] - }, - { - "id": 1239, - "data_source": "Mono3DRefer", - "scene_name": "000832", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[748.39, 211.53], [791.01, 211.46], [825.56, 218.04], [775.72, 218.14], [748.39, 162.64], [791.01, 162.66], [825.56, 160.92], [775.72, 160.9]]}]\n```" - } - ] - }, - { - "id": 1240, - "data_source": "Mono3DRefer", - "scene_name": "000832", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[676.85, 193.44], [697.26, 193.4], [707.71, 195.05], [685.69, 195.09], [676.85, 174.9], [697.26, 174.89], [707.71, 175.06], [685.69, 175.06]]}, {\"category\": \"van\", \"corners_3d\": [[748.39, 211.53], [791.01, 211.46], [825.56, 218.04], [775.72, 218.14], [748.39, 162.64], [791.01, 162.66], [825.56, 160.92], [775.72, 160.9]]}, {\"category\": \"car\", \"corners_3d\": [[800.72, 200.28], [816.39, 199.8], [866.76, 201.77], [850.89, 202.32], [800.72, 174.91], [816.39, 174.87], [866.76, 175.02], [850.89, 175.06]]}]\n```" - } - ] - }, - { - "id": 1241, - "data_source": "Mono3DRefer", - "scene_name": "000834", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[570.55, 200.07], [547.67, 200.09], [550.97, 198.17], [572.24, 198.16], [570.55, 177.5], [547.67, 177.5], [550.97, 177.17], [572.24, 177.17]]}]\n```" - } - ] - }, - { - "id": 1242, - "data_source": "Mono3DRefer", - "scene_name": "000835", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[568.86, 204.21], [543.75, 204.21], [548.92, 201.72], [572.04, 201.72], [568.86, 182.14], [543.75, 182.14], [548.92, 181.4], [572.04, 181.4]]}]\n```" - } - ] - }, - { - "id": 1243, - "data_source": "Mono3DRefer", - "scene_name": "000836", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[618.96, 219.98], [665.67, 220.19], [666.38, 226.94], [613.04, 226.66], [618.96, 174.81], [665.67, 174.82], [666.38, 175.1], [613.04, 175.08]]}, {\"category\": \"car\", \"corners_3d\": [[526.06, 226.22], [476.9, 226.04], [499.57, 218.87], [542.15, 219.01], [526.06, 181.79], [476.9, 181.76], [499.57, 180.56], [542.15, 180.58]]}, {\"category\": \"car\", \"corners_3d\": [[466.26, 220.03], [427.13, 219.77], [454.75, 214.68], [489.74, 214.88], [466.26, 182.86], [427.13, 182.81], [454.75, 181.73], [489.74, 181.77]]}]\n```" - } - ] - }, - { - "id": 1244, - "data_source": "Mono3DRefer", - "scene_name": "000839", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[491.16, 238.13], [436.12, 238.23], [460.48, 228.18], [507.04, 228.11], [491.16, 183.99], [436.12, 184.01], [460.48, 182.29], [507.04, 182.28]]}]\n```" - } - ] - }, - { - "id": 1245, - "data_source": "Mono3DRefer", - "scene_name": "000842", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[566.81, 203.04], [536.86, 203.08], [541.47, 200.32], [568.68, 200.29], [566.81, 171.57], [536.86, 171.57], [541.47, 171.69], [568.68, 171.69]]}, {\"category\": \"car\", \"corners_3d\": [[576.16, 192.03], [560.33, 192.02], [563.61, 191.02], [578.62, 191.03], [576.16, 174.41], [560.33, 174.4], [563.61, 174.32], [578.62, 174.32]]}]\n```" - } - ] - }, - { - "id": 1246, - "data_source": "Mono3DRefer", - "scene_name": "000845", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[370.16, 237.62], [348.85, 243.39], [197.11, 243.39], [230.83, 237.62], [370.16, 184.78], [348.85, 185.85], [197.11, 185.85], [230.83, 184.78]]}, {\"category\": \"car\", \"corners_3d\": [[253.45, 232.62], [281.73, 228.25], [407.22, 228.65], [388.71, 233.1], [253.45, 174.87], [281.73, 174.72], [407.22, 174.73], [388.71, 174.88]]}, {\"category\": \"car\", \"corners_3d\": [[308.61, 227.23], [326.8, 223.85], [428.83, 223.77], [417.42, 227.14], [308.61, 182.35], [326.8, 181.76], [428.83, 181.75], [417.42, 182.34]]}, {\"category\": \"car\", \"corners_3d\": [[336.04, 218.0], [350.37, 215.7], [440.35, 215.76], [430.83, 218.07], [336.04, 181.62], [350.37, 181.18], [440.35, 181.19], [430.83, 181.64]]}, {\"category\": \"car\", \"corners_3d\": [[366.07, 219.21], [375.45, 216.94], [449.34, 216.63], [443.8, 218.86], [366.07, 181.11], [375.45, 180.71], [449.34, 180.65], [443.8, 181.05]]}, {\"category\": \"car\", \"corners_3d\": [[391.5, 213.2], [399.06, 211.44], [477.77, 211.19], [473.82, 212.92], [391.5, 184.02], [399.06, 183.53], [477.77, 183.46], [473.82, 183.94]]}, {\"category\": \"car\", \"corners_3d\": [[411.06, 206.63], [420.46, 205.09], [485.37, 205.12], [479.05, 206.66], [411.06, 173.82], [420.46, 173.77], [485.37, 173.77], [479.05, 173.82]]}, {\"category\": \"car\", \"corners_3d\": [[428.07, 206.22], [434.68, 204.89], [501.73, 204.82], [497.91, 206.15], [428.07, 177.18], [434.68, 177.0], [501.73, 177.0], [497.91, 177.17]]}, {\"category\": \"car\", \"corners_3d\": [[452.37, 203.84], [442.87, 202.85], [500.55, 201.53], [511.44, 202.44], [452.37, 177.63], [442.87, 177.48], [500.55, 177.27], [511.44, 177.41]]}, {\"category\": \"car\", \"corners_3d\": [[-366.97, 341.0], [-204.44, 312.56], [81.01, 311.39], [-22.36, 339.31], [-366.97, 201.82], [-204.44, 196.92], [81.01, 196.72], [-22.36, 201.53]]}, {\"category\": \"car\", \"corners_3d\": [[34.9, 277.41], [94.58, 266.39], [269.11, 266.13], [230.11, 277.08], [34.9, 188.49], [94.58, 186.84], [269.11, 186.81], [230.11, 188.44]]}, {\"category\": \"car\", \"corners_3d\": [[140.86, 252.66], [183.21, 245.58], [330.91, 245.75], [302.87, 252.87], [140.86, 181.0], [183.21, 180.28], [330.91, 180.3], [302.87, 181.02]]}]\n```" - } - ] - }, - { - "id": 1247, - "data_source": "Mono3DRefer", - "scene_name": "000846", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1175.71, 188.22], [1161.33, 187.65], [1235.67, 187.97], [1253.29, 188.56], [1175.71, 157.64], [1161.33, 158.21], [1235.67, 157.89], [1253.29, 157.3]]}]\n```" - } - ] - }, - { - "id": 1248, - "data_source": "Mono3DRefer", - "scene_name": "000846", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[735.99, 236.25], [799.89, 236.43], [851.02, 255.76], [767.58, 255.47], [735.99, 179.22], [799.89, 179.24], [851.02, 181.18], [767.58, 181.15]]}, {\"category\": \"car\", \"corners_3d\": [[661.73, 195.32], [651.91, 194.6], [707.01, 194.0], [718.31, 194.67], [661.73, 173.01], [651.91, 173.0], [707.01, 173.0], [718.31, 173.0]]}, {\"category\": \"car\", \"corners_3d\": [[829.32, 191.4], [828.22, 190.81], [891.71, 191.18], [894.99, 191.8], [829.32, 167.42], [828.22, 167.59], [891.71, 167.48], [894.99, 167.3]]}, {\"category\": \"car\", \"corners_3d\": [[985.8, 190.19], [979.68, 189.61], [1059.18, 190.04], [1068.29, 190.65], [985.8, 167.17], [979.68, 167.36], [1059.18, 167.22], [1068.29, 167.02]]}, {\"category\": \"car\", \"corners_3d\": [[427.74, 202.81], [413.98, 203.83], [366.21, 203.13], [381.21, 202.16], [427.74, 179.56], [413.98, 179.79], [366.21, 179.63], [381.21, 179.41]]}, {\"category\": \"car\", \"corners_3d\": [[608.13, 190.16], [595.63, 190.09], [604.55, 189.21], [616.44, 189.27], [608.13, 175.82], [595.63, 175.81], [604.55, 175.66], [616.44, 175.67]]}, {\"category\": \"car\", \"corners_3d\": [[573.61, 190.35], [560.34, 190.26], [572.02, 189.47], [584.75, 189.55], [573.61, 175.69], [560.34, 175.67], [572.02, 175.54], [584.75, 175.56]]}]\n```" - } - ] - }, - { - "id": 1249, - "data_source": "Mono3DRefer", - "scene_name": "000846", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[516.14, 210.25], [513.73, 211.14], [500.67, 211.13], [503.39, 210.24], [516.14, 171.02], [513.73, 170.98], [500.67, 170.98], [503.39, 171.02]]}]\n```" - } - ] - }, - { - "id": 1250, - "data_source": "Mono3DRefer", - "scene_name": "000846", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[824.96, 218.29], [826.26, 216.86], [883.03, 217.97], [883.59, 219.48], [824.96, 172.84], [826.26, 172.84], [883.03, 172.84], [883.59, 172.84]]}]\n```" - } - ] - }, - { - "id": 1251, - "data_source": "Mono3DRefer", - "scene_name": "000846", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[735.99, 236.25], [799.89, 236.43], [851.02, 255.76], [767.58, 255.47], [735.99, 179.22], [799.89, 179.24], [851.02, 181.18], [767.58, 181.15]]}, {\"category\": \"car\", \"corners_3d\": [[661.73, 195.32], [651.91, 194.6], [707.01, 194.0], [718.31, 194.67], [661.73, 173.01], [651.91, 173.0], [707.01, 173.0], [718.31, 173.0]]}, {\"category\": \"car\", \"corners_3d\": [[829.32, 191.4], [828.22, 190.81], [891.71, 191.18], [894.99, 191.8], [829.32, 167.42], [828.22, 167.59], [891.71, 167.48], [894.99, 167.3]]}, {\"category\": \"car\", \"corners_3d\": [[985.8, 190.19], [979.68, 189.61], [1059.18, 190.04], [1068.29, 190.65], [985.8, 167.17], [979.68, 167.36], [1059.18, 167.22], [1068.29, 167.02]]}, {\"category\": \"van\", \"corners_3d\": [[1175.71, 188.22], [1161.33, 187.65], [1235.67, 187.97], [1253.29, 188.56], [1175.71, 157.64], [1161.33, 158.21], [1235.67, 157.89], [1253.29, 157.3]]}, {\"category\": \"car\", \"corners_3d\": [[427.74, 202.81], [413.98, 203.83], [366.21, 203.13], [381.21, 202.16], [427.74, 179.56], [413.98, 179.79], [366.21, 179.63], [381.21, 179.41]]}, {\"category\": \"car\", \"corners_3d\": [[608.13, 190.16], [595.63, 190.09], [604.55, 189.21], [616.44, 189.27], [608.13, 175.82], [595.63, 175.81], [604.55, 175.66], [616.44, 175.67]]}, {\"category\": \"car\", \"corners_3d\": [[573.61, 190.35], [560.34, 190.26], [572.02, 189.47], [584.75, 189.55], [573.61, 175.69], [560.34, 175.67], [572.02, 175.54], [584.75, 175.56]]}]\n```" - } - ] - }, - { - "id": 1252, - "data_source": "Mono3DRefer", - "scene_name": "000846", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[516.14, 210.25], [513.73, 211.14], [500.67, 211.13], [503.39, 210.24], [516.14, 171.02], [513.73, 170.98], [500.67, 170.98], [503.39, 171.02]]}, {\"category\": \"van\", \"corners_3d\": [[1175.71, 188.22], [1161.33, 187.65], [1235.67, 187.97], [1253.29, 188.56], [1175.71, 157.64], [1161.33, 158.21], [1235.67, 157.89], [1253.29, 157.3]]}]\n```" - } - ] - }, - { - "id": 1253, - "data_source": "Mono3DRefer", - "scene_name": "000846", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[824.96, 218.29], [826.26, 216.86], [883.03, 217.97], [883.59, 219.48], [824.96, 172.84], [826.26, 172.84], [883.03, 172.84], [883.59, 172.84]]}, {\"category\": \"van\", \"corners_3d\": [[1175.71, 188.22], [1161.33, 187.65], [1235.67, 187.97], [1253.29, 188.56], [1175.71, 157.64], [1161.33, 158.21], [1235.67, 157.89], [1253.29, 157.3]]}]\n```" - } - ] - }, - { - "id": 1254, - "data_source": "Mono3DRefer", - "scene_name": "000846", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[735.99, 236.25], [799.89, 236.43], [851.02, 255.76], [767.58, 255.47], [735.99, 179.22], [799.89, 179.24], [851.02, 181.18], [767.58, 181.15]]}, {\"category\": \"car\", \"corners_3d\": [[661.73, 195.32], [651.91, 194.6], [707.01, 194.0], [718.31, 194.67], [661.73, 173.01], [651.91, 173.0], [707.01, 173.0], [718.31, 173.0]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[516.14, 210.25], [513.73, 211.14], [500.67, 211.13], [503.39, 210.24], [516.14, 171.02], [513.73, 170.98], [500.67, 170.98], [503.39, 171.02]]}, {\"category\": \"car\", \"corners_3d\": [[829.32, 191.4], [828.22, 190.81], [891.71, 191.18], [894.99, 191.8], [829.32, 167.42], [828.22, 167.59], [891.71, 167.48], [894.99, 167.3]]}, {\"category\": \"car\", \"corners_3d\": [[985.8, 190.19], [979.68, 189.61], [1059.18, 190.04], [1068.29, 190.65], [985.8, 167.17], [979.68, 167.36], [1059.18, 167.22], [1068.29, 167.02]]}, {\"category\": \"car\", \"corners_3d\": [[427.74, 202.81], [413.98, 203.83], [366.21, 203.13], [381.21, 202.16], [427.74, 179.56], [413.98, 179.79], [366.21, 179.63], [381.21, 179.41]]}, {\"category\": \"car\", \"corners_3d\": [[608.13, 190.16], [595.63, 190.09], [604.55, 189.21], [616.44, 189.27], [608.13, 175.82], [595.63, 175.81], [604.55, 175.66], [616.44, 175.67]]}, {\"category\": \"car\", \"corners_3d\": [[573.61, 190.35], [560.34, 190.26], [572.02, 189.47], [584.75, 189.55], [573.61, 175.69], [560.34, 175.67], [572.02, 175.54], [584.75, 175.56]]}]\n```" - } - ] - }, - { - "id": 1255, - "data_source": "Mono3DRefer", - "scene_name": "000846", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[735.99, 236.25], [799.89, 236.43], [851.02, 255.76], [767.58, 255.47], [735.99, 179.22], [799.89, 179.24], [851.02, 181.18], [767.58, 181.15]]}, {\"category\": \"car\", \"corners_3d\": [[661.73, 195.32], [651.91, 194.6], [707.01, 194.0], [718.31, 194.67], [661.73, 173.01], [651.91, 173.0], [707.01, 173.0], [718.31, 173.0]]}, {\"category\": \"cyclist\", \"corners_3d\": [[824.96, 218.29], [826.26, 216.86], [883.03, 217.97], [883.59, 219.48], [824.96, 172.84], [826.26, 172.84], [883.03, 172.84], [883.59, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[829.32, 191.4], [828.22, 190.81], [891.71, 191.18], [894.99, 191.8], [829.32, 167.42], [828.22, 167.59], [891.71, 167.48], [894.99, 167.3]]}, {\"category\": \"car\", \"corners_3d\": [[985.8, 190.19], [979.68, 189.61], [1059.18, 190.04], [1068.29, 190.65], [985.8, 167.17], [979.68, 167.36], [1059.18, 167.22], [1068.29, 167.02]]}, {\"category\": \"car\", \"corners_3d\": [[427.74, 202.81], [413.98, 203.83], [366.21, 203.13], [381.21, 202.16], [427.74, 179.56], [413.98, 179.79], [366.21, 179.63], [381.21, 179.41]]}, {\"category\": \"car\", \"corners_3d\": [[608.13, 190.16], [595.63, 190.09], [604.55, 189.21], [616.44, 189.27], [608.13, 175.82], [595.63, 175.81], [604.55, 175.66], [616.44, 175.67]]}, {\"category\": \"car\", \"corners_3d\": [[573.61, 190.35], [560.34, 190.26], [572.02, 189.47], [584.75, 189.55], [573.61, 175.69], [560.34, 175.67], [572.02, 175.54], [584.75, 175.56]]}]\n```" - } - ] - }, - { - "id": 1256, - "data_source": "Mono3DRefer", - "scene_name": "000846", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[516.14, 210.25], [513.73, 211.14], [500.67, 211.13], [503.39, 210.24], [516.14, 171.02], [513.73, 170.98], [500.67, 170.98], [503.39, 171.02]]}, {\"category\": \"cyclist\", \"corners_3d\": [[824.96, 218.29], [826.26, 216.86], [883.03, 217.97], [883.59, 219.48], [824.96, 172.84], [826.26, 172.84], [883.03, 172.84], [883.59, 172.84]]}]\n```" - } - ] - }, - { - "id": 1257, - "data_source": "Mono3DRefer", - "scene_name": "000846", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[735.99, 236.25], [799.89, 236.43], [851.02, 255.76], [767.58, 255.47], [735.99, 179.22], [799.89, 179.24], [851.02, 181.18], [767.58, 181.15]]}, {\"category\": \"car\", \"corners_3d\": [[661.73, 195.32], [651.91, 194.6], [707.01, 194.0], [718.31, 194.67], [661.73, 173.01], [651.91, 173.0], [707.01, 173.0], [718.31, 173.0]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[516.14, 210.25], [513.73, 211.14], [500.67, 211.13], [503.39, 210.24], [516.14, 171.02], [513.73, 170.98], [500.67, 170.98], [503.39, 171.02]]}, {\"category\": \"car\", \"corners_3d\": [[829.32, 191.4], [828.22, 190.81], [891.71, 191.18], [894.99, 191.8], [829.32, 167.42], [828.22, 167.59], [891.71, 167.48], [894.99, 167.3]]}, {\"category\": \"car\", \"corners_3d\": [[985.8, 190.19], [979.68, 189.61], [1059.18, 190.04], [1068.29, 190.65], [985.8, 167.17], [979.68, 167.36], [1059.18, 167.22], [1068.29, 167.02]]}, {\"category\": \"van\", \"corners_3d\": [[1175.71, 188.22], [1161.33, 187.65], [1235.67, 187.97], [1253.29, 188.56], [1175.71, 157.64], [1161.33, 158.21], [1235.67, 157.89], [1253.29, 157.3]]}, {\"category\": \"car\", \"corners_3d\": [[427.74, 202.81], [413.98, 203.83], [366.21, 203.13], [381.21, 202.16], [427.74, 179.56], [413.98, 179.79], [366.21, 179.63], [381.21, 179.41]]}, {\"category\": \"car\", \"corners_3d\": [[608.13, 190.16], [595.63, 190.09], [604.55, 189.21], [616.44, 189.27], [608.13, 175.82], [595.63, 175.81], [604.55, 175.66], [616.44, 175.67]]}, {\"category\": \"car\", \"corners_3d\": [[573.61, 190.35], [560.34, 190.26], [572.02, 189.47], [584.75, 189.55], [573.61, 175.69], [560.34, 175.67], [572.02, 175.54], [584.75, 175.56]]}]\n```" - } - ] - }, - { - "id": 1258, - "data_source": "Mono3DRefer", - "scene_name": "000846", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[735.99, 236.25], [799.89, 236.43], [851.02, 255.76], [767.58, 255.47], [735.99, 179.22], [799.89, 179.24], [851.02, 181.18], [767.58, 181.15]]}, {\"category\": \"car\", \"corners_3d\": [[661.73, 195.32], [651.91, 194.6], [707.01, 194.0], [718.31, 194.67], [661.73, 173.01], [651.91, 173.0], [707.01, 173.0], [718.31, 173.0]]}, {\"category\": \"cyclist\", \"corners_3d\": [[824.96, 218.29], [826.26, 216.86], [883.03, 217.97], [883.59, 219.48], [824.96, 172.84], [826.26, 172.84], [883.03, 172.84], [883.59, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[829.32, 191.4], [828.22, 190.81], [891.71, 191.18], [894.99, 191.8], [829.32, 167.42], [828.22, 167.59], [891.71, 167.48], [894.99, 167.3]]}, {\"category\": \"car\", \"corners_3d\": [[985.8, 190.19], [979.68, 189.61], [1059.18, 190.04], [1068.29, 190.65], [985.8, 167.17], [979.68, 167.36], [1059.18, 167.22], [1068.29, 167.02]]}, {\"category\": \"van\", \"corners_3d\": [[1175.71, 188.22], [1161.33, 187.65], [1235.67, 187.97], [1253.29, 188.56], [1175.71, 157.64], [1161.33, 158.21], [1235.67, 157.89], [1253.29, 157.3]]}, {\"category\": \"car\", \"corners_3d\": [[427.74, 202.81], [413.98, 203.83], [366.21, 203.13], [381.21, 202.16], [427.74, 179.56], [413.98, 179.79], [366.21, 179.63], [381.21, 179.41]]}, {\"category\": \"car\", \"corners_3d\": [[608.13, 190.16], [595.63, 190.09], [604.55, 189.21], [616.44, 189.27], [608.13, 175.82], [595.63, 175.81], [604.55, 175.66], [616.44, 175.67]]}, {\"category\": \"car\", \"corners_3d\": [[573.61, 190.35], [560.34, 190.26], [572.02, 189.47], [584.75, 189.55], [573.61, 175.69], [560.34, 175.67], [572.02, 175.54], [584.75, 175.56]]}]\n```" - } - ] - }, - { - "id": 1259, - "data_source": "Mono3DRefer", - "scene_name": "000846", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[516.14, 210.25], [513.73, 211.14], [500.67, 211.13], [503.39, 210.24], [516.14, 171.02], [513.73, 170.98], [500.67, 170.98], [503.39, 171.02]]}, {\"category\": \"cyclist\", \"corners_3d\": [[824.96, 218.29], [826.26, 216.86], [883.03, 217.97], [883.59, 219.48], [824.96, 172.84], [826.26, 172.84], [883.03, 172.84], [883.59, 172.84]]}, {\"category\": \"van\", \"corners_3d\": [[1175.71, 188.22], [1161.33, 187.65], [1235.67, 187.97], [1253.29, 188.56], [1175.71, 157.64], [1161.33, 158.21], [1235.67, 157.89], [1253.29, 157.3]]}]\n```" - } - ] - }, - { - "id": 1260, - "data_source": "Mono3DRefer", - "scene_name": "000846", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[735.99, 236.25], [799.89, 236.43], [851.02, 255.76], [767.58, 255.47], [735.99, 179.22], [799.89, 179.24], [851.02, 181.18], [767.58, 181.15]]}, {\"category\": \"car\", \"corners_3d\": [[661.73, 195.32], [651.91, 194.6], [707.01, 194.0], [718.31, 194.67], [661.73, 173.01], [651.91, 173.0], [707.01, 173.0], [718.31, 173.0]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[516.14, 210.25], [513.73, 211.14], [500.67, 211.13], [503.39, 210.24], [516.14, 171.02], [513.73, 170.98], [500.67, 170.98], [503.39, 171.02]]}, {\"category\": \"cyclist\", \"corners_3d\": [[824.96, 218.29], [826.26, 216.86], [883.03, 217.97], [883.59, 219.48], [824.96, 172.84], [826.26, 172.84], [883.03, 172.84], [883.59, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[829.32, 191.4], [828.22, 190.81], [891.71, 191.18], [894.99, 191.8], [829.32, 167.42], [828.22, 167.59], [891.71, 167.48], [894.99, 167.3]]}, {\"category\": \"car\", \"corners_3d\": [[985.8, 190.19], [979.68, 189.61], [1059.18, 190.04], [1068.29, 190.65], [985.8, 167.17], [979.68, 167.36], [1059.18, 167.22], [1068.29, 167.02]]}, {\"category\": \"car\", \"corners_3d\": [[427.74, 202.81], [413.98, 203.83], [366.21, 203.13], [381.21, 202.16], [427.74, 179.56], [413.98, 179.79], [366.21, 179.63], [381.21, 179.41]]}, {\"category\": \"car\", \"corners_3d\": [[608.13, 190.16], [595.63, 190.09], [604.55, 189.21], [616.44, 189.27], [608.13, 175.82], [595.63, 175.81], [604.55, 175.66], [616.44, 175.67]]}, {\"category\": \"car\", \"corners_3d\": [[573.61, 190.35], [560.34, 190.26], [572.02, 189.47], [584.75, 189.55], [573.61, 175.69], [560.34, 175.67], [572.02, 175.54], [584.75, 175.56]]}]\n```" - } - ] - }, - { - "id": 1261, - "data_source": "Mono3DRefer", - "scene_name": "000846", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[735.99, 236.25], [799.89, 236.43], [851.02, 255.76], [767.58, 255.47], [735.99, 179.22], [799.89, 179.24], [851.02, 181.18], [767.58, 181.15]]}, {\"category\": \"car\", \"corners_3d\": [[661.73, 195.32], [651.91, 194.6], [707.01, 194.0], [718.31, 194.67], [661.73, 173.01], [651.91, 173.0], [707.01, 173.0], [718.31, 173.0]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[516.14, 210.25], [513.73, 211.14], [500.67, 211.13], [503.39, 210.24], [516.14, 171.02], [513.73, 170.98], [500.67, 170.98], [503.39, 171.02]]}, {\"category\": \"cyclist\", \"corners_3d\": [[824.96, 218.29], [826.26, 216.86], [883.03, 217.97], [883.59, 219.48], [824.96, 172.84], [826.26, 172.84], [883.03, 172.84], [883.59, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[829.32, 191.4], [828.22, 190.81], [891.71, 191.18], [894.99, 191.8], [829.32, 167.42], [828.22, 167.59], [891.71, 167.48], [894.99, 167.3]]}, {\"category\": \"car\", \"corners_3d\": [[985.8, 190.19], [979.68, 189.61], [1059.18, 190.04], [1068.29, 190.65], [985.8, 167.17], [979.68, 167.36], [1059.18, 167.22], [1068.29, 167.02]]}, {\"category\": \"van\", \"corners_3d\": [[1175.71, 188.22], [1161.33, 187.65], [1235.67, 187.97], [1253.29, 188.56], [1175.71, 157.64], [1161.33, 158.21], [1235.67, 157.89], [1253.29, 157.3]]}, {\"category\": \"car\", \"corners_3d\": [[427.74, 202.81], [413.98, 203.83], [366.21, 203.13], [381.21, 202.16], [427.74, 179.56], [413.98, 179.79], [366.21, 179.63], [381.21, 179.41]]}, {\"category\": \"car\", \"corners_3d\": [[608.13, 190.16], [595.63, 190.09], [604.55, 189.21], [616.44, 189.27], [608.13, 175.82], [595.63, 175.81], [604.55, 175.66], [616.44, 175.67]]}, {\"category\": \"car\", \"corners_3d\": [[573.61, 190.35], [560.34, 190.26], [572.02, 189.47], [584.75, 189.55], [573.61, 175.69], [560.34, 175.67], [572.02, 175.54], [584.75, 175.56]]}]\n```" - } - ] - }, - { - "id": 1262, - "data_source": "Mono3DRefer", - "scene_name": "000848", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[838.63, 217.45], [808.14, 214.1], [934.12, 212.2], [972.8, 215.23], [838.63, 144.55], [808.14, 146.68], [934.12, 147.88], [972.8, 145.96]]}]\n```" - } - ] - }, - { - "id": 1263, - "data_source": "Mono3DRefer", - "scene_name": "000848", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-5.68, 256.13], [-6.33, 248.36], [252.32, 234.92], [274.31, 240.07], [-5.68, 168.73], [-6.33, 169.11], [252.32, 169.78], [274.31, 169.52]]}, {\"category\": \"car\", \"corners_3d\": [[1136.4, 238.64], [1238.43, 246.08], [1127.01, 256.85], [1024.48, 247.21], [1136.4, 157.0], [1238.43, 155.21], [1127.01, 152.62], [1024.48, 154.94]]}, {\"category\": \"car\", \"corners_3d\": [[1210.86, 207.49], [1162.78, 205.33], [1252.83, 204.17], [1304.83, 206.17], [1210.86, 156.95], [1162.78, 157.94], [1252.83, 158.48], [1304.83, 157.56]]}]\n```" - } - ] - }, - { - "id": 1264, - "data_source": "Mono3DRefer", - "scene_name": "000848", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-5.68, 256.13], [-6.33, 248.36], [252.32, 234.92], [274.31, 240.07], [-5.68, 168.73], [-6.33, 169.11], [252.32, 169.78], [274.31, 169.52]]}, {\"category\": \"car\", \"corners_3d\": [[1136.4, 238.64], [1238.43, 246.08], [1127.01, 256.85], [1024.48, 247.21], [1136.4, 157.0], [1238.43, 155.21], [1127.01, 152.62], [1024.48, 154.94]]}, {\"category\": \"van\", \"corners_3d\": [[838.63, 217.45], [808.14, 214.1], [934.12, 212.2], [972.8, 215.23], [838.63, 144.55], [808.14, 146.68], [934.12, 147.88], [972.8, 145.96]]}, {\"category\": \"car\", \"corners_3d\": [[1210.86, 207.49], [1162.78, 205.33], [1252.83, 204.17], [1304.83, 206.17], [1210.86, 156.95], [1162.78, 157.94], [1252.83, 158.48], [1304.83, 157.56]]}]\n```" - } - ] - }, - { - "id": 1265, - "data_source": "Mono3DRefer", - "scene_name": "000849", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1137.91, 411.42], [906.59, 409.23], [790.22, 309.56], [922.86, 310.28], [1137.91, 199.79], [906.59, 199.54], [790.22, 188.29], [922.86, 188.37]]}, {\"category\": \"car\", \"corners_3d\": [[626.5, 207.03], [590.24, 206.91], [598.63, 202.79], [630.51, 202.88], [626.5, 174.41], [590.24, 174.4], [598.63, 174.22], [630.51, 174.22]]}, {\"category\": \"car\", \"corners_3d\": [[374.09, 224.93], [391.58, 221.08], [526.74, 221.09], [520.04, 224.94], [374.09, 174.56], [391.58, 174.43], [526.74, 174.43], [520.04, 174.56]]}]\n```" - } - ] - }, - { - "id": 1266, - "data_source": "Mono3DRefer", - "scene_name": "000849", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[595.98, 221.57], [540.75, 221.2], [561.1, 214.2], [608.42, 214.47], [595.98, 165.23], [540.75, 165.29], [561.1, 166.38], [608.42, 166.34]]}, {\"category\": \"van\", \"corners_3d\": [[710.32, 216.25], [754.45, 216.41], [770.87, 223.5], [719.53, 223.28], [710.32, 168.83], [754.45, 168.81], [770.87, 168.15], [719.53, 168.17]]}]\n```" - } - ] - }, - { - "id": 1267, - "data_source": "Mono3DRefer", - "scene_name": "000849", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1137.91, 411.42], [906.59, 409.23], [790.22, 309.56], [922.86, 310.28], [1137.91, 199.79], [906.59, 199.54], [790.22, 188.29], [922.86, 188.37]]}, {\"category\": \"van\", \"corners_3d\": [[595.98, 221.57], [540.75, 221.2], [561.1, 214.2], [608.42, 214.47], [595.98, 165.23], [540.75, 165.29], [561.1, 166.38], [608.42, 166.34]]}, {\"category\": \"van\", \"corners_3d\": [[710.32, 216.25], [754.45, 216.41], [770.87, 223.5], [719.53, 223.28], [710.32, 168.83], [754.45, 168.81], [770.87, 168.15], [719.53, 168.17]]}, {\"category\": \"car\", \"corners_3d\": [[626.5, 207.03], [590.24, 206.91], [598.63, 202.79], [630.51, 202.88], [626.5, 174.41], [590.24, 174.4], [598.63, 174.22], [630.51, 174.22]]}, {\"category\": \"car\", \"corners_3d\": [[374.09, 224.93], [391.58, 221.08], [526.74, 221.09], [520.04, 224.94], [374.09, 174.56], [391.58, 174.43], [526.74, 174.43], [520.04, 174.56]]}]\n```" - } - ] - }, - { - "id": 1268, - "data_source": "Mono3DRefer", - "scene_name": "000851", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[533.37, 215.66], [569.13, 215.26], [581.35, 220.57], [541.21, 221.07], [533.37, 181.0], [569.13, 180.93], [581.35, 181.94], [541.21, 182.03]]}, {\"category\": \"car\", \"corners_3d\": [[462.59, 205.34], [434.21, 205.48], [442.03, 202.79], [468.05, 202.68], [462.59, 177.02], [434.21, 177.04], [442.03, 176.69], [468.05, 176.68]]}]\n```" - } - ] - }, - { - "id": 1269, - "data_source": "Mono3DRefer", - "scene_name": "000854", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[281.83, 227.29], [236.87, 227.23], [281.38, 220.97], [321.2, 221.02], [281.83, 187.98], [236.87, 187.97], [281.38, 186.23], [321.2, 186.24]]}, {\"category\": \"car\", \"corners_3d\": [[434.96, 197.46], [412.77, 197.44], [427.4, 195.78], [448.11, 195.8], [434.96, 179.18], [412.77, 179.17], [427.4, 178.75], [448.11, 178.75]]}]\n```" - } - ] - }, - { - "id": 1270, - "data_source": "Mono3DRefer", - "scene_name": "000854", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[665.59, 213.78], [729.68, 213.78], [815.09, 243.02], [705.21, 243.01], [665.59, 144.34], [729.68, 144.34], [815.09, 123.96], [705.21, 123.97]]}]\n```" - } - ] - }, - { - "id": 1271, - "data_source": "Mono3DRefer", - "scene_name": "000854", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[665.59, 213.78], [729.68, 213.78], [815.09, 243.02], [705.21, 243.01], [665.59, 144.34], [729.68, 144.34], [815.09, 123.96], [705.21, 123.97]]}, {\"category\": \"car\", \"corners_3d\": [[281.83, 227.29], [236.87, 227.23], [281.38, 220.97], [321.2, 221.02], [281.83, 187.98], [236.87, 187.97], [281.38, 186.23], [321.2, 186.24]]}, {\"category\": \"car\", \"corners_3d\": [[434.96, 197.46], [412.77, 197.44], [427.4, 195.78], [448.11, 195.8], [434.96, 179.18], [412.77, 179.17], [427.4, 178.75], [448.11, 178.75]]}]\n```" - } - ] - }, - { - "id": 1272, - "data_source": "Mono3DRefer", - "scene_name": "000856", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[758.85, 326.3], [894.34, 326.61], [1115.38, 449.31], [871.36, 448.31], [758.85, 201.5], [894.34, 201.55], [1115.38, 224.46], [871.36, 224.27]]}, {\"category\": \"car\", \"corners_3d\": [[530.66, 264.97], [459.37, 264.52], [494.5, 247.18], [552.39, 247.47], [530.66, 192.38], [459.37, 192.28], [494.5, 188.61], [552.39, 188.67]]}, {\"category\": \"car\", \"corners_3d\": [[573.22, 241.44], [515.76, 241.22], [536.35, 229.69], [584.15, 229.84], [573.22, 174.93], [515.76, 174.92], [536.35, 174.58], [584.15, 174.58]]}, {\"category\": \"car\", \"corners_3d\": [[603.95, 229.22], [558.54, 228.94], [573.43, 221.22], [612.63, 221.43], [603.95, 179.62], [558.54, 179.59], [573.43, 178.66], [612.63, 178.69]]}, {\"category\": \"car\", \"corners_3d\": [[648.94, 212.81], [622.01, 212.37], [638.85, 209.08], [663.69, 209.45], [648.94, 186.91], [622.01, 186.76], [638.85, 185.6], [663.69, 185.73]]}]\n```" - } - ] - }, - { - "id": 1273, - "data_source": "Mono3DRefer", - "scene_name": "000858", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[376.02, 239.65], [441.53, 239.59], [390.78, 260.58], [304.6, 260.67], [376.02, 156.68], [441.53, 156.69], [390.78, 151.61], [304.6, 151.59]]}]\n```" - } - ] - }, - { - "id": 1274, - "data_source": "Mono3DRefer", - "scene_name": "000858", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[115.8, 265.1], [49.45, 265.43], [144.1, 248.92], [198.37, 248.69], [115.8, 207.17], [49.45, 207.29], [144.1, 201.15], [198.37, 201.07]]}]\n```" - } - ] - }, - { - "id": 1275, - "data_source": "Mono3DRefer", - "scene_name": "000858", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[376.02, 239.65], [441.53, 239.59], [390.78, 260.58], [304.6, 260.67], [376.02, 156.68], [441.53, 156.69], [390.78, 151.61], [304.6, 151.59]]}, {\"category\": \"car\", \"corners_3d\": [[115.8, 265.1], [49.45, 265.43], [144.1, 248.92], [198.37, 248.69], [115.8, 207.17], [49.45, 207.29], [144.1, 201.15], [198.37, 201.07]]}]\n```" - } - ] - }, - { - "id": 1276, - "data_source": "Mono3DRefer", - "scene_name": "000859", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[791.86, 260.85], [883.26, 262.17], [962.12, 300.01], [831.16, 297.37], [791.86, 177.61], [883.26, 177.68], [962.12, 179.72], [831.16, 179.58]]}, {\"category\": \"car\", \"corners_3d\": [[831.71, 250.3], [756.05, 249.67], [738.88, 234.89], [799.78, 235.3], [831.71, 179.75], [756.05, 179.7], [738.88, 178.38], [799.78, 178.42]]}, {\"category\": \"car\", \"corners_3d\": [[618.03, 208.4], [585.16, 208.02], [604.76, 204.45], [634.44, 204.75], [618.03, 174.58], [585.16, 174.56], [604.76, 174.39], [634.44, 174.4]]}, {\"category\": \"car\", \"corners_3d\": [[648.63, 202.67], [622.87, 202.41], [637.49, 199.87], [661.15, 200.09], [648.63, 178.94], [622.87, 178.89], [637.49, 178.37], [661.15, 178.41]]}, {\"category\": \"car\", \"corners_3d\": [[701.52, 194.01], [682.08, 193.84], [690.15, 192.71], [708.6, 192.86], [701.52, 177.13], [682.08, 177.09], [690.15, 176.87], [708.6, 176.9]]}]\n```" - } - ] - }, - { - "id": 1277, - "data_source": "Mono3DRefer", - "scene_name": "000859", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[716.27, 210.63], [731.04, 210.75], [731.69, 212.85], [716.1, 212.72], [716.27, 173.06], [731.04, 173.06], [731.69, 173.07], [716.1, 173.07]]}]\n```" - } - ] - }, - { - "id": 1278, - "data_source": "Mono3DRefer", - "scene_name": "000859", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[791.86, 260.85], [883.26, 262.17], [962.12, 300.01], [831.16, 297.37], [791.86, 177.61], [883.26, 177.68], [962.12, 179.72], [831.16, 179.58]]}, {\"category\": \"car\", \"corners_3d\": [[831.71, 250.3], [756.05, 249.67], [738.88, 234.89], [799.78, 235.3], [831.71, 179.75], [756.05, 179.7], [738.88, 178.38], [799.78, 178.42]]}, {\"category\": \"car\", \"corners_3d\": [[618.03, 208.4], [585.16, 208.02], [604.76, 204.45], [634.44, 204.75], [618.03, 174.58], [585.16, 174.56], [604.76, 174.39], [634.44, 174.4]]}, {\"category\": \"car\", \"corners_3d\": [[648.63, 202.67], [622.87, 202.41], [637.49, 199.87], [661.15, 200.09], [648.63, 178.94], [622.87, 178.89], [637.49, 178.37], [661.15, 178.41]]}, {\"category\": \"car\", \"corners_3d\": [[701.52, 194.01], [682.08, 193.84], [690.15, 192.71], [708.6, 192.86], [701.52, 177.13], [682.08, 177.09], [690.15, 176.87], [708.6, 176.9]]}, {\"category\": \"cyclist\", \"corners_3d\": [[716.27, 210.63], [731.04, 210.75], [731.69, 212.85], [716.1, 212.72], [716.27, 173.06], [731.04, 173.06], [731.69, 173.07], [716.1, 173.07]]}]\n```" - } - ] - }, - { - "id": 1279, - "data_source": "Mono3DRefer", - "scene_name": "000861", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[534.24, 227.2], [555.44, 227.17], [553.84, 229.32], [531.8, 229.35], [534.24, 177.27], [555.44, 177.26], [553.84, 177.44], [531.8, 177.44]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[518.88, 242.74], [488.9, 243.0], [491.04, 240.27], [519.85, 240.03], [518.88, 170.72], [488.9, 170.71], [491.04, 170.8], [519.85, 170.8]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[740.01, 221.77], [721.32, 221.74], [717.23, 219.7], [735.14, 219.73], [740.01, 165.85], [721.32, 165.86], [717.23, 166.15], [735.14, 166.15]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[760.43, 219.89], [738.16, 219.92], [732.55, 218.08], [753.95, 218.05], [760.43, 166.75], [738.16, 166.74], [732.55, 166.98], [753.95, 166.98]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1193.75, 283.71], [1251.61, 294.3], [1210.01, 294.52], [1155.69, 283.89], [1193.75, 154.85], [1251.61, 153.13], [1210.01, 153.09], [1155.69, 154.82]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1164.54, 277.6], [1202.96, 284.85], [1160.57, 284.85], [1124.9, 277.6], [1164.54, 163.63], [1202.96, 162.99], [1160.57, 162.99], [1124.9, 163.63]]}]\n```" - } - ] - }, - { - "id": 1280, - "data_source": "Mono3DRefer", - "scene_name": "000865", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[379.02, 386.62], [203.15, 385.61], [356.61, 307.93], [468.52, 308.33], [379.02, 217.27], [203.15, 217.06], [356.61, 200.92], [468.52, 201.0]]}, {\"category\": \"car\", \"corners_3d\": [[569.37, 228.23], [535.81, 228.23], [543.37, 222.51], [573.46, 222.51], [569.37, 199.53], [535.81, 199.53], [543.37, 196.77], [573.46, 196.77]]}]\n```" - } - ] - }, - { - "id": 1281, - "data_source": "Mono3DRefer", - "scene_name": "000870", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[657.74, 208.3], [695.08, 208.33], [704.93, 213.29], [662.37, 213.24], [657.74, 165.84], [695.08, 165.83], [704.93, 164.85], [662.37, 164.86]]}, {\"category\": \"van\", \"corners_3d\": [[458.7, 202.16], [436.28, 202.16], [449.38, 199.94], [470.1, 199.94], [458.7, 177.59], [436.28, 177.59], [449.38, 177.23], [470.1, 177.23]]}]\n```" - } - ] - }, - { - "id": 1282, - "data_source": "Mono3DRefer", - "scene_name": "000870", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[775.6, 221.19], [771.57, 219.59], [828.44, 219.81], [834.43, 221.42], [775.6, 172.84], [771.57, 172.84], [828.44, 172.84], [834.43, 172.84]]}]\n```" - } - ] - }, - { - "id": 1283, - "data_source": "Mono3DRefer", - "scene_name": "000870", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1206.85, 280.02], [1171.07, 280.54], [1122.6, 272.22], [1155.85, 271.78], [1206.85, 159.94], [1171.07, 159.88], [1122.6, 160.88], [1155.85, 160.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[533.99, 218.98], [534.43, 220.06], [516.95, 220.21], [516.91, 219.12], [533.99, 170.24], [534.43, 170.18], [516.95, 170.17], [516.91, 170.23]]}]\n```" - } - ] - }, - { - "id": 1284, - "data_source": "Mono3DRefer", - "scene_name": "000870", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.43, 232.33], [666.67, 232.34], [678.69, 244.98], [602.01, 244.98], [603.43, 181.45], [666.67, 181.45], [678.69, 183.27], [602.01, 183.27]]}, {\"category\": \"car\", \"corners_3d\": [[1197.46, 230.22], [1148.32, 225.87], [1279.35, 225.03], [1338.25, 229.24], [1197.46, 178.54], [1148.32, 178.11], [1279.35, 178.03], [1338.25, 178.45]]}, {\"category\": \"car\", \"corners_3d\": [[376.04, 216.96], [338.24, 216.94], [367.56, 212.29], [401.38, 212.3], [376.04, 181.89], [338.24, 181.88], [367.56, 180.93], [401.38, 180.94]]}, {\"category\": \"car\", \"corners_3d\": [[476.01, 209.73], [500.26, 208.5], [550.07, 210.83], [526.0, 212.22], [476.01, 178.37], [500.26, 178.19], [550.07, 178.53], [526.0, 178.74]]}, {\"category\": \"car\", \"corners_3d\": [[588.41, 216.65], [627.34, 216.7], [627.71, 222.5], [583.64, 222.43], [588.41, 181.23], [627.34, 181.24], [627.71, 182.35], [583.64, 182.34]]}]\n```" - } - ] - }, - { - "id": 1285, - "data_source": "Mono3DRefer", - "scene_name": "000870", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[775.6, 221.19], [771.57, 219.59], [828.44, 219.81], [834.43, 221.42], [775.6, 172.84], [771.57, 172.84], [828.44, 172.84], [834.43, 172.84]]}, {\"category\": \"van\", \"corners_3d\": [[657.74, 208.3], [695.08, 208.33], [704.93, 213.29], [662.37, 213.24], [657.74, 165.84], [695.08, 165.83], [704.93, 164.85], [662.37, 164.86]]}, {\"category\": \"van\", \"corners_3d\": [[458.7, 202.16], [436.28, 202.16], [449.38, 199.94], [470.1, 199.94], [458.7, 177.59], [436.28, 177.59], [449.38, 177.23], [470.1, 177.23]]}]\n```" - } - ] - }, - { - "id": 1286, - "data_source": "Mono3DRefer", - "scene_name": "000870", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1206.85, 280.02], [1171.07, 280.54], [1122.6, 272.22], [1155.85, 271.78], [1206.85, 159.94], [1171.07, 159.88], [1122.6, 160.88], [1155.85, 160.93]]}, {\"category\": \"van\", \"corners_3d\": [[657.74, 208.3], [695.08, 208.33], [704.93, 213.29], [662.37, 213.24], [657.74, 165.84], [695.08, 165.83], [704.93, 164.85], [662.37, 164.86]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[533.99, 218.98], [534.43, 220.06], [516.95, 220.21], [516.91, 219.12], [533.99, 170.24], [534.43, 170.18], [516.95, 170.17], [516.91, 170.23]]}, {\"category\": \"van\", \"corners_3d\": [[458.7, 202.16], [436.28, 202.16], [449.38, 199.94], [470.1, 199.94], [458.7, 177.59], [436.28, 177.59], [449.38, 177.23], [470.1, 177.23]]}]\n```" - } - ] - }, - { - "id": 1287, - "data_source": "Mono3DRefer", - "scene_name": "000870", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.43, 232.33], [666.67, 232.34], [678.69, 244.98], [602.01, 244.98], [603.43, 181.45], [666.67, 181.45], [678.69, 183.27], [602.01, 183.27]]}, {\"category\": \"car\", \"corners_3d\": [[1197.46, 230.22], [1148.32, 225.87], [1279.35, 225.03], [1338.25, 229.24], [1197.46, 178.54], [1148.32, 178.11], [1279.35, 178.03], [1338.25, 178.45]]}, {\"category\": \"car\", \"corners_3d\": [[376.04, 216.96], [338.24, 216.94], [367.56, 212.29], [401.38, 212.3], [376.04, 181.89], [338.24, 181.88], [367.56, 180.93], [401.38, 180.94]]}, {\"category\": \"van\", \"corners_3d\": [[657.74, 208.3], [695.08, 208.33], [704.93, 213.29], [662.37, 213.24], [657.74, 165.84], [695.08, 165.83], [704.93, 164.85], [662.37, 164.86]]}, {\"category\": \"van\", \"corners_3d\": [[458.7, 202.16], [436.28, 202.16], [449.38, 199.94], [470.1, 199.94], [458.7, 177.59], [436.28, 177.59], [449.38, 177.23], [470.1, 177.23]]}, {\"category\": \"car\", \"corners_3d\": [[476.01, 209.73], [500.26, 208.5], [550.07, 210.83], [526.0, 212.22], [476.01, 178.37], [500.26, 178.19], [550.07, 178.53], [526.0, 178.74]]}, {\"category\": \"car\", \"corners_3d\": [[588.41, 216.65], [627.34, 216.7], [627.71, 222.5], [583.64, 222.43], [588.41, 181.23], [627.34, 181.24], [627.71, 182.35], [583.64, 182.34]]}]\n```" - } - ] - }, - { - "id": 1288, - "data_source": "Mono3DRefer", - "scene_name": "000870", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1206.85, 280.02], [1171.07, 280.54], [1122.6, 272.22], [1155.85, 271.78], [1206.85, 159.94], [1171.07, 159.88], [1122.6, 160.88], [1155.85, 160.93]]}, {\"category\": \"cyclist\", \"corners_3d\": [[775.6, 221.19], [771.57, 219.59], [828.44, 219.81], [834.43, 221.42], [775.6, 172.84], [771.57, 172.84], [828.44, 172.84], [834.43, 172.84]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[533.99, 218.98], [534.43, 220.06], [516.95, 220.21], [516.91, 219.12], [533.99, 170.24], [534.43, 170.18], [516.95, 170.17], [516.91, 170.23]]}]\n```" - } - ] - }, - { - "id": 1289, - "data_source": "Mono3DRefer", - "scene_name": "000870", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.43, 232.33], [666.67, 232.34], [678.69, 244.98], [602.01, 244.98], [603.43, 181.45], [666.67, 181.45], [678.69, 183.27], [602.01, 183.27]]}, {\"category\": \"car\", \"corners_3d\": [[1197.46, 230.22], [1148.32, 225.87], [1279.35, 225.03], [1338.25, 229.24], [1197.46, 178.54], [1148.32, 178.11], [1279.35, 178.03], [1338.25, 178.45]]}, {\"category\": \"car\", \"corners_3d\": [[376.04, 216.96], [338.24, 216.94], [367.56, 212.29], [401.38, 212.3], [376.04, 181.89], [338.24, 181.88], [367.56, 180.93], [401.38, 180.94]]}, {\"category\": \"cyclist\", \"corners_3d\": [[775.6, 221.19], [771.57, 219.59], [828.44, 219.81], [834.43, 221.42], [775.6, 172.84], [771.57, 172.84], [828.44, 172.84], [834.43, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[476.01, 209.73], [500.26, 208.5], [550.07, 210.83], [526.0, 212.22], [476.01, 178.37], [500.26, 178.19], [550.07, 178.53], [526.0, 178.74]]}, {\"category\": \"car\", \"corners_3d\": [[588.41, 216.65], [627.34, 216.7], [627.71, 222.5], [583.64, 222.43], [588.41, 181.23], [627.34, 181.24], [627.71, 182.35], [583.64, 182.34]]}]\n```" - } - ] - }, - { - "id": 1290, - "data_source": "Mono3DRefer", - "scene_name": "000870", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.43, 232.33], [666.67, 232.34], [678.69, 244.98], [602.01, 244.98], [603.43, 181.45], [666.67, 181.45], [678.69, 183.27], [602.01, 183.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1206.85, 280.02], [1171.07, 280.54], [1122.6, 272.22], [1155.85, 271.78], [1206.85, 159.94], [1171.07, 159.88], [1122.6, 160.88], [1155.85, 160.93]]}, {\"category\": \"car\", \"corners_3d\": [[1197.46, 230.22], [1148.32, 225.87], [1279.35, 225.03], [1338.25, 229.24], [1197.46, 178.54], [1148.32, 178.11], [1279.35, 178.03], [1338.25, 178.45]]}, {\"category\": \"car\", \"corners_3d\": [[376.04, 216.96], [338.24, 216.94], [367.56, 212.29], [401.38, 212.3], [376.04, 181.89], [338.24, 181.88], [367.56, 180.93], [401.38, 180.94]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[533.99, 218.98], [534.43, 220.06], [516.95, 220.21], [516.91, 219.12], [533.99, 170.24], [534.43, 170.18], [516.95, 170.17], [516.91, 170.23]]}, {\"category\": \"car\", \"corners_3d\": [[476.01, 209.73], [500.26, 208.5], [550.07, 210.83], [526.0, 212.22], [476.01, 178.37], [500.26, 178.19], [550.07, 178.53], [526.0, 178.74]]}, {\"category\": \"car\", \"corners_3d\": [[588.41, 216.65], [627.34, 216.7], [627.71, 222.5], [583.64, 222.43], [588.41, 181.23], [627.34, 181.24], [627.71, 182.35], [583.64, 182.34]]}]\n```" - } - ] - }, - { - "id": 1291, - "data_source": "Mono3DRefer", - "scene_name": "000870", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1206.85, 280.02], [1171.07, 280.54], [1122.6, 272.22], [1155.85, 271.78], [1206.85, 159.94], [1171.07, 159.88], [1122.6, 160.88], [1155.85, 160.93]]}, {\"category\": \"cyclist\", \"corners_3d\": [[775.6, 221.19], [771.57, 219.59], [828.44, 219.81], [834.43, 221.42], [775.6, 172.84], [771.57, 172.84], [828.44, 172.84], [834.43, 172.84]]}, {\"category\": \"van\", \"corners_3d\": [[657.74, 208.3], [695.08, 208.33], [704.93, 213.29], [662.37, 213.24], [657.74, 165.84], [695.08, 165.83], [704.93, 164.85], [662.37, 164.86]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[533.99, 218.98], [534.43, 220.06], [516.95, 220.21], [516.91, 219.12], [533.99, 170.24], [534.43, 170.18], [516.95, 170.17], [516.91, 170.23]]}, {\"category\": \"van\", \"corners_3d\": [[458.7, 202.16], [436.28, 202.16], [449.38, 199.94], [470.1, 199.94], [458.7, 177.59], [436.28, 177.59], [449.38, 177.23], [470.1, 177.23]]}]\n```" - } - ] - }, - { - "id": 1292, - "data_source": "Mono3DRefer", - "scene_name": "000870", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.43, 232.33], [666.67, 232.34], [678.69, 244.98], [602.01, 244.98], [603.43, 181.45], [666.67, 181.45], [678.69, 183.27], [602.01, 183.27]]}, {\"category\": \"car\", \"corners_3d\": [[1197.46, 230.22], [1148.32, 225.87], [1279.35, 225.03], [1338.25, 229.24], [1197.46, 178.54], [1148.32, 178.11], [1279.35, 178.03], [1338.25, 178.45]]}, {\"category\": \"car\", \"corners_3d\": [[376.04, 216.96], [338.24, 216.94], [367.56, 212.29], [401.38, 212.3], [376.04, 181.89], [338.24, 181.88], [367.56, 180.93], [401.38, 180.94]]}, {\"category\": \"cyclist\", \"corners_3d\": [[775.6, 221.19], [771.57, 219.59], [828.44, 219.81], [834.43, 221.42], [775.6, 172.84], [771.57, 172.84], [828.44, 172.84], [834.43, 172.84]]}, {\"category\": \"van\", \"corners_3d\": [[657.74, 208.3], [695.08, 208.33], [704.93, 213.29], [662.37, 213.24], [657.74, 165.84], [695.08, 165.83], [704.93, 164.85], [662.37, 164.86]]}, {\"category\": \"van\", \"corners_3d\": [[458.7, 202.16], [436.28, 202.16], [449.38, 199.94], [470.1, 199.94], [458.7, 177.59], [436.28, 177.59], [449.38, 177.23], [470.1, 177.23]]}, {\"category\": \"car\", \"corners_3d\": [[476.01, 209.73], [500.26, 208.5], [550.07, 210.83], [526.0, 212.22], [476.01, 178.37], [500.26, 178.19], [550.07, 178.53], [526.0, 178.74]]}, {\"category\": \"car\", \"corners_3d\": [[588.41, 216.65], [627.34, 216.7], [627.71, 222.5], [583.64, 222.43], [588.41, 181.23], [627.34, 181.24], [627.71, 182.35], [583.64, 182.34]]}]\n```" - } - ] - }, - { - "id": 1293, - "data_source": "Mono3DRefer", - "scene_name": "000870", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.43, 232.33], [666.67, 232.34], [678.69, 244.98], [602.01, 244.98], [603.43, 181.45], [666.67, 181.45], [678.69, 183.27], [602.01, 183.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1206.85, 280.02], [1171.07, 280.54], [1122.6, 272.22], [1155.85, 271.78], [1206.85, 159.94], [1171.07, 159.88], [1122.6, 160.88], [1155.85, 160.93]]}, {\"category\": \"car\", \"corners_3d\": [[1197.46, 230.22], [1148.32, 225.87], [1279.35, 225.03], [1338.25, 229.24], [1197.46, 178.54], [1148.32, 178.11], [1279.35, 178.03], [1338.25, 178.45]]}, {\"category\": \"car\", \"corners_3d\": [[376.04, 216.96], [338.24, 216.94], [367.56, 212.29], [401.38, 212.3], [376.04, 181.89], [338.24, 181.88], [367.56, 180.93], [401.38, 180.94]]}, {\"category\": \"van\", \"corners_3d\": [[657.74, 208.3], [695.08, 208.33], [704.93, 213.29], [662.37, 213.24], [657.74, 165.84], [695.08, 165.83], [704.93, 164.85], [662.37, 164.86]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[533.99, 218.98], [534.43, 220.06], [516.95, 220.21], [516.91, 219.12], [533.99, 170.24], [534.43, 170.18], [516.95, 170.17], [516.91, 170.23]]}, {\"category\": \"van\", \"corners_3d\": [[458.7, 202.16], [436.28, 202.16], [449.38, 199.94], [470.1, 199.94], [458.7, 177.59], [436.28, 177.59], [449.38, 177.23], [470.1, 177.23]]}, {\"category\": \"car\", \"corners_3d\": [[476.01, 209.73], [500.26, 208.5], [550.07, 210.83], [526.0, 212.22], [476.01, 178.37], [500.26, 178.19], [550.07, 178.53], [526.0, 178.74]]}, {\"category\": \"car\", \"corners_3d\": [[588.41, 216.65], [627.34, 216.7], [627.71, 222.5], [583.64, 222.43], [588.41, 181.23], [627.34, 181.24], [627.71, 182.35], [583.64, 182.34]]}]\n```" - } - ] - }, - { - "id": 1294, - "data_source": "Mono3DRefer", - "scene_name": "000870", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.43, 232.33], [666.67, 232.34], [678.69, 244.98], [602.01, 244.98], [603.43, 181.45], [666.67, 181.45], [678.69, 183.27], [602.01, 183.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1206.85, 280.02], [1171.07, 280.54], [1122.6, 272.22], [1155.85, 271.78], [1206.85, 159.94], [1171.07, 159.88], [1122.6, 160.88], [1155.85, 160.93]]}, {\"category\": \"car\", \"corners_3d\": [[1197.46, 230.22], [1148.32, 225.87], [1279.35, 225.03], [1338.25, 229.24], [1197.46, 178.54], [1148.32, 178.11], [1279.35, 178.03], [1338.25, 178.45]]}, {\"category\": \"car\", \"corners_3d\": [[376.04, 216.96], [338.24, 216.94], [367.56, 212.29], [401.38, 212.3], [376.04, 181.89], [338.24, 181.88], [367.56, 180.93], [401.38, 180.94]]}, {\"category\": \"cyclist\", \"corners_3d\": [[775.6, 221.19], [771.57, 219.59], [828.44, 219.81], [834.43, 221.42], [775.6, 172.84], [771.57, 172.84], [828.44, 172.84], [834.43, 172.84]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[533.99, 218.98], [534.43, 220.06], [516.95, 220.21], [516.91, 219.12], [533.99, 170.24], [534.43, 170.18], [516.95, 170.17], [516.91, 170.23]]}, {\"category\": \"car\", \"corners_3d\": [[476.01, 209.73], [500.26, 208.5], [550.07, 210.83], [526.0, 212.22], [476.01, 178.37], [500.26, 178.19], [550.07, 178.53], [526.0, 178.74]]}, {\"category\": \"car\", \"corners_3d\": [[588.41, 216.65], [627.34, 216.7], [627.71, 222.5], [583.64, 222.43], [588.41, 181.23], [627.34, 181.24], [627.71, 182.35], [583.64, 182.34]]}]\n```" - } - ] - }, - { - "id": 1295, - "data_source": "Mono3DRefer", - "scene_name": "000870", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.43, 232.33], [666.67, 232.34], [678.69, 244.98], [602.01, 244.98], [603.43, 181.45], [666.67, 181.45], [678.69, 183.27], [602.01, 183.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1206.85, 280.02], [1171.07, 280.54], [1122.6, 272.22], [1155.85, 271.78], [1206.85, 159.94], [1171.07, 159.88], [1122.6, 160.88], [1155.85, 160.93]]}, {\"category\": \"car\", \"corners_3d\": [[1197.46, 230.22], [1148.32, 225.87], [1279.35, 225.03], [1338.25, 229.24], [1197.46, 178.54], [1148.32, 178.11], [1279.35, 178.03], [1338.25, 178.45]]}, {\"category\": \"car\", \"corners_3d\": [[376.04, 216.96], [338.24, 216.94], [367.56, 212.29], [401.38, 212.3], [376.04, 181.89], [338.24, 181.88], [367.56, 180.93], [401.38, 180.94]]}, {\"category\": \"cyclist\", \"corners_3d\": [[775.6, 221.19], [771.57, 219.59], [828.44, 219.81], [834.43, 221.42], [775.6, 172.84], [771.57, 172.84], [828.44, 172.84], [834.43, 172.84]]}, {\"category\": \"van\", \"corners_3d\": [[657.74, 208.3], [695.08, 208.33], [704.93, 213.29], [662.37, 213.24], [657.74, 165.84], [695.08, 165.83], [704.93, 164.85], [662.37, 164.86]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[533.99, 218.98], [534.43, 220.06], [516.95, 220.21], [516.91, 219.12], [533.99, 170.24], [534.43, 170.18], [516.95, 170.17], [516.91, 170.23]]}, {\"category\": \"van\", \"corners_3d\": [[458.7, 202.16], [436.28, 202.16], [449.38, 199.94], [470.1, 199.94], [458.7, 177.59], [436.28, 177.59], [449.38, 177.23], [470.1, 177.23]]}, {\"category\": \"car\", \"corners_3d\": [[476.01, 209.73], [500.26, 208.5], [550.07, 210.83], [526.0, 212.22], [476.01, 178.37], [500.26, 178.19], [550.07, 178.53], [526.0, 178.74]]}, {\"category\": \"car\", \"corners_3d\": [[588.41, 216.65], [627.34, 216.7], [627.71, 222.5], [583.64, 222.43], [588.41, 181.23], [627.34, 181.24], [627.71, 182.35], [583.64, 182.34]]}]\n```" - } - ] - }, - { - "id": 1296, - "data_source": "Mono3DRefer", - "scene_name": "000875", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-154.71, 295.9], [-44.71, 276.71], [206.12, 273.96], [144.46, 292.06], [-154.71, 182.0], [-44.71, 180.57], [206.12, 180.37], [144.46, 181.71]]}, {\"category\": \"car\", \"corners_3d\": [[16.48, 266.73], [75.7, 256.02], [269.67, 253.9], [236.38, 264.03], [16.48, 181.94], [75.7, 180.9], [269.67, 180.7], [236.38, 181.68]]}, {\"category\": \"car\", \"corners_3d\": [[314.89, 240.42], [302.17, 246.86], [114.18, 250.8], [144.67, 243.69], [314.89, 169.22], [302.17, 168.87], [114.18, 168.66], [144.67, 169.04]]}, {\"category\": \"car\", \"corners_3d\": [[511.95, 213.93], [547.79, 213.34], [567.56, 218.58], [527.33, 219.33], [511.95, 177.46], [547.79, 177.39], [567.56, 177.98], [527.33, 178.06]]}]\n```" - } - ] - }, - { - "id": 1297, - "data_source": "Mono3DRefer", - "scene_name": "000877", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[214.43, 268.68], [117.67, 269.07], [223.94, 247.3], [298.54, 247.07], [214.43, 183.66], [117.67, 183.7], [223.94, 181.25], [298.54, 181.22]]}]\n```" - } - ] - }, - { - "id": 1298, - "data_source": "Mono3DRefer", - "scene_name": "000877", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[501.1, 202.19], [467.89, 202.24], [479.41, 199.22], [509.2, 199.18], [501.1, 161.94], [467.89, 161.92], [479.41, 163.04], [509.2, 163.06]]}]\n```" - } - ] - }, - { - "id": 1299, - "data_source": "Mono3DRefer", - "scene_name": "000877", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[410.0, 225.32], [369.96, 225.41], [396.48, 219.0], [431.6, 218.93], [410.0, 189.57], [369.96, 189.6], [396.48, 187.56], [431.6, 187.54]]}, {\"category\": \"car\", \"corners_3d\": [[461.41, 209.92], [434.54, 209.94], [443.85, 207.87], [469.22, 207.86], [461.41, 184.24], [434.54, 184.25], [443.85, 183.61], [469.22, 183.61]]}, {\"category\": \"car\", \"corners_3d\": [[594.11, 193.26], [610.67, 193.24], [611.87, 194.34], [594.41, 194.35], [594.11, 180.2], [610.67, 180.2], [611.87, 180.59], [594.41, 180.6]]}]\n```" - } - ] - }, - { - "id": 1300, - "data_source": "Mono3DRefer", - "scene_name": "000877", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[214.43, 268.68], [117.67, 269.07], [223.94, 247.3], [298.54, 247.07], [214.43, 183.66], [117.67, 183.7], [223.94, 181.25], [298.54, 181.22]]}, {\"category\": \"truck\", \"corners_3d\": [[501.1, 202.19], [467.89, 202.24], [479.41, 199.22], [509.2, 199.18], [501.1, 161.94], [467.89, 161.92], [479.41, 163.04], [509.2, 163.06]]}]\n```" - } - ] - }, - { - "id": 1301, - "data_source": "Mono3DRefer", - "scene_name": "000877", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[214.43, 268.68], [117.67, 269.07], [223.94, 247.3], [298.54, 247.07], [214.43, 183.66], [117.67, 183.7], [223.94, 181.25], [298.54, 181.22]]}, {\"category\": \"car\", \"corners_3d\": [[410.0, 225.32], [369.96, 225.41], [396.48, 219.0], [431.6, 218.93], [410.0, 189.57], [369.96, 189.6], [396.48, 187.56], [431.6, 187.54]]}, {\"category\": \"car\", \"corners_3d\": [[461.41, 209.92], [434.54, 209.94], [443.85, 207.87], [469.22, 207.86], [461.41, 184.24], [434.54, 184.25], [443.85, 183.61], [469.22, 183.61]]}, {\"category\": \"car\", \"corners_3d\": [[594.11, 193.26], [610.67, 193.24], [611.87, 194.34], [594.41, 194.35], [594.11, 180.2], [610.67, 180.2], [611.87, 180.59], [594.41, 180.6]]}]\n```" - } - ] - }, - { - "id": 1302, - "data_source": "Mono3DRefer", - "scene_name": "000877", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[410.0, 225.32], [369.96, 225.41], [396.48, 219.0], [431.6, 218.93], [410.0, 189.57], [369.96, 189.6], [396.48, 187.56], [431.6, 187.54]]}, {\"category\": \"car\", \"corners_3d\": [[461.41, 209.92], [434.54, 209.94], [443.85, 207.87], [469.22, 207.86], [461.41, 184.24], [434.54, 184.25], [443.85, 183.61], [469.22, 183.61]]}, {\"category\": \"truck\", \"corners_3d\": [[501.1, 202.19], [467.89, 202.24], [479.41, 199.22], [509.2, 199.18], [501.1, 161.94], [467.89, 161.92], [479.41, 163.04], [509.2, 163.06]]}, {\"category\": \"car\", \"corners_3d\": [[594.11, 193.26], [610.67, 193.24], [611.87, 194.34], [594.41, 194.35], [594.11, 180.2], [610.67, 180.2], [611.87, 180.59], [594.41, 180.6]]}]\n```" - } - ] - }, - { - "id": 1303, - "data_source": "Mono3DRefer", - "scene_name": "000877", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[214.43, 268.68], [117.67, 269.07], [223.94, 247.3], [298.54, 247.07], [214.43, 183.66], [117.67, 183.7], [223.94, 181.25], [298.54, 181.22]]}, {\"category\": \"car\", \"corners_3d\": [[410.0, 225.32], [369.96, 225.41], [396.48, 219.0], [431.6, 218.93], [410.0, 189.57], [369.96, 189.6], [396.48, 187.56], [431.6, 187.54]]}, {\"category\": \"car\", \"corners_3d\": [[461.41, 209.92], [434.54, 209.94], [443.85, 207.87], [469.22, 207.86], [461.41, 184.24], [434.54, 184.25], [443.85, 183.61], [469.22, 183.61]]}, {\"category\": \"truck\", \"corners_3d\": [[501.1, 202.19], [467.89, 202.24], [479.41, 199.22], [509.2, 199.18], [501.1, 161.94], [467.89, 161.92], [479.41, 163.04], [509.2, 163.06]]}, {\"category\": \"car\", \"corners_3d\": [[594.11, 193.26], [610.67, 193.24], [611.87, 194.34], [594.41, 194.35], [594.11, 180.2], [610.67, 180.2], [611.87, 180.59], [594.41, 180.6]]}]\n```" - } - ] - }, - { - "id": 1304, - "data_source": "Mono3DRefer", - "scene_name": "000878", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[408.52, 211.74], [376.82, 211.69], [400.44, 208.07], [429.21, 208.11], [408.52, 183.86], [376.82, 183.85], [400.44, 182.82], [429.21, 182.84]]}]\n```" - } - ] - }, - { - "id": 1305, - "data_source": "Mono3DRefer", - "scene_name": "000878", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[64.94, 249.84], [-19.89, 250.2], [129.03, 231.0], [192.35, 230.8], [64.94, 140.32], [-19.89, 140.17], [129.03, 148.28], [192.35, 148.37]]}]\n```" - } - ] - }, - { - "id": 1306, - "data_source": "Mono3DRefer", - "scene_name": "000878", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[64.94, 249.84], [-19.89, 250.2], [129.03, 231.0], [192.35, 230.8], [64.94, 140.32], [-19.89, 140.17], [129.03, 148.28], [192.35, 148.37]]}, {\"category\": \"car\", \"corners_3d\": [[408.52, 211.74], [376.82, 211.69], [400.44, 208.07], [429.21, 208.11], [408.52, 183.86], [376.82, 183.85], [400.44, 182.82], [429.21, 182.84]]}]\n```" - } - ] - }, - { - "id": 1307, - "data_source": "Mono3DRefer", - "scene_name": "000879", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[596.8, 194.14], [626.12, 194.16], [624.79, 199.19], [588.57, 199.15], [596.8, 162.37], [626.12, 162.36], [624.79, 159.89], [588.57, 159.91]]}]\n```" - } - ] - }, - { - "id": 1308, - "data_source": "Mono3DRefer", - "scene_name": "000879", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[400.95, 210.69], [369.54, 210.7], [391.84, 207.17], [420.32, 207.17], [400.95, 186.54], [369.54, 186.54], [391.84, 185.26], [420.32, 185.26]]}, {\"category\": \"car\", \"corners_3d\": [[492.8, 196.78], [474.62, 196.78], [481.06, 195.57], [498.32, 195.57], [492.8, 180.41], [474.62, 180.41], [481.06, 180.03], [498.32, 180.03]]}]\n```" - } - ] - }, - { - "id": 1309, - "data_source": "Mono3DRefer", - "scene_name": "000879", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[596.8, 194.14], [626.12, 194.16], [624.79, 199.19], [588.57, 199.15], [596.8, 162.37], [626.12, 162.36], [624.79, 159.89], [588.57, 159.91]]}, {\"category\": \"car\", \"corners_3d\": [[400.95, 210.69], [369.54, 210.7], [391.84, 207.17], [420.32, 207.17], [400.95, 186.54], [369.54, 186.54], [391.84, 185.26], [420.32, 185.26]]}, {\"category\": \"car\", \"corners_3d\": [[492.8, 196.78], [474.62, 196.78], [481.06, 195.57], [498.32, 195.57], [492.8, 180.41], [474.62, 180.41], [481.06, 180.03], [498.32, 180.03]]}]\n```" - } - ] - }, - { - "id": 1310, - "data_source": "Mono3DRefer", - "scene_name": "000881", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[353.56, 210.53], [316.18, 210.7], [406.5, 195.69], [428.85, 195.62], [353.56, 162.22], [316.18, 162.17], [406.5, 166.41], [428.85, 166.43]]}]\n```" - } - ] - }, - { - "id": 1311, - "data_source": "Mono3DRefer", - "scene_name": "000881", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[579.77, 206.66], [604.64, 206.63], [605.37, 209.38], [578.48, 209.4], [579.77, 182.08], [604.64, 182.08], [605.37, 182.83], [578.48, 182.83]]}]\n```" - } - ] - }, - { - "id": 1312, - "data_source": "Mono3DRefer", - "scene_name": "000881", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[579.77, 206.66], [604.64, 206.63], [605.37, 209.38], [578.48, 209.4], [579.77, 182.08], [604.64, 182.08], [605.37, 182.83], [578.48, 182.83]]}, {\"category\": \"tram\", \"corners_3d\": [[353.56, 210.53], [316.18, 210.7], [406.5, 195.69], [428.85, 195.62], [353.56, 162.22], [316.18, 162.17], [406.5, 166.41], [428.85, 166.43]]}]\n```" - } - ] - }, - { - "id": 1313, - "data_source": "Mono3DRefer", - "scene_name": "000884", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[853.75, 384.0], [1062.36, 382.82], [2401.23, 985.22], [1610.48, 1003.18], [853.75, 162.63], [1062.36, 162.69], [2401.23, 133.52], [1610.48, 132.65]]}, {\"category\": \"van\", \"corners_3d\": [[693.42, 239.56], [755.48, 239.39], [804.27, 258.83], [724.17, 259.11], [693.42, 168.58], [755.48, 168.59], [804.27, 167.35], [724.17, 167.33]]}]\n```" - } - ] - }, - { - "id": 1314, - "data_source": "Mono3DRefer", - "scene_name": "000884", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[531.4, 255.16], [454.97, 255.17], [488.54, 237.23], [548.31, 237.23], [531.4, 184.66], [454.97, 184.66], [488.54, 182.09], [548.31, 182.09]]}, {\"category\": \"car\", \"corners_3d\": [[670.06, 224.12], [709.7, 224.04], [731.18, 233.11], [684.54, 233.22], [670.06, 182.48], [709.7, 182.46], [731.18, 184.16], [684.54, 184.18]]}]\n```" - } - ] - }, - { - "id": 1315, - "data_source": "Mono3DRefer", - "scene_name": "000884", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[853.75, 384.0], [1062.36, 382.82], [2401.23, 985.22], [1610.48, 1003.18], [853.75, 162.63], [1062.36, 162.69], [2401.23, 133.52], [1610.48, 132.65]]}, {\"category\": \"car\", \"corners_3d\": [[531.4, 255.16], [454.97, 255.17], [488.54, 237.23], [548.31, 237.23], [531.4, 184.66], [454.97, 184.66], [488.54, 182.09], [548.31, 182.09]]}, {\"category\": \"van\", \"corners_3d\": [[693.42, 239.56], [755.48, 239.39], [804.27, 258.83], [724.17, 259.11], [693.42, 168.58], [755.48, 168.59], [804.27, 167.35], [724.17, 167.33]]}, {\"category\": \"car\", \"corners_3d\": [[670.06, 224.12], [709.7, 224.04], [731.18, 233.11], [684.54, 233.22], [670.06, 182.48], [709.7, 182.46], [731.18, 184.16], [684.54, 184.18]]}]\n```" - } - ] - }, - { - "id": 1316, - "data_source": "Mono3DRefer", - "scene_name": "000885", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[511.55, 262.06], [430.84, 262.17], [471.73, 240.8], [533.12, 240.74], [511.55, 184.12], [430.84, 184.13], [471.73, 181.43], [533.12, 181.42]]}, {\"category\": \"car\", \"corners_3d\": [[844.92, 294.4], [733.89, 294.99], [693.1, 260.99], [773.48, 260.69], [844.92, 187.67], [733.89, 187.74], [693.1, 183.6], [773.48, 183.56]]}, {\"category\": \"car\", \"corners_3d\": [[775.56, 250.98], [709.85, 251.2], [688.62, 237.62], [743.01, 237.47], [775.56, 181.37], [709.85, 181.39], [688.62, 179.91], [743.01, 179.9]]}, {\"category\": \"car\", \"corners_3d\": [[564.0, 218.24], [525.19, 218.37], [529.97, 213.8], [564.89, 213.7], [564.0, 182.99], [525.19, 183.01], [529.97, 181.99], [564.89, 181.97]]}, {\"category\": \"car\", \"corners_3d\": [[647.19, 212.58], [683.51, 212.59], [692.1, 217.24], [651.53, 217.24], [647.19, 177.55], [683.51, 177.55], [692.1, 178.1], [651.53, 178.1]]}]\n```" - } - ] - }, - { - "id": 1317, - "data_source": "Mono3DRefer", - "scene_name": "000887", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[524.02, 205.22], [573.82, 205.04], [581.21, 216.1], [514.23, 216.42], [524.02, 144.61], [573.82, 144.76], [581.21, 135.11], [514.23, 134.83]]}]\n```" - } - ] - }, - { - "id": 1318, - "data_source": "Mono3DRefer", - "scene_name": "000887", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[38.78, 230.18], [2.8, 230.5], [51.01, 225.14], [83.43, 224.88], [38.78, 198.4], [2.8, 198.54], [51.01, 196.16], [83.43, 196.04]]}, {\"category\": \"car\", \"corners_3d\": [[113.02, 222.35], [83.44, 222.29], [137.35, 217.42], [164.06, 217.47], [113.02, 191.74], [83.44, 191.72], [137.35, 189.86], [164.06, 189.88]]}, {\"category\": \"car\", \"corners_3d\": [[171.06, 216.23], [149.09, 216.25], [184.95, 212.76], [205.13, 212.74], [171.06, 192.45], [149.09, 192.46], [184.95, 190.88], [205.13, 190.87]]}, {\"category\": \"car\", \"corners_3d\": [[209.06, 210.22], [184.14, 210.33], [211.64, 207.47], [234.58, 207.37], [209.06, 188.24], [184.14, 188.28], [211.64, 187.1], [234.58, 187.06]]}, {\"category\": \"car\", \"corners_3d\": [[245.06, 211.79], [223.62, 211.9], [243.76, 209.39], [263.78, 209.29], [245.06, 190.3], [223.62, 190.35], [243.76, 189.22], [263.78, 189.18]]}]\n```" - } - ] - }, - { - "id": 1319, - "data_source": "Mono3DRefer", - "scene_name": "000887", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[524.02, 205.22], [573.82, 205.04], [581.21, 216.1], [514.23, 216.42], [524.02, 144.61], [573.82, 144.76], [581.21, 135.11], [514.23, 134.83]]}, {\"category\": \"car\", \"corners_3d\": [[38.78, 230.18], [2.8, 230.5], [51.01, 225.14], [83.43, 224.88], [38.78, 198.4], [2.8, 198.54], [51.01, 196.16], [83.43, 196.04]]}, {\"category\": \"car\", \"corners_3d\": [[113.02, 222.35], [83.44, 222.29], [137.35, 217.42], [164.06, 217.47], [113.02, 191.74], [83.44, 191.72], [137.35, 189.86], [164.06, 189.88]]}, {\"category\": \"car\", \"corners_3d\": [[171.06, 216.23], [149.09, 216.25], [184.95, 212.76], [205.13, 212.74], [171.06, 192.45], [149.09, 192.46], [184.95, 190.88], [205.13, 190.87]]}, {\"category\": \"car\", \"corners_3d\": [[209.06, 210.22], [184.14, 210.33], [211.64, 207.47], [234.58, 207.37], [209.06, 188.24], [184.14, 188.28], [211.64, 187.1], [234.58, 187.06]]}, {\"category\": \"car\", \"corners_3d\": [[245.06, 211.79], [223.62, 211.9], [243.76, 209.39], [263.78, 209.29], [245.06, 190.3], [223.62, 190.35], [243.76, 189.22], [263.78, 189.18]]}]\n```" - } - ] - }, - { - "id": 1320, - "data_source": "Mono3DRefer", - "scene_name": "000890", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[730.94, 258.26], [807.66, 257.9], [878.56, 284.54], [778.08, 285.15], [730.94, 186.38], [807.66, 186.32], [878.56, 190.54], [778.08, 190.64]]}, {\"category\": \"car\", \"corners_3d\": [[736.05, 205.61], [765.54, 205.53], [785.37, 208.78], [752.98, 208.88], [736.05, 176.69], [765.54, 176.68], [785.37, 177.06], [752.98, 177.07]]}, {\"category\": \"car\", \"corners_3d\": [[814.23, 223.81], [861.8, 223.54], [925.73, 234.0], [868.69, 234.4], [814.23, 178.51], [861.8, 178.48], [925.73, 179.64], [868.69, 179.68]]}, {\"category\": \"car\", \"corners_3d\": [[651.06, 203.64], [677.8, 203.59], [687.07, 206.33], [657.95, 206.4], [651.06, 177.3], [677.8, 177.29], [687.07, 177.69], [657.95, 177.7]]}]\n```" - } - ] - }, - { - "id": 1321, - "data_source": "Mono3DRefer", - "scene_name": "000893", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[176.69, 248.71], [113.43, 248.72], [193.2, 236.5], [246.28, 236.5], [176.69, 196.47], [113.43, 196.47], [193.2, 192.67], [246.28, 192.67]]}]\n```" - } - ] - }, - { - "id": 1322, - "data_source": "Mono3DRefer", - "scene_name": "000893", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[639.71, 202.1], [682.8, 202.15], [702.62, 213.53], [642.78, 213.43], [639.71, 155.53], [682.8, 155.5], [702.62, 148.75], [642.78, 148.81]]}]\n```" - } - ] - }, - { - "id": 1323, - "data_source": "Mono3DRefer", - "scene_name": "000893", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[639.71, 202.1], [682.8, 202.15], [702.62, 213.53], [642.78, 213.43], [639.71, 155.53], [682.8, 155.5], [702.62, 148.75], [642.78, 148.81]]}, {\"category\": \"car\", \"corners_3d\": [[176.69, 248.71], [113.43, 248.72], [193.2, 236.5], [246.28, 236.5], [176.69, 196.47], [113.43, 196.47], [193.2, 192.67], [246.28, 192.67]]}]\n```" - } - ] - }, - { - "id": 1324, - "data_source": "Mono3DRefer", - "scene_name": "000895", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[547.8, 230.16], [494.54, 230.58], [501.71, 220.73], [545.89, 220.44], [547.8, 183.44], [494.54, 183.52], [501.71, 181.7], [545.89, 181.65]]}]\n```" - } - ] - }, - { - "id": 1325, - "data_source": "Mono3DRefer", - "scene_name": "000901", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[626.73, 225.04], [590.17, 225.05], [592.51, 218.54], [624.52, 218.54], [626.73, 192.62], [590.17, 192.62], [592.51, 190.16], [624.52, 190.16]]}]\n```" - } - ] - }, - { - "id": 1326, - "data_source": "Mono3DRefer", - "scene_name": "000905", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[453.78, 233.31], [412.08, 234.12], [414.03, 228.24], [451.75, 227.57], [453.78, 195.38], [412.08, 195.68], [414.03, 193.49], [451.75, 193.24]]}]\n```" - } - ] - }, - { - "id": 1327, - "data_source": "Mono3DRefer", - "scene_name": "000908", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[599.53, 187.07], [584.15, 186.99], [594.7, 186.23], [609.28, 186.29], [599.53, 172.55], [584.15, 172.55], [594.7, 172.57], [609.28, 172.57]]}, {\"category\": \"car\", \"corners_3d\": [[721.41, 186.26], [738.59, 186.32], [737.43, 187.2], [719.14, 187.13], [721.41, 171.04], [738.59, 171.03], [737.43, 170.91], [719.14, 170.92]]}]\n```" - } - ] - }, - { - "id": 1328, - "data_source": "Mono3DRefer", - "scene_name": "000920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[358.77, 332.53], [205.16, 334.56], [330.83, 278.18], [430.29, 277.32], [358.77, 200.95], [205.16, 201.31], [330.83, 191.39], [430.29, 191.24]]}, {\"category\": \"car\", \"corners_3d\": [[725.77, 244.11], [665.3, 244.48], [648.11, 232.13], [698.27, 231.87], [725.77, 181.7], [665.3, 181.75], [648.11, 180.21], [698.27, 180.18]]}, {\"category\": \"car\", \"corners_3d\": [[517.98, 228.0], [469.66, 228.04], [489.57, 219.72], [530.6, 219.69], [517.98, 186.13], [469.66, 186.14], [489.57, 184.14], [530.6, 184.13]]}, {\"category\": \"car\", \"corners_3d\": [[630.82, 219.59], [669.62, 219.34], [688.08, 225.84], [643.93, 226.17], [630.82, 183.23], [669.62, 183.18], [688.08, 184.62], [643.93, 184.69]]}]\n```" - } - ] - }, - { - "id": 1329, - "data_source": "Mono3DRefer", - "scene_name": "000921", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[564.24, 207.51], [540.72, 207.49], [546.95, 204.88], [568.7, 204.9], [564.24, 181.32], [540.72, 181.32], [546.95, 180.68], [568.7, 180.68]]}, {\"category\": \"car\", \"corners_3d\": [[530.62, 239.91], [470.09, 240.03], [491.07, 228.7], [541.38, 228.62], [530.62, 176.39], [470.09, 176.4], [491.07, 175.8], [541.38, 175.79]]}]\n```" - } - ] - }, - { - "id": 1330, - "data_source": "Mono3DRefer", - "scene_name": "000930", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[306.73, 405.58], [75.23, 405.64], [312.92, 301.98], [441.31, 301.96], [306.73, 164.29], [75.23, 164.28], [312.92, 168.1], [441.31, 168.1]]}, {\"category\": \"car\", \"corners_3d\": [[452.53, 245.86], [523.54, 246.08], [492.88, 266.82], [401.95, 266.45], [452.53, 186.08], [523.54, 186.12], [492.88, 189.87], [401.95, 189.81]]}, {\"category\": \"car\", \"corners_3d\": [[682.02, 216.51], [643.5, 216.6], [636.48, 211.76], [670.76, 211.68], [682.02, 180.76], [643.5, 180.78], [636.48, 179.9], [670.76, 179.89]]}]\n```" - } - ] - }, - { - "id": 1331, - "data_source": "Mono3DRefer", - "scene_name": "000930", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[575.14, 198.76], [547.89, 198.71], [555.38, 196.72], [580.54, 196.76], [575.14, 171.21], [547.89, 171.21], [555.38, 171.34], [580.54, 171.34]]}]\n```" - } - ] - }, - { - "id": 1332, - "data_source": "Mono3DRefer", - "scene_name": "000930", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[306.73, 405.58], [75.23, 405.64], [312.92, 301.98], [441.31, 301.96], [306.73, 164.29], [75.23, 164.28], [312.92, 168.1], [441.31, 168.1]]}, {\"category\": \"car\", \"corners_3d\": [[452.53, 245.86], [523.54, 246.08], [492.88, 266.82], [401.95, 266.45], [452.53, 186.08], [523.54, 186.12], [492.88, 189.87], [401.95, 189.81]]}, {\"category\": \"car\", \"corners_3d\": [[682.02, 216.51], [643.5, 216.6], [636.48, 211.76], [670.76, 211.68], [682.02, 180.76], [643.5, 180.78], [636.48, 179.9], [670.76, 179.89]]}, {\"category\": \"van\", \"corners_3d\": [[575.14, 198.76], [547.89, 198.71], [555.38, 196.72], [580.54, 196.76], [575.14, 171.21], [547.89, 171.21], [555.38, 171.34], [580.54, 171.34]]}]\n```" - } - ] - }, - { - "id": 1333, - "data_source": "Mono3DRefer", - "scene_name": "000931", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[997.71, 237.76], [1020.14, 237.76], [1044.61, 241.64], [1020.84, 241.63], [997.71, 158.36], [1020.14, 158.36], [1044.61, 157.49], [1020.84, 157.49]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[957.87, 238.9], [987.74, 238.55], [1008.77, 241.53], [977.66, 241.91], [957.87, 159.81], [987.74, 159.88], [1008.77, 159.29], [977.66, 159.22]]}]\n```" - } - ] - }, - { - "id": 1334, - "data_source": "Mono3DRefer", - "scene_name": "000931", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[779.4, 207.49], [790.2, 207.54], [796.08, 209.44], [784.69, 209.38], [779.4, 168.03], [790.2, 168.03], [796.08, 167.76], [784.69, 167.77]]}, {\"category\": \"cyclist\", \"corners_3d\": [[922.7, 249.1], [952.03, 249.4], [992.67, 260.95], [958.75, 260.55], [922.7, 151.74], [952.03, 151.66], [992.67, 148.46], [958.75, 148.57]]}]\n```" - } - ] - }, - { - "id": 1335, - "data_source": "Mono3DRefer", - "scene_name": "000931", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[997.71, 237.76], [1020.14, 237.76], [1044.61, 241.64], [1020.84, 241.63], [997.71, 158.36], [1020.14, 158.36], [1044.61, 157.49], [1020.84, 157.49]]}, {\"category\": \"cyclist\", \"corners_3d\": [[779.4, 207.49], [790.2, 207.54], [796.08, 209.44], [784.69, 209.38], [779.4, 168.03], [790.2, 168.03], [796.08, 167.76], [784.69, 167.77]]}, {\"category\": \"cyclist\", \"corners_3d\": [[922.7, 249.1], [952.03, 249.4], [992.67, 260.95], [958.75, 260.55], [922.7, 151.74], [952.03, 151.66], [992.67, 148.46], [958.75, 148.57]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[957.87, 238.9], [987.74, 238.55], [1008.77, 241.53], [977.66, 241.91], [957.87, 159.81], [987.74, 159.88], [1008.77, 159.29], [977.66, 159.22]]}]\n```" - } - ] - }, - { - "id": 1336, - "data_source": "Mono3DRefer", - "scene_name": "000933", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[531.02, 227.65], [481.14, 227.65], [499.38, 219.83], [542.14, 219.83], [531.02, 185.89], [481.14, 185.89], [499.38, 184.03], [542.14, 184.03]]}]\n```" - } - ] - }, - { - "id": 1337, - "data_source": "Mono3DRefer", - "scene_name": "000934", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[629.07, 208.62], [611.78, 208.61], [612.06, 206.36], [628.26, 206.37], [629.07, 193.29], [611.78, 193.29], [612.06, 192.0], [628.26, 192.0]]}]\n```" - } - ] - }, - { - "id": 1338, - "data_source": "Mono3DRefer", - "scene_name": "000937", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[770.14, 215.0], [749.86, 215.0], [744.85, 213.5], [764.41, 213.5], [770.14, 166.79], [749.86, 166.79], [744.85, 167.0], [764.41, 167.0]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1116.61, 264.05], [1157.21, 271.15], [1122.93, 271.24], [1084.78, 264.13], [1116.61, 157.32], [1157.21, 156.11], [1122.93, 156.09], [1084.78, 157.3]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1092.94, 258.97], [1120.1, 263.88], [1084.94, 263.84], [1059.69, 258.93], [1092.94, 163.98], [1120.1, 163.48], [1084.94, 163.48], [1059.69, 163.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[563.88, 223.92], [583.56, 223.85], [583.91, 225.72], [563.51, 225.79], [563.88, 177.54], [583.56, 177.53], [583.91, 177.71], [563.51, 177.71]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[556.83, 233.95], [531.36, 233.95], [533.97, 231.89], [558.58, 231.89], [556.83, 172.11], [531.36, 172.11], [533.97, 172.14], [558.58, 172.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[748.84, 217.14], [731.85, 217.09], [728.56, 215.41], [744.91, 215.46], [748.84, 166.52], [731.85, 166.52], [728.56, 166.76], [744.91, 166.76]]}]\n```" - } - ] - }, - { - "id": 1339, - "data_source": "Mono3DRefer", - "scene_name": "000941", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[280.27, 334.27], [425.85, 334.96], [189.18, 528.03], [-126.03, 524.76], [280.27, 186.21], [425.85, 186.27], [189.18, 202.24], [-126.03, 201.97]]}, {\"category\": \"car\", \"corners_3d\": [[513.36, 261.5], [433.67, 261.31], [472.79, 243.07], [536.09, 243.19], [513.36, 185.23], [433.67, 185.2], [472.79, 182.66], [536.09, 182.67]]}, {\"category\": \"car\", \"corners_3d\": [[586.15, 207.25], [558.97, 207.24], [563.85, 204.31], [588.71, 204.32], [586.15, 183.99], [558.97, 183.99], [563.85, 183.04], [588.71, 183.04]]}]\n```" - } - ] - }, - { - "id": 1340, - "data_source": "Mono3DRefer", - "scene_name": "000945", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[480.0, 302.67], [365.36, 302.27], [435.28, 267.2], [518.95, 267.41], [480.0, 185.09], [365.36, 185.05], [435.28, 181.75], [518.95, 181.77]]}, {\"category\": \"car\", \"corners_3d\": [[879.17, 408.49], [1088.88, 415.17], [1720.82, 792.02], [1164.35, 750.23], [879.17, 217.57], [1088.88, 218.84], [1720.82, 290.36], [1164.35, 282.43]]}, {\"category\": \"car\", \"corners_3d\": [[722.77, 284.41], [814.33, 283.99], [904.68, 328.45], [776.78, 329.27], [722.77, 185.85], [814.33, 185.8], [904.68, 190.98], [776.78, 191.07]]}, {\"category\": \"car\", \"corners_3d\": [[671.12, 254.28], [738.26, 254.06], [776.93, 275.03], [692.53, 275.38], [671.12, 189.3], [738.26, 189.25], [776.93, 193.49], [692.53, 193.56]]}, {\"category\": \"car\", \"corners_3d\": [[477.76, 245.47], [534.85, 245.41], [519.58, 261.69], [449.66, 261.77], [477.76, 186.91], [534.85, 186.9], [519.58, 190.05], [449.66, 190.07]]}, {\"category\": \"car\", \"corners_3d\": [[536.61, 239.15], [479.82, 239.1], [498.67, 229.94], [547.62, 229.98], [536.61, 188.89], [479.82, 188.88], [498.67, 186.67], [547.62, 186.68]]}, {\"category\": \"car\", \"corners_3d\": [[654.99, 231.6], [701.45, 231.26], [726.2, 240.5], [672.5, 240.95], [654.99, 184.41], [701.45, 184.34], [726.2, 186.16], [672.5, 186.25]]}, {\"category\": \"car\", \"corners_3d\": [[636.96, 223.59], [676.78, 223.48], [691.17, 231.11], [645.37, 231.25], [636.96, 186.59], [676.78, 186.56], [691.17, 188.63], [645.37, 188.67]]}, {\"category\": \"car\", \"corners_3d\": [[556.31, 218.22], [525.04, 218.16], [534.81, 213.97], [563.2, 214.02], [556.31, 186.62], [525.04, 186.6], [534.81, 185.33], [563.2, 185.34]]}, {\"category\": \"car\", \"corners_3d\": [[576.98, 211.07], [547.77, 211.05], [554.16, 207.49], [580.64, 207.5], [576.98, 177.43], [547.77, 177.43], [554.16, 177.0], [580.64, 177.0]]}, {\"category\": \"car\", \"corners_3d\": [[586.74, 208.28], [560.83, 208.21], [567.86, 205.26], [591.62, 205.31], [586.74, 180.09], [560.83, 180.08], [567.86, 179.47], [591.62, 179.48]]}]\n```" - } - ] - }, - { - "id": 1341, - "data_source": "Mono3DRefer", - "scene_name": "000950", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[568.53, 214.63], [541.6, 214.8], [540.6, 211.38], [565.34, 211.24], [568.53, 186.99], [541.6, 187.05], [540.6, 185.9], [565.34, 185.85]]}]\n```" - } - ] - }, - { - "id": 1342, - "data_source": "Mono3DRefer", - "scene_name": "000951", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[251.92, 240.5], [316.62, 240.04], [244.38, 260.63], [159.01, 261.42], [251.92, 186.08], [316.62, 185.99], [244.38, 190.01], [159.01, 190.17]]}, {\"category\": \"car\", \"corners_3d\": [[138.75, 243.42], [200.75, 243.25], [81.84, 264.84], [0.37, 265.14], [138.75, 168.43], [200.75, 168.44], [81.84, 167.09], [0.37, 167.07]]}, {\"category\": \"car\", \"corners_3d\": [[424.67, 211.64], [470.76, 211.4], [458.94, 218.2], [404.57, 218.54], [424.67, 172.29], [470.76, 172.29], [458.94, 172.19], [404.57, 172.19]]}, {\"category\": \"car\", \"corners_3d\": [[729.88, 224.2], [786.29, 223.79], [851.57, 237.17], [780.85, 237.82], [729.88, 173.18], [786.29, 173.18], [851.57, 173.26], [780.85, 173.27]]}, {\"category\": \"car\", \"corners_3d\": [[573.87, 196.04], [608.06, 196.0], [611.65, 199.1], [572.87, 199.16], [573.87, 163.72], [608.06, 163.74], [611.65, 162.52], [572.87, 162.5]]}, {\"category\": \"car\", \"corners_3d\": [[806.67, 219.45], [862.18, 219.23], [936.14, 230.85], [867.08, 231.19], [806.67, 165.62], [862.18, 165.65], [936.14, 163.85], [867.08, 163.8]]}, {\"category\": \"car\", \"corners_3d\": [[892.87, 218.6], [950.04, 218.34], [1049.62, 229.94], [978.47, 230.34], [892.87, 173.17], [950.04, 173.17], [1049.62, 173.25], [978.47, 173.25]]}]\n```" - } - ] - }, - { - "id": 1343, - "data_source": "Mono3DRefer", - "scene_name": "000951", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1287.69, 242.09], [1415.25, 253.97], [1050.51, 258.09], [973.19, 245.07], [1287.69, 109.56], [1415.25, 98.7], [1050.51, 94.94], [973.19, 106.84]]}, {\"category\": \"van\", \"corners_3d\": [[663.96, 194.01], [630.46, 194.11], [619.91, 191.52], [649.39, 191.45], [663.96, 148.54], [630.46, 148.44], [619.91, 151.41], [649.39, 151.49]]}, {\"category\": \"van\", \"corners_3d\": [[507.44, 233.4], [572.41, 233.08], [574.04, 250.59], [490.06, 251.13], [507.44, 162.03], [572.41, 162.08], [574.04, 158.95], [490.06, 158.86]]}]\n```" - } - ] - }, - { - "id": 1344, - "data_source": "Mono3DRefer", - "scene_name": "000951", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[607.79, 197.02], [566.84, 197.13], [564.21, 193.07], [598.35, 192.99], [607.79, 132.14], [566.84, 131.95], [564.21, 138.8], [598.35, 138.93]]}]\n```" - } - ] - }, - { - "id": 1345, - "data_source": "Mono3DRefer", - "scene_name": "000951", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1287.69, 242.09], [1415.25, 253.97], [1050.51, 258.09], [973.19, 245.07], [1287.69, 109.56], [1415.25, 98.7], [1050.51, 94.94], [973.19, 106.84]]}, {\"category\": \"car\", \"corners_3d\": [[251.92, 240.5], [316.62, 240.04], [244.38, 260.63], [159.01, 261.42], [251.92, 186.08], [316.62, 185.99], [244.38, 190.01], [159.01, 190.17]]}, {\"category\": \"car\", \"corners_3d\": [[138.75, 243.42], [200.75, 243.25], [81.84, 264.84], [0.37, 265.14], [138.75, 168.43], [200.75, 168.44], [81.84, 167.09], [0.37, 167.07]]}, {\"category\": \"car\", \"corners_3d\": [[424.67, 211.64], [470.76, 211.4], [458.94, 218.2], [404.57, 218.54], [424.67, 172.29], [470.76, 172.29], [458.94, 172.19], [404.57, 172.19]]}, {\"category\": \"car\", \"corners_3d\": [[729.88, 224.2], [786.29, 223.79], [851.57, 237.17], [780.85, 237.82], [729.88, 173.18], [786.29, 173.18], [851.57, 173.26], [780.85, 173.27]]}, {\"category\": \"car\", \"corners_3d\": [[573.87, 196.04], [608.06, 196.0], [611.65, 199.1], [572.87, 199.16], [573.87, 163.72], [608.06, 163.74], [611.65, 162.52], [572.87, 162.5]]}, {\"category\": \"van\", \"corners_3d\": [[663.96, 194.01], [630.46, 194.11], [619.91, 191.52], [649.39, 191.45], [663.96, 148.54], [630.46, 148.44], [619.91, 151.41], [649.39, 151.49]]}, {\"category\": \"car\", \"corners_3d\": [[806.67, 219.45], [862.18, 219.23], [936.14, 230.85], [867.08, 231.19], [806.67, 165.62], [862.18, 165.65], [936.14, 163.85], [867.08, 163.8]]}, {\"category\": \"car\", \"corners_3d\": [[892.87, 218.6], [950.04, 218.34], [1049.62, 229.94], [978.47, 230.34], [892.87, 173.17], [950.04, 173.17], [1049.62, 173.25], [978.47, 173.25]]}, {\"category\": \"van\", \"corners_3d\": [[507.44, 233.4], [572.41, 233.08], [574.04, 250.59], [490.06, 251.13], [507.44, 162.03], [572.41, 162.08], [574.04, 158.95], [490.06, 158.86]]}]\n```" - } - ] - }, - { - "id": 1346, - "data_source": "Mono3DRefer", - "scene_name": "000951", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[251.92, 240.5], [316.62, 240.04], [244.38, 260.63], [159.01, 261.42], [251.92, 186.08], [316.62, 185.99], [244.38, 190.01], [159.01, 190.17]]}, {\"category\": \"car\", \"corners_3d\": [[138.75, 243.42], [200.75, 243.25], [81.84, 264.84], [0.37, 265.14], [138.75, 168.43], [200.75, 168.44], [81.84, 167.09], [0.37, 167.07]]}, {\"category\": \"car\", \"corners_3d\": [[424.67, 211.64], [470.76, 211.4], [458.94, 218.2], [404.57, 218.54], [424.67, 172.29], [470.76, 172.29], [458.94, 172.19], [404.57, 172.19]]}, {\"category\": \"car\", \"corners_3d\": [[729.88, 224.2], [786.29, 223.79], [851.57, 237.17], [780.85, 237.82], [729.88, 173.18], [786.29, 173.18], [851.57, 173.26], [780.85, 173.27]]}, {\"category\": \"car\", \"corners_3d\": [[573.87, 196.04], [608.06, 196.0], [611.65, 199.1], [572.87, 199.16], [573.87, 163.72], [608.06, 163.74], [611.65, 162.52], [572.87, 162.5]]}, {\"category\": \"car\", \"corners_3d\": [[806.67, 219.45], [862.18, 219.23], [936.14, 230.85], [867.08, 231.19], [806.67, 165.62], [862.18, 165.65], [936.14, 163.85], [867.08, 163.8]]}, {\"category\": \"car\", \"corners_3d\": [[892.87, 218.6], [950.04, 218.34], [1049.62, 229.94], [978.47, 230.34], [892.87, 173.17], [950.04, 173.17], [1049.62, 173.25], [978.47, 173.25]]}, {\"category\": \"truck\", \"corners_3d\": [[607.79, 197.02], [566.84, 197.13], [564.21, 193.07], [598.35, 192.99], [607.79, 132.14], [566.84, 131.95], [564.21, 138.8], [598.35, 138.93]]}]\n```" - } - ] - }, - { - "id": 1347, - "data_source": "Mono3DRefer", - "scene_name": "000951", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1287.69, 242.09], [1415.25, 253.97], [1050.51, 258.09], [973.19, 245.07], [1287.69, 109.56], [1415.25, 98.7], [1050.51, 94.94], [973.19, 106.84]]}, {\"category\": \"van\", \"corners_3d\": [[663.96, 194.01], [630.46, 194.11], [619.91, 191.52], [649.39, 191.45], [663.96, 148.54], [630.46, 148.44], [619.91, 151.41], [649.39, 151.49]]}, {\"category\": \"truck\", \"corners_3d\": [[607.79, 197.02], [566.84, 197.13], [564.21, 193.07], [598.35, 192.99], [607.79, 132.14], [566.84, 131.95], [564.21, 138.8], [598.35, 138.93]]}, {\"category\": \"van\", \"corners_3d\": [[507.44, 233.4], [572.41, 233.08], [574.04, 250.59], [490.06, 251.13], [507.44, 162.03], [572.41, 162.08], [574.04, 158.95], [490.06, 158.86]]}]\n```" - } - ] - }, - { - "id": 1348, - "data_source": "Mono3DRefer", - "scene_name": "000951", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1287.69, 242.09], [1415.25, 253.97], [1050.51, 258.09], [973.19, 245.07], [1287.69, 109.56], [1415.25, 98.7], [1050.51, 94.94], [973.19, 106.84]]}, {\"category\": \"car\", \"corners_3d\": [[251.92, 240.5], [316.62, 240.04], [244.38, 260.63], [159.01, 261.42], [251.92, 186.08], [316.62, 185.99], [244.38, 190.01], [159.01, 190.17]]}, {\"category\": \"car\", \"corners_3d\": [[138.75, 243.42], [200.75, 243.25], [81.84, 264.84], [0.37, 265.14], [138.75, 168.43], [200.75, 168.44], [81.84, 167.09], [0.37, 167.07]]}, {\"category\": \"car\", \"corners_3d\": [[424.67, 211.64], [470.76, 211.4], [458.94, 218.2], [404.57, 218.54], [424.67, 172.29], [470.76, 172.29], [458.94, 172.19], [404.57, 172.19]]}, {\"category\": \"car\", \"corners_3d\": [[729.88, 224.2], [786.29, 223.79], [851.57, 237.17], [780.85, 237.82], [729.88, 173.18], [786.29, 173.18], [851.57, 173.26], [780.85, 173.27]]}, {\"category\": \"car\", \"corners_3d\": [[573.87, 196.04], [608.06, 196.0], [611.65, 199.1], [572.87, 199.16], [573.87, 163.72], [608.06, 163.74], [611.65, 162.52], [572.87, 162.5]]}, {\"category\": \"van\", \"corners_3d\": [[663.96, 194.01], [630.46, 194.11], [619.91, 191.52], [649.39, 191.45], [663.96, 148.54], [630.46, 148.44], [619.91, 151.41], [649.39, 151.49]]}, {\"category\": \"car\", \"corners_3d\": [[806.67, 219.45], [862.18, 219.23], [936.14, 230.85], [867.08, 231.19], [806.67, 165.62], [862.18, 165.65], [936.14, 163.85], [867.08, 163.8]]}, {\"category\": \"car\", \"corners_3d\": [[892.87, 218.6], [950.04, 218.34], [1049.62, 229.94], [978.47, 230.34], [892.87, 173.17], [950.04, 173.17], [1049.62, 173.25], [978.47, 173.25]]}, {\"category\": \"truck\", \"corners_3d\": [[607.79, 197.02], [566.84, 197.13], [564.21, 193.07], [598.35, 192.99], [607.79, 132.14], [566.84, 131.95], [564.21, 138.8], [598.35, 138.93]]}, {\"category\": \"van\", \"corners_3d\": [[507.44, 233.4], [572.41, 233.08], [574.04, 250.59], [490.06, 251.13], [507.44, 162.03], [572.41, 162.08], [574.04, 158.95], [490.06, 158.86]]}]\n```" - } - ] - }, - { - "id": 1349, - "data_source": "Mono3DRefer", - "scene_name": "000952", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1119.85, 244.26], [1209.43, 250.23], [1151.87, 263.74], [1053.21, 255.6], [1119.85, 179.45], [1209.43, 180.0], [1151.87, 181.25], [1053.21, 180.5]]}, {\"category\": \"car\", \"corners_3d\": [[778.84, 295.2], [893.1, 313.47], [692.33, 361.27], [579.21, 329.84], [778.84, 171.28], [893.1, 171.04], [692.33, 170.43], [579.21, 170.83]]}, {\"category\": \"car\", \"corners_3d\": [[951.68, 300.48], [840.94, 285.25], [967.28, 263.66], [1067.98, 273.36], [951.68, 168.62], [840.94, 169.13], [967.28, 169.84], [1067.98, 169.52]]}]\n```" - } - ] - }, - { - "id": 1350, - "data_source": "Mono3DRefer", - "scene_name": "000954", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[763.47, 272.05], [872.28, 271.76], [988.39, 313.29], [834.19, 313.87], [763.47, 165.7], [872.28, 165.72], [988.39, 162.72], [834.19, 162.68]]}]\n```" - } - ] - }, - { - "id": 1351, - "data_source": "Mono3DRefer", - "scene_name": "000955", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1212.89, 256.27], [1248.97, 256.59], [1367.87, 273.73], [1323.9, 273.27], [1212.89, 135.74], [1248.97, 135.6], [1367.87, 127.98], [1323.9, 128.18]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1032.78, 232.82], [1054.86, 232.93], [1107.94, 240.88], [1082.84, 240.74], [1032.78, 152.51], [1054.86, 152.47], [1107.94, 149.78], [1082.84, 149.83]]}, {\"category\": \"cyclist\", \"corners_3d\": [[998.2, 223.26], [1016.85, 223.92], [1008.76, 228.14], [988.71, 227.37], [998.2, 163.2], [1016.85, 163.07], [1008.76, 162.27], [988.71, 162.41]]}]\n```" - } - ] - }, - { - "id": 1352, - "data_source": "Mono3DRefer", - "scene_name": "000955", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[390.28, 235.15], [373.81, 240.34], [236.03, 240.72], [263.24, 235.48], [390.28, 183.68], [373.81, 184.58], [236.03, 184.65], [263.24, 183.73]]}, {\"category\": \"car\", \"corners_3d\": [[307.89, 228.29], [330.32, 224.91], [428.18, 225.58], [411.88, 229.05], [307.89, 184.75], [330.32, 184.02], [428.18, 184.17], [411.88, 184.91]]}, {\"category\": \"car\", \"corners_3d\": [[921.52, 185.75], [894.32, 185.64], [886.07, 184.6], [910.96, 184.69], [921.52, 163.1], [894.32, 163.18], [886.07, 163.97], [910.96, 163.9]]}]\n```" - } - ] - }, - { - "id": 1353, - "data_source": "Mono3DRefer", - "scene_name": "000955", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[390.28, 235.15], [373.81, 240.34], [236.03, 240.72], [263.24, 235.48], [390.28, 183.68], [373.81, 184.58], [236.03, 184.65], [263.24, 183.73]]}, {\"category\": \"car\", \"corners_3d\": [[307.89, 228.29], [330.32, 224.91], [428.18, 225.58], [411.88, 229.05], [307.89, 184.75], [330.32, 184.02], [428.18, 184.17], [411.88, 184.91]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1212.89, 256.27], [1248.97, 256.59], [1367.87, 273.73], [1323.9, 273.27], [1212.89, 135.74], [1248.97, 135.6], [1367.87, 127.98], [1323.9, 128.18]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1032.78, 232.82], [1054.86, 232.93], [1107.94, 240.88], [1082.84, 240.74], [1032.78, 152.51], [1054.86, 152.47], [1107.94, 149.78], [1082.84, 149.83]]}, {\"category\": \"cyclist\", \"corners_3d\": [[998.2, 223.26], [1016.85, 223.92], [1008.76, 228.14], [988.71, 227.37], [998.2, 163.2], [1016.85, 163.07], [1008.76, 162.27], [988.71, 162.41]]}, {\"category\": \"car\", \"corners_3d\": [[921.52, 185.75], [894.32, 185.64], [886.07, 184.6], [910.96, 184.69], [921.52, 163.1], [894.32, 163.18], [886.07, 163.97], [910.96, 163.9]]}]\n```" - } - ] - }, - { - "id": 1354, - "data_source": "Mono3DRefer", - "scene_name": "000957", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[910.89, 290.19], [892.65, 284.36], [955.09, 283.88], [976.72, 289.65], [910.89, 145.32], [892.65, 147.19], [955.09, 147.34], [976.72, 145.49]]}]\n```" - } - ] - }, - { - "id": 1355, - "data_source": "Mono3DRefer", - "scene_name": "000960", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[573.75, 197.45], [551.3, 197.45], [555.13, 195.82], [576.09, 195.82], [573.75, 179.06], [551.3, 179.06], [555.13, 178.65], [576.09, 178.65]]}]\n```" - } - ] - }, - { - "id": 1356, - "data_source": "Mono3DRefer", - "scene_name": "000962", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[636.94, 196.83], [658.07, 196.8], [663.39, 198.46], [640.8, 198.49], [636.94, 176.07], [658.07, 176.06], [663.39, 176.28], [640.8, 176.29]]}, {\"category\": \"car\", \"corners_3d\": [[685.14, 230.13], [736.59, 230.01], [764.66, 240.85], [703.51, 241.02], [685.14, 182.22], [736.59, 182.2], [764.66, 183.97], [703.51, 184.0]]}, {\"category\": \"car\", \"corners_3d\": [[831.82, 305.06], [959.76, 304.37], [1239.42, 403.43], [1016.81, 405.56], [831.82, 184.85], [959.76, 184.78], [1239.42, 193.77], [1016.81, 193.96]]}, {\"category\": \"car\", \"corners_3d\": [[701.55, 198.31], [724.98, 198.26], [737.23, 200.23], [712.01, 200.28], [701.55, 175.41], [724.98, 175.41], [737.23, 175.6], [712.01, 175.61]]}, {\"category\": \"car\", \"corners_3d\": [[752.02, 209.39], [786.64, 209.25], [819.72, 214.39], [780.34, 214.58], [752.02, 176.71], [786.64, 176.69], [819.72, 177.23], [780.34, 177.25]]}]\n```" - } - ] - }, - { - "id": 1357, - "data_source": "Mono3DRefer", - "scene_name": "000962", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[925.76, 256.76], [1055.92, 256.31], [1672.79, 366.44], [1376.44, 368.87], [925.76, 72.86], [1055.92, 73.39], [1672.79, -57.87], [1376.44, -60.76]]}]\n```" - } - ] - }, - { - "id": 1358, - "data_source": "Mono3DRefer", - "scene_name": "000962", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[636.94, 196.83], [658.07, 196.8], [663.39, 198.46], [640.8, 198.49], [636.94, 176.07], [658.07, 176.06], [663.39, 176.28], [640.8, 176.29]]}, {\"category\": \"car\", \"corners_3d\": [[685.14, 230.13], [736.59, 230.01], [764.66, 240.85], [703.51, 241.02], [685.14, 182.22], [736.59, 182.2], [764.66, 183.97], [703.51, 184.0]]}, {\"category\": \"car\", \"corners_3d\": [[831.82, 305.06], [959.76, 304.37], [1239.42, 403.43], [1016.81, 405.56], [831.82, 184.85], [959.76, 184.78], [1239.42, 193.77], [1016.81, 193.96]]}, {\"category\": \"car\", \"corners_3d\": [[701.55, 198.31], [724.98, 198.26], [737.23, 200.23], [712.01, 200.28], [701.55, 175.41], [724.98, 175.41], [737.23, 175.6], [712.01, 175.61]]}, {\"category\": \"car\", \"corners_3d\": [[752.02, 209.39], [786.64, 209.25], [819.72, 214.39], [780.34, 214.58], [752.02, 176.71], [786.64, 176.69], [819.72, 177.23], [780.34, 177.25]]}, {\"category\": \"truck\", \"corners_3d\": [[925.76, 256.76], [1055.92, 256.31], [1672.79, 366.44], [1376.44, 368.87], [925.76, 72.86], [1055.92, 73.39], [1672.79, -57.87], [1376.44, -60.76]]}]\n```" - } - ] - }, - { - "id": 1359, - "data_source": "Mono3DRefer", - "scene_name": "000968", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[620.67, 201.3], [592.71, 201.16], [602.86, 198.56], [628.29, 198.68], [620.67, 177.06], [592.71, 177.04], [602.86, 176.65], [628.29, 176.67]]}]\n```" - } - ] - }, - { - "id": 1360, - "data_source": "Mono3DRefer", - "scene_name": "000973", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[748.08, 286.5], [842.26, 285.61], [1015.27, 356.42], [863.42, 358.79], [748.08, 191.57], [842.26, 191.42], [1015.27, 203.08], [863.42, 203.47]]}]\n```" - } - ] - }, - { - "id": 1361, - "data_source": "Mono3DRefer", - "scene_name": "000973", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[645.61, 222.46], [691.15, 222.34], [711.33, 231.42], [657.46, 231.6], [645.61, 169.52], [691.15, 169.53], [711.33, 168.92], [657.46, 168.91]]}]\n```" - } - ] - }, - { - "id": 1362, - "data_source": "Mono3DRefer", - "scene_name": "000973", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[748.08, 286.5], [842.26, 285.61], [1015.27, 356.42], [863.42, 358.79], [748.08, 191.57], [842.26, 191.42], [1015.27, 203.08], [863.42, 203.47]]}, {\"category\": \"van\", \"corners_3d\": [[645.61, 222.46], [691.15, 222.34], [711.33, 231.42], [657.46, 231.6], [645.61, 169.52], [691.15, 169.53], [711.33, 168.92], [657.46, 168.91]]}]\n```" - } - ] - }, - { - "id": 1363, - "data_source": "Mono3DRefer", - "scene_name": "000974", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[642.01, 190.43], [665.63, 190.39], [686.28, 193.58], [658.41, 193.65], [642.01, 164.6], [665.63, 164.62], [686.28, 163.13], [658.41, 163.1]]}]\n```" - } - ] - }, - { - "id": 1364, - "data_source": "Mono3DRefer", - "scene_name": "000974", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[601.01, 202.57], [628.21, 202.57], [629.5, 204.68], [600.36, 204.68], [601.01, 176.46], [628.21, 176.46], [629.5, 176.72], [600.36, 176.72]]}, {\"category\": \"car\", \"corners_3d\": [[387.36, 212.82], [356.87, 212.8], [378.1, 209.53], [406.1, 209.54], [387.36, 185.8], [356.87, 185.79], [378.1, 184.73], [406.1, 184.74]]}]\n```" - } - ] - }, - { - "id": 1365, - "data_source": "Mono3DRefer", - "scene_name": "000974", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[601.01, 202.57], [628.21, 202.57], [629.5, 204.68], [600.36, 204.68], [601.01, 176.46], [628.21, 176.46], [629.5, 176.72], [600.36, 176.72]]}, {\"category\": \"bus\", \"corners_3d\": [[642.01, 190.43], [665.63, 190.39], [686.28, 193.58], [658.41, 193.65], [642.01, 164.6], [665.63, 164.62], [686.28, 163.13], [658.41, 163.1]]}, {\"category\": \"car\", \"corners_3d\": [[387.36, 212.82], [356.87, 212.8], [378.1, 209.53], [406.1, 209.54], [387.36, 185.8], [356.87, 185.79], [378.1, 184.73], [406.1, 184.74]]}]\n```" - } - ] - }, - { - "id": 1366, - "data_source": "Mono3DRefer", - "scene_name": "000976", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[505.21, 278.78], [400.67, 279.41], [451.48, 249.23], [526.33, 248.91], [505.21, 186.15], [400.67, 186.23], [451.48, 182.44], [526.33, 182.4]]}, {\"category\": \"car\", \"corners_3d\": [[559.52, 224.83], [502.33, 224.5], [529.16, 216.08], [577.11, 216.31], [559.52, 170.61], [502.33, 170.62], [529.16, 170.99], [577.11, 170.98]]}, {\"category\": \"car\", \"corners_3d\": [[662.27, 199.85], [686.65, 200.44], [662.81, 202.02], [637.62, 201.36], [662.27, 172.14], [686.65, 172.13], [662.81, 172.08], [637.62, 172.1]]}]\n```" - } - ] - }, - { - "id": 1367, - "data_source": "Mono3DRefer", - "scene_name": "000976", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[159.92, 556.87], [-295.49, 566.52], [208.8, 339.79], [399.37, 338.03], [159.92, 120.93], [-295.49, 119.62], [208.8, 150.28], [399.37, 150.52]]}, {\"category\": \"van\", \"corners_3d\": [[814.69, 357.73], [994.44, 356.38], [1553.87, 609.28], [1131.94, 617.03], [814.69, 159.37], [994.44, 159.46], [1553.87, 141.01], [1131.94, 140.45]]}, {\"category\": \"van\", \"corners_3d\": [[537.84, 244.76], [467.3, 244.91], [492.72, 230.51], [549.15, 230.42], [537.84, 168.29], [467.3, 168.28], [492.72, 169.2], [549.15, 169.2]]}]\n```" - } - ] - }, - { - "id": 1368, - "data_source": "Mono3DRefer", - "scene_name": "000976", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[159.92, 556.87], [-295.49, 566.52], [208.8, 339.79], [399.37, 338.03], [159.92, 120.93], [-295.49, 119.62], [208.8, 150.28], [399.37, 150.52]]}, {\"category\": \"van\", \"corners_3d\": [[814.69, 357.73], [994.44, 356.38], [1553.87, 609.28], [1131.94, 617.03], [814.69, 159.37], [994.44, 159.46], [1553.87, 141.01], [1131.94, 140.45]]}, {\"category\": \"car\", \"corners_3d\": [[505.21, 278.78], [400.67, 279.41], [451.48, 249.23], [526.33, 248.91], [505.21, 186.15], [400.67, 186.23], [451.48, 182.44], [526.33, 182.4]]}, {\"category\": \"van\", \"corners_3d\": [[537.84, 244.76], [467.3, 244.91], [492.72, 230.51], [549.15, 230.42], [537.84, 168.29], [467.3, 168.28], [492.72, 169.2], [549.15, 169.2]]}, {\"category\": \"car\", \"corners_3d\": [[559.52, 224.83], [502.33, 224.5], [529.16, 216.08], [577.11, 216.31], [559.52, 170.61], [502.33, 170.62], [529.16, 170.99], [577.11, 170.98]]}, {\"category\": \"car\", \"corners_3d\": [[662.27, 199.85], [686.65, 200.44], [662.81, 202.02], [637.62, 201.36], [662.27, 172.14], [686.65, 172.13], [662.81, 172.08], [637.62, 172.1]]}]\n```" - } - ] - }, - { - "id": 1369, - "data_source": "Mono3DRefer", - "scene_name": "000977", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-161.39, 600.16], [-599.7, 612.68], [77.78, 358.56], [257.66, 356.29], [-161.39, 237.11], [-599.7, 238.99], [77.78, 200.78], [257.66, 200.44]]}, {\"category\": \"car\", \"corners_3d\": [[1331.5, 218.74], [1434.0, 225.6], [1137.6, 224.58], [1074.72, 217.97], [1331.5, 118.48], [1434.0, 110.36], [1137.6, 111.57], [1074.72, 119.41]]}, {\"category\": \"car\", \"corners_3d\": [[1229.11, 186.61], [1280.11, 186.63], [1397.69, 189.07], [1337.55, 189.05], [1229.11, 131.53], [1280.11, 131.5], [1397.69, 124.16], [1337.55, 124.2]]}, {\"category\": \"car\", \"corners_3d\": [[458.45, 261.19], [379.89, 260.91], [429.92, 243.34], [492.9, 243.52], [458.45, 193.74], [379.89, 193.67], [429.92, 189.52], [492.9, 189.56]]}, {\"category\": \"car\", \"corners_3d\": [[345.63, 318.49], [210.22, 318.23], [344.74, 270.06], [435.38, 270.18], [345.63, 195.11], [210.22, 195.07], [344.74, 187.71], [435.38, 187.73]]}, {\"category\": \"car\", \"corners_3d\": [[819.05, 224.84], [758.33, 224.76], [734.25, 215.49], [784.09, 215.55], [819.05, 171.78], [758.33, 171.78], [734.25, 171.97], [784.09, 171.97]]}, {\"category\": \"car\", \"corners_3d\": [[507.55, 236.07], [453.53, 235.79], [481.88, 226.82], [528.28, 227.02], [507.55, 184.28], [453.53, 184.23], [481.88, 182.6], [528.28, 182.64]]}, {\"category\": \"car\", \"corners_3d\": [[542.85, 222.78], [504.99, 222.73], [518.68, 216.96], [552.17, 217.0], [542.85, 185.85], [504.99, 185.84], [518.68, 184.33], [552.17, 184.34]]}, {\"category\": \"car\", \"corners_3d\": [[554.65, 216.93], [517.87, 216.94], [528.26, 211.91], [560.84, 211.91], [554.65, 185.25], [517.87, 185.25], [528.26, 183.84], [560.84, 183.84]]}, {\"category\": \"car\", \"corners_3d\": [[654.97, 209.96], [689.23, 210.26], [684.08, 214.55], [645.97, 214.18], [654.97, 178.43], [689.23, 178.48], [684.08, 179.12], [645.97, 179.07]]}, {\"category\": \"car\", \"corners_3d\": [[839.02, 211.06], [849.43, 213.38], [729.14, 212.85], [725.65, 210.59], [839.02, 171.79], [849.43, 171.73], [729.14, 171.74], [725.65, 171.8]]}, {\"category\": \"car\", \"corners_3d\": [[717.8, 202.78], [740.59, 203.53], [711.09, 205.28], [687.8, 204.45], [717.8, 176.3], [740.59, 176.39], [711.09, 176.59], [687.8, 176.49]]}, {\"category\": \"car\", \"corners_3d\": [[688.4, 183.33], [694.87, 183.53], [655.59, 183.74], [649.78, 183.53], [688.4, 170.53], [694.87, 170.49], [655.59, 170.44], [649.78, 170.49]]}, {\"category\": \"car\", \"corners_3d\": [[572.2, 186.36], [575.38, 186.62], [539.78, 186.81], [537.21, 186.55], [572.2, 174.26], [575.38, 174.28], [539.78, 174.3], [537.21, 174.28]]}]\n```" - } - ] - }, - { - "id": 1370, - "data_source": "Mono3DRefer", - "scene_name": "000978", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[518.23, 269.95], [440.97, 271.4], [452.2, 254.72], [516.4, 253.72], [518.23, 198.74], [440.97, 199.12], [452.2, 194.67], [516.4, 194.41]]}, {\"category\": \"car\", \"corners_3d\": [[523.84, 245.83], [476.92, 245.88], [494.98, 235.32], [535.11, 235.28], [523.84, 198.08], [476.92, 198.1], [494.98, 194.45], [535.11, 194.44]]}]\n```" - } - ] - }, - { - "id": 1371, - "data_source": "Mono3DRefer", - "scene_name": "000983", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[716.47, 267.53], [808.53, 268.26], [861.9, 301.1], [737.93, 299.79], [716.47, 189.38], [808.53, 189.5], [861.9, 195.24], [737.93, 195.01]]}, {\"category\": \"car\", \"corners_3d\": [[528.16, 217.09], [570.35, 217.27], [556.69, 223.99], [508.21, 223.74], [528.16, 181.49], [570.35, 181.53], [556.69, 182.84], [508.21, 182.79]]}, {\"category\": \"car\", \"corners_3d\": [[673.92, 204.28], [645.96, 204.28], [642.91, 201.68], [668.55, 201.68], [673.92, 178.4], [645.96, 178.4], [642.91, 177.95], [668.55, 177.95]]}, {\"category\": \"car\", \"corners_3d\": [[26.58, 270.05], [99.97, 257.99], [342.29, 258.32], [303.06, 270.49], [26.58, 170.32], [99.97, 170.64], [342.29, 170.63], [303.06, 170.31]]}, {\"category\": \"car\", \"corners_3d\": [[519.36, 247.09], [448.66, 246.88], [484.61, 232.25], [541.39, 232.39], [519.36, 172.84], [448.66, 172.84], [484.61, 172.84], [541.39, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[778.74, 338.04], [934.68, 338.78], [1090.29, 422.04], [855.58, 420.38], [778.74, 183.51], [934.68, 183.56], [1090.29, 188.93], [855.58, 188.82]]}]\n```" - } - ] - }, - { - "id": 1372, - "data_source": "Mono3DRefer", - "scene_name": "000983", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[594.67, 250.96], [571.33, 250.93], [574.27, 245.83], [596.08, 245.85], [594.67, 174.11], [571.33, 174.11], [574.27, 174.03], [596.08, 174.03]]}]\n```" - } - ] - }, - { - "id": 1373, - "data_source": "Mono3DRefer", - "scene_name": "000983", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[716.47, 267.53], [808.53, 268.26], [861.9, 301.1], [737.93, 299.79], [716.47, 189.38], [808.53, 189.5], [861.9, 195.24], [737.93, 195.01]]}, {\"category\": \"car\", \"corners_3d\": [[528.16, 217.09], [570.35, 217.27], [556.69, 223.99], [508.21, 223.74], [528.16, 181.49], [570.35, 181.53], [556.69, 182.84], [508.21, 182.79]]}, {\"category\": \"car\", \"corners_3d\": [[673.92, 204.28], [645.96, 204.28], [642.91, 201.68], [668.55, 201.68], [673.92, 178.4], [645.96, 178.4], [642.91, 177.95], [668.55, 177.95]]}, {\"category\": \"car\", \"corners_3d\": [[26.58, 270.05], [99.97, 257.99], [342.29, 258.32], [303.06, 270.49], [26.58, 170.32], [99.97, 170.64], [342.29, 170.63], [303.06, 170.31]]}, {\"category\": \"car\", \"corners_3d\": [[519.36, 247.09], [448.66, 246.88], [484.61, 232.25], [541.39, 232.39], [519.36, 172.84], [448.66, 172.84], [484.61, 172.84], [541.39, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[778.74, 338.04], [934.68, 338.78], [1090.29, 422.04], [855.58, 420.38], [778.74, 183.51], [934.68, 183.56], [1090.29, 188.93], [855.58, 188.82]]}, {\"category\": \"cyclist\", \"corners_3d\": [[594.67, 250.96], [571.33, 250.93], [574.27, 245.83], [596.08, 245.85], [594.67, 174.11], [571.33, 174.11], [574.27, 174.03], [596.08, 174.03]]}]\n```" - } - ] - }, - { - "id": 1374, - "data_source": "Mono3DRefer", - "scene_name": "000989", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[479.94, 222.24], [425.27, 222.08], [453.68, 216.61], [501.04, 216.72], [479.94, 178.04], [425.27, 178.06], [453.68, 178.52], [501.04, 178.51]]}]\n```" - } - ] - }, - { - "id": 1375, - "data_source": "Mono3DRefer", - "scene_name": "000993", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[757.95, 200.14], [778.84, 199.65], [829.43, 202.06], [807.62, 202.64], [757.95, 174.43], [778.84, 174.41], [829.43, 174.55], [807.62, 174.58]]}]\n```" - } - ] - }, - { - "id": 1376, - "data_source": "Mono3DRefer", - "scene_name": "001000", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[829.62, 250.03], [799.93, 241.13], [1005.13, 240.11], [1060.63, 248.73], [829.62, 171.77], [799.93, 171.9], [1005.13, 171.91], [1060.63, 171.79]]}, {\"category\": \"car\", \"corners_3d\": [[561.17, 205.63], [583.73, 205.53], [588.08, 208.62], [563.38, 208.74], [561.17, 184.43], [583.73, 184.39], [588.08, 185.48], [563.38, 185.52]]}, {\"category\": \"car\", \"corners_3d\": [[495.49, 211.11], [468.64, 211.2], [477.45, 207.72], [501.84, 207.65], [495.49, 188.38], [468.64, 188.41], [477.45, 187.0], [501.84, 186.97]]}]\n```" - } - ] - }, - { - "id": 1377, - "data_source": "Mono3DRefer", - "scene_name": "001001", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[441.09, 229.79], [400.88, 230.61], [401.65, 225.3], [438.2, 224.61], [441.09, 193.37], [400.88, 193.67], [401.65, 191.75], [438.2, 191.51]]}]\n```" - } - ] - }, - { - "id": 1378, - "data_source": "Mono3DRefer", - "scene_name": "001002", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[609.57, 212.18], [646.14, 212.24], [647.47, 215.91], [607.51, 215.83], [609.57, 175.42], [646.14, 175.43], [647.47, 175.66], [607.51, 175.66]]}, {\"category\": \"car\", \"corners_3d\": [[468.77, 257.1], [391.82, 256.65], [445.44, 238.91], [506.26, 239.19], [468.77, 190.53], [391.82, 190.43], [445.44, 186.71], [506.26, 186.77]]}, {\"category\": \"car\", \"corners_3d\": [[519.47, 228.39], [467.84, 228.51], [487.04, 219.57], [530.35, 219.48], [519.47, 181.24], [467.84, 181.26], [487.04, 179.91], [530.35, 179.89]]}, {\"category\": \"car\", \"corners_3d\": [[545.61, 208.19], [511.69, 208.14], [524.25, 204.41], [554.61, 204.45], [545.61, 179.16], [511.69, 179.15], [524.25, 178.48], [554.61, 178.49]]}, {\"category\": \"car\", \"corners_3d\": [[563.48, 194.64], [541.42, 194.6], [549.43, 193.03], [569.91, 193.07], [563.48, 172.99], [541.42, 172.99], [549.43, 172.98], [569.91, 172.98]]}]\n```" - } - ] - }, - { - "id": 1379, - "data_source": "Mono3DRefer", - "scene_name": "001003", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[513.62, 212.11], [528.11, 211.72], [538.44, 213.05], [523.57, 213.47], [513.62, 168.69], [528.11, 168.73], [538.44, 168.59], [523.57, 168.54]]}]\n```" - } - ] - }, - { - "id": 1380, - "data_source": "Mono3DRefer", - "scene_name": "001003", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[324.38, 354.19], [212.78, 411.96], [-323.02, 395.1], [-95.07, 344.32], [324.38, 208.67], [212.78, 220.08], [-323.02, 216.75], [-95.07, 206.72]]}, {\"category\": \"car\", \"corners_3d\": [[1468.43, 365.46], [1809.29, 441.9], [1055.45, 441.9], [928.77, 365.46], [1468.43, 150.58], [1809.29, 141.74], [1055.45, 141.74], [928.77, 150.58]]}, {\"category\": \"car\", \"corners_3d\": [[273.78, 249.46], [311.22, 242.33], [480.57, 243.88], [460.03, 251.36], [273.78, 183.97], [311.22, 182.94], [480.57, 183.17], [460.03, 184.25]]}, {\"category\": \"car\", \"corners_3d\": [[363.35, 230.19], [376.55, 225.98], [516.66, 225.06], [514.6, 229.12], [363.35, 182.8], [376.55, 182.07], [516.66, 181.91], [514.6, 182.61]]}, {\"category\": \"car\", \"corners_3d\": [[1005.5, 259.03], [1060.21, 271.16], [853.86, 270.88], [824.68, 258.82], [1005.5, 172.27], [1060.21, 172.19], [853.86, 172.19], [824.68, 172.27]]}, {\"category\": \"car\", \"corners_3d\": [[1130.93, 276.37], [1231.38, 296.6], [899.74, 296.03], [853.66, 275.98], [1130.93, 172.83], [1231.38, 172.82], [899.74, 172.82], [853.66, 172.83]]}]\n```" - } - ] - }, - { - "id": 1381, - "data_source": "Mono3DRefer", - "scene_name": "001003", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[324.38, 354.19], [212.78, 411.96], [-323.02, 395.1], [-95.07, 344.32], [324.38, 208.67], [212.78, 220.08], [-323.02, 216.75], [-95.07, 206.72]]}, {\"category\": \"car\", \"corners_3d\": [[1468.43, 365.46], [1809.29, 441.9], [1055.45, 441.9], [928.77, 365.46], [1468.43, 150.58], [1809.29, 141.74], [1055.45, 141.74], [928.77, 150.58]]}, {\"category\": \"car\", \"corners_3d\": [[273.78, 249.46], [311.22, 242.33], [480.57, 243.88], [460.03, 251.36], [273.78, 183.97], [311.22, 182.94], [480.57, 183.17], [460.03, 184.25]]}, {\"category\": \"car\", \"corners_3d\": [[363.35, 230.19], [376.55, 225.98], [516.66, 225.06], [514.6, 229.12], [363.35, 182.8], [376.55, 182.07], [516.66, 181.91], [514.6, 182.61]]}, {\"category\": \"car\", \"corners_3d\": [[1005.5, 259.03], [1060.21, 271.16], [853.86, 270.88], [824.68, 258.82], [1005.5, 172.27], [1060.21, 172.19], [853.86, 172.19], [824.68, 172.27]]}, {\"category\": \"car\", \"corners_3d\": [[1130.93, 276.37], [1231.38, 296.6], [899.74, 296.03], [853.66, 275.98], [1130.93, 172.83], [1231.38, 172.82], [899.74, 172.82], [853.66, 172.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[513.62, 212.11], [528.11, 211.72], [538.44, 213.05], [523.57, 213.47], [513.62, 168.69], [528.11, 168.73], [538.44, 168.59], [523.57, 168.54]]}]\n```" - } - ] - }, - { - "id": 1382, - "data_source": "Mono3DRefer", - "scene_name": "001005", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[711.0, 245.14], [785.39, 244.92], [835.29, 263.18], [742.16, 263.52], [711.0, 172.39], [785.39, 172.4], [835.29, 172.28], [742.16, 172.28]]}]\n```" - } - ] - }, - { - "id": 1383, - "data_source": "Mono3DRefer", - "scene_name": "001008", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[165.26, 595.99], [-41.0, 587.38], [226.45, 430.33], [355.32, 433.54], [165.26, 150.01], [-41.0, 150.67], [226.45, 162.6], [355.32, 162.36]]}]\n```" - } - ] - }, - { - "id": 1384, - "data_source": "Mono3DRefer", - "scene_name": "001008", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[674.94, 198.96], [720.98, 199.05], [734.79, 202.92], [678.54, 202.79], [674.94, 142.32], [720.98, 142.12], [734.79, 133.41], [678.54, 133.71]]}]\n```" - } - ] - }, - { - "id": 1385, - "data_source": "Mono3DRefer", - "scene_name": "001008", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[907.69, 276.86], [978.62, 277.24], [1016.12, 287.53], [937.43, 287.07], [907.69, 153.9], [978.62, 153.8], [1016.12, 150.82], [937.43, 150.95]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[768.05, 222.37], [805.28, 222.58], [809.11, 223.77], [770.79, 223.55], [768.05, 154.17], [805.28, 154.03], [809.11, 153.24], [770.79, 153.38]]}]\n```" - } - ] - }, - { - "id": 1386, - "data_source": "Mono3DRefer", - "scene_name": "001008", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[647.68, 195.15], [610.74, 195.09], [616.69, 193.59], [649.56, 193.64], [647.68, 162.29], [610.74, 162.38], [616.69, 164.5], [649.56, 164.43]]}]\n```" - } - ] - }, - { - "id": 1387, - "data_source": "Mono3DRefer", - "scene_name": "001008", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[165.26, 595.99], [-41.0, 587.38], [226.45, 430.33], [355.32, 433.54], [165.26, 150.01], [-41.0, 150.67], [226.45, 162.6], [355.32, 162.36]]}, {\"category\": \"van\", \"corners_3d\": [[674.94, 198.96], [720.98, 199.05], [734.79, 202.92], [678.54, 202.79], [674.94, 142.32], [720.98, 142.12], [734.79, 133.41], [678.54, 133.71]]}]\n```" - } - ] - }, - { - "id": 1388, - "data_source": "Mono3DRefer", - "scene_name": "001008", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[907.69, 276.86], [978.62, 277.24], [1016.12, 287.53], [937.43, 287.07], [907.69, 153.9], [978.62, 153.8], [1016.12, 150.82], [937.43, 150.95]]}, {\"category\": \"cyclist\", \"corners_3d\": [[165.26, 595.99], [-41.0, 587.38], [226.45, 430.33], [355.32, 433.54], [165.26, 150.01], [-41.0, 150.67], [226.45, 162.6], [355.32, 162.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[768.05, 222.37], [805.28, 222.58], [809.11, 223.77], [770.79, 223.55], [768.05, 154.17], [805.28, 154.03], [809.11, 153.24], [770.79, 153.38]]}]\n```" - } - ] - }, - { - "id": 1389, - "data_source": "Mono3DRefer", - "scene_name": "001008", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[165.26, 595.99], [-41.0, 587.38], [226.45, 430.33], [355.32, 433.54], [165.26, 150.01], [-41.0, 150.67], [226.45, 162.6], [355.32, 162.36]]}, {\"category\": \"car\", \"corners_3d\": [[647.68, 195.15], [610.74, 195.09], [616.69, 193.59], [649.56, 193.64], [647.68, 162.29], [610.74, 162.38], [616.69, 164.5], [649.56, 164.43]]}]\n```" - } - ] - }, - { - "id": 1390, - "data_source": "Mono3DRefer", - "scene_name": "001008", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[907.69, 276.86], [978.62, 277.24], [1016.12, 287.53], [937.43, 287.07], [907.69, 153.9], [978.62, 153.8], [1016.12, 150.82], [937.43, 150.95]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[768.05, 222.37], [805.28, 222.58], [809.11, 223.77], [770.79, 223.55], [768.05, 154.17], [805.28, 154.03], [809.11, 153.24], [770.79, 153.38]]}, {\"category\": \"van\", \"corners_3d\": [[674.94, 198.96], [720.98, 199.05], [734.79, 202.92], [678.54, 202.79], [674.94, 142.32], [720.98, 142.12], [734.79, 133.41], [678.54, 133.71]]}]\n```" - } - ] - }, - { - "id": 1391, - "data_source": "Mono3DRefer", - "scene_name": "001008", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[647.68, 195.15], [610.74, 195.09], [616.69, 193.59], [649.56, 193.64], [647.68, 162.29], [610.74, 162.38], [616.69, 164.5], [649.56, 164.43]]}, {\"category\": \"van\", \"corners_3d\": [[674.94, 198.96], [720.98, 199.05], [734.79, 202.92], [678.54, 202.79], [674.94, 142.32], [720.98, 142.12], [734.79, 133.41], [678.54, 133.71]]}]\n```" - } - ] - }, - { - "id": 1392, - "data_source": "Mono3DRefer", - "scene_name": "001008", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[907.69, 276.86], [978.62, 277.24], [1016.12, 287.53], [937.43, 287.07], [907.69, 153.9], [978.62, 153.8], [1016.12, 150.82], [937.43, 150.95]]}, {\"category\": \"car\", \"corners_3d\": [[647.68, 195.15], [610.74, 195.09], [616.69, 193.59], [649.56, 193.64], [647.68, 162.29], [610.74, 162.38], [616.69, 164.5], [649.56, 164.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[768.05, 222.37], [805.28, 222.58], [809.11, 223.77], [770.79, 223.55], [768.05, 154.17], [805.28, 154.03], [809.11, 153.24], [770.79, 153.38]]}]\n```" - } - ] - }, - { - "id": 1393, - "data_source": "Mono3DRefer", - "scene_name": "001008", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[907.69, 276.86], [978.62, 277.24], [1016.12, 287.53], [937.43, 287.07], [907.69, 153.9], [978.62, 153.8], [1016.12, 150.82], [937.43, 150.95]]}, {\"category\": \"cyclist\", \"corners_3d\": [[165.26, 595.99], [-41.0, 587.38], [226.45, 430.33], [355.32, 433.54], [165.26, 150.01], [-41.0, 150.67], [226.45, 162.6], [355.32, 162.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[768.05, 222.37], [805.28, 222.58], [809.11, 223.77], [770.79, 223.55], [768.05, 154.17], [805.28, 154.03], [809.11, 153.24], [770.79, 153.38]]}, {\"category\": \"van\", \"corners_3d\": [[674.94, 198.96], [720.98, 199.05], [734.79, 202.92], [678.54, 202.79], [674.94, 142.32], [720.98, 142.12], [734.79, 133.41], [678.54, 133.71]]}]\n```" - } - ] - }, - { - "id": 1394, - "data_source": "Mono3DRefer", - "scene_name": "001008", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[165.26, 595.99], [-41.0, 587.38], [226.45, 430.33], [355.32, 433.54], [165.26, 150.01], [-41.0, 150.67], [226.45, 162.6], [355.32, 162.36]]}, {\"category\": \"car\", \"corners_3d\": [[647.68, 195.15], [610.74, 195.09], [616.69, 193.59], [649.56, 193.64], [647.68, 162.29], [610.74, 162.38], [616.69, 164.5], [649.56, 164.43]]}, {\"category\": \"van\", \"corners_3d\": [[674.94, 198.96], [720.98, 199.05], [734.79, 202.92], [678.54, 202.79], [674.94, 142.32], [720.98, 142.12], [734.79, 133.41], [678.54, 133.71]]}]\n```" - } - ] - }, - { - "id": 1395, - "data_source": "Mono3DRefer", - "scene_name": "001008", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[907.69, 276.86], [978.62, 277.24], [1016.12, 287.53], [937.43, 287.07], [907.69, 153.9], [978.62, 153.8], [1016.12, 150.82], [937.43, 150.95]]}, {\"category\": \"cyclist\", \"corners_3d\": [[165.26, 595.99], [-41.0, 587.38], [226.45, 430.33], [355.32, 433.54], [165.26, 150.01], [-41.0, 150.67], [226.45, 162.6], [355.32, 162.36]]}, {\"category\": \"car\", \"corners_3d\": [[647.68, 195.15], [610.74, 195.09], [616.69, 193.59], [649.56, 193.64], [647.68, 162.29], [610.74, 162.38], [616.69, 164.5], [649.56, 164.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[768.05, 222.37], [805.28, 222.58], [809.11, 223.77], [770.79, 223.55], [768.05, 154.17], [805.28, 154.03], [809.11, 153.24], [770.79, 153.38]]}]\n```" - } - ] - }, - { - "id": 1396, - "data_source": "Mono3DRefer", - "scene_name": "001008", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[907.69, 276.86], [978.62, 277.24], [1016.12, 287.53], [937.43, 287.07], [907.69, 153.9], [978.62, 153.8], [1016.12, 150.82], [937.43, 150.95]]}, {\"category\": \"car\", \"corners_3d\": [[647.68, 195.15], [610.74, 195.09], [616.69, 193.59], [649.56, 193.64], [647.68, 162.29], [610.74, 162.38], [616.69, 164.5], [649.56, 164.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[768.05, 222.37], [805.28, 222.58], [809.11, 223.77], [770.79, 223.55], [768.05, 154.17], [805.28, 154.03], [809.11, 153.24], [770.79, 153.38]]}, {\"category\": \"van\", \"corners_3d\": [[674.94, 198.96], [720.98, 199.05], [734.79, 202.92], [678.54, 202.79], [674.94, 142.32], [720.98, 142.12], [734.79, 133.41], [678.54, 133.71]]}]\n```" - } - ] - }, - { - "id": 1397, - "data_source": "Mono3DRefer", - "scene_name": "001008", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[907.69, 276.86], [978.62, 277.24], [1016.12, 287.53], [937.43, 287.07], [907.69, 153.9], [978.62, 153.8], [1016.12, 150.82], [937.43, 150.95]]}, {\"category\": \"cyclist\", \"corners_3d\": [[165.26, 595.99], [-41.0, 587.38], [226.45, 430.33], [355.32, 433.54], [165.26, 150.01], [-41.0, 150.67], [226.45, 162.6], [355.32, 162.36]]}, {\"category\": \"car\", \"corners_3d\": [[647.68, 195.15], [610.74, 195.09], [616.69, 193.59], [649.56, 193.64], [647.68, 162.29], [610.74, 162.38], [616.69, 164.5], [649.56, 164.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[768.05, 222.37], [805.28, 222.58], [809.11, 223.77], [770.79, 223.55], [768.05, 154.17], [805.28, 154.03], [809.11, 153.24], [770.79, 153.38]]}, {\"category\": \"van\", \"corners_3d\": [[674.94, 198.96], [720.98, 199.05], [734.79, 202.92], [678.54, 202.79], [674.94, 142.32], [720.98, 142.12], [734.79, 133.41], [678.54, 133.71]]}]\n```" - } - ] - }, - { - "id": 1398, - "data_source": "Mono3DRefer", - "scene_name": "001009", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[538.98, 203.57], [574.37, 203.61], [566.45, 206.78], [526.02, 206.72], [538.98, 170.77], [574.37, 170.75], [566.45, 169.21], [526.02, 169.23]]}, {\"category\": \"car\", \"corners_3d\": [[560.93, 195.49], [586.89, 195.51], [583.84, 196.71], [555.66, 196.69], [560.93, 171.7], [586.89, 171.69], [583.84, 170.85], [555.66, 170.86]]}, {\"category\": \"car\", \"corners_3d\": [[613.19, 191.87], [640.32, 191.89], [639.9, 193.03], [609.81, 193.0], [613.19, 167.22], [640.32, 167.18], [639.9, 165.61], [609.81, 165.65]]}, {\"category\": \"car\", \"corners_3d\": [[451.79, 222.08], [399.17, 222.03], [432.96, 215.66], [477.34, 215.7], [451.79, 172.93], [399.17, 172.94], [432.96, 174.29], [477.34, 174.28]]}, {\"category\": \"car\", \"corners_3d\": [[584.52, 185.99], [605.57, 186.0], [604.34, 186.32], [581.78, 186.32], [584.52, 165.34], [605.57, 165.32], [604.34, 164.16], [581.78, 164.17]]}]\n```" - } - ] - }, - { - "id": 1399, - "data_source": "Mono3DRefer", - "scene_name": "001012", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[174.68, 452.68], [-112.83, 457.12], [208.99, 325.1], [361.39, 323.82], [174.68, 191.68], [-112.83, 191.98], [208.99, 183.1], [361.39, 183.01]]}, {\"category\": \"car\", \"corners_3d\": [[1192.13, 445.31], [929.54, 449.49], [756.93, 312.84], [892.14, 311.76], [1192.13, 200.37], [929.54, 200.79], [756.93, 186.99], [892.14, 186.88]]}, {\"category\": \"car\", \"corners_3d\": [[468.01, 260.86], [378.04, 260.98], [429.63, 240.66], [498.82, 240.59], [468.01, 179.94], [378.04, 179.95], [429.63, 178.32], [498.82, 178.31]]}, {\"category\": \"car\", \"corners_3d\": [[683.86, 242.55], [751.51, 242.55], [791.28, 262.15], [704.59, 262.15], [683.86, 186.86], [751.51, 186.86], [791.28, 190.8], [704.59, 190.8]]}, {\"category\": \"car\", \"corners_3d\": [[711.76, 226.72], [663.64, 226.72], [655.47, 218.67], [696.4, 218.67], [711.76, 183.74], [663.64, 183.74], [655.47, 182.11], [696.4, 182.11]]}, {\"category\": \"car\", \"corners_3d\": [[530.4, 221.86], [481.96, 221.99], [494.61, 215.66], [536.79, 215.56], [530.4, 180.18], [481.96, 180.2], [494.61, 179.25], [536.79, 179.24]]}]\n```" - } - ] - }, - { - "id": 1400, - "data_source": "Mono3DRefer", - "scene_name": "001018", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[717.62, 259.22], [803.83, 258.92], [907.86, 300.5], [780.33, 301.17], [717.62, 183.63], [803.83, 183.6], [907.86, 188.78], [780.33, 188.87]]}]\n```" - } - ] - }, - { - "id": 1401, - "data_source": "Mono3DRefer", - "scene_name": "001020", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[305.88, 320.51], [446.44, 320.8], [285.91, 459.09], [7.06, 457.92], [305.88, 190.24], [446.44, 190.26], [285.91, 198.95], [7.06, 198.87]]}, {\"category\": \"car\", \"corners_3d\": [[458.37, 237.4], [521.65, 237.45], [501.02, 250.8], [422.69, 250.72], [458.37, 179.39], [521.65, 179.39], [501.02, 178.89], [422.69, 178.89]]}, {\"category\": \"car\", \"corners_3d\": [[586.13, 214.01], [631.81, 214.03], [636.55, 220.54], [581.72, 220.51], [586.13, 172.45], [631.81, 172.44], [636.55, 170.63], [581.72, 170.64]]}, {\"category\": \"car\", \"corners_3d\": [[544.22, 204.5], [578.71, 204.52], [574.16, 207.37], [535.4, 207.34], [544.22, 170.62], [578.71, 170.61], [574.16, 169.26], [535.4, 169.27]]}, {\"category\": \"car\", \"corners_3d\": [[523.87, 194.19], [496.34, 194.18], [505.76, 193.17], [531.09, 193.18], [523.87, 173.18], [496.34, 173.18], [505.76, 173.85], [531.09, 173.84]]}]\n```" - } - ] - }, - { - "id": 1402, - "data_source": "Mono3DRefer", - "scene_name": "001021", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[454.74, 246.72], [384.3, 247.43], [411.23, 234.24], [469.1, 233.76], [454.74, 185.08], [384.3, 185.2], [411.23, 183.01], [469.1, 182.93]]}, {\"category\": \"car\", \"corners_3d\": [[518.78, 203.43], [503.72, 202.41], [565.9, 200.69], [582.09, 201.6], [518.78, 176.67], [503.72, 176.54], [565.9, 176.33], [582.09, 176.44]]}, {\"category\": \"car\", \"corners_3d\": [[434.51, 221.4], [392.46, 223.59], [344.16, 218.37], [383.93, 216.6], [434.51, 183.93], [392.46, 184.43], [344.16, 183.24], [383.93, 182.83]]}, {\"category\": \"car\", \"corners_3d\": [[632.95, 198.63], [620.64, 197.64], [687.6, 196.69], [701.98, 197.6], [632.95, 174.01], [620.64, 173.97], [687.6, 173.93], [701.98, 173.97]]}]\n```" - } - ] - }, - { - "id": 1403, - "data_source": "Mono3DRefer", - "scene_name": "001021", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[363.76, 204.93], [374.78, 203.48], [468.44, 203.47], [461.84, 204.92], [363.76, 154.78], [374.78, 155.59], [468.44, 155.6], [461.84, 154.78]]}]\n```" - } - ] - }, - { - "id": 1404, - "data_source": "Mono3DRefer", - "scene_name": "001021", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[748.82, 222.35], [726.33, 222.61], [717.62, 220.75], [739.32, 220.51], [748.82, 165.04], [726.33, 165.0], [717.62, 165.29], [739.32, 165.33]]}]\n```" - } - ] - }, - { - "id": 1405, - "data_source": "Mono3DRefer", - "scene_name": "001021", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[363.76, 204.93], [374.78, 203.48], [468.44, 203.47], [461.84, 204.92], [363.76, 154.78], [374.78, 155.59], [468.44, 155.6], [461.84, 154.78]]}, {\"category\": \"car\", \"corners_3d\": [[454.74, 246.72], [384.3, 247.43], [411.23, 234.24], [469.1, 233.76], [454.74, 185.08], [384.3, 185.2], [411.23, 183.01], [469.1, 182.93]]}, {\"category\": \"car\", \"corners_3d\": [[518.78, 203.43], [503.72, 202.41], [565.9, 200.69], [582.09, 201.6], [518.78, 176.67], [503.72, 176.54], [565.9, 176.33], [582.09, 176.44]]}, {\"category\": \"car\", \"corners_3d\": [[434.51, 221.4], [392.46, 223.59], [344.16, 218.37], [383.93, 216.6], [434.51, 183.93], [392.46, 184.43], [344.16, 183.24], [383.93, 182.83]]}, {\"category\": \"car\", \"corners_3d\": [[632.95, 198.63], [620.64, 197.64], [687.6, 196.69], [701.98, 197.6], [632.95, 174.01], [620.64, 173.97], [687.6, 173.93], [701.98, 173.97]]}]\n```" - } - ] - }, - { - "id": 1406, - "data_source": "Mono3DRefer", - "scene_name": "001021", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[454.74, 246.72], [384.3, 247.43], [411.23, 234.24], [469.1, 233.76], [454.74, 185.08], [384.3, 185.2], [411.23, 183.01], [469.1, 182.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[748.82, 222.35], [726.33, 222.61], [717.62, 220.75], [739.32, 220.51], [748.82, 165.04], [726.33, 165.0], [717.62, 165.29], [739.32, 165.33]]}, {\"category\": \"car\", \"corners_3d\": [[518.78, 203.43], [503.72, 202.41], [565.9, 200.69], [582.09, 201.6], [518.78, 176.67], [503.72, 176.54], [565.9, 176.33], [582.09, 176.44]]}, {\"category\": \"car\", \"corners_3d\": [[434.51, 221.4], [392.46, 223.59], [344.16, 218.37], [383.93, 216.6], [434.51, 183.93], [392.46, 184.43], [344.16, 183.24], [383.93, 182.83]]}, {\"category\": \"car\", \"corners_3d\": [[632.95, 198.63], [620.64, 197.64], [687.6, 196.69], [701.98, 197.6], [632.95, 174.01], [620.64, 173.97], [687.6, 173.93], [701.98, 173.97]]}]\n```" - } - ] - }, - { - "id": 1407, - "data_source": "Mono3DRefer", - "scene_name": "001021", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[363.76, 204.93], [374.78, 203.48], [468.44, 203.47], [461.84, 204.92], [363.76, 154.78], [374.78, 155.59], [468.44, 155.6], [461.84, 154.78]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[748.82, 222.35], [726.33, 222.61], [717.62, 220.75], [739.32, 220.51], [748.82, 165.04], [726.33, 165.0], [717.62, 165.29], [739.32, 165.33]]}]\n```" - } - ] - }, - { - "id": 1408, - "data_source": "Mono3DRefer", - "scene_name": "001021", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[363.76, 204.93], [374.78, 203.48], [468.44, 203.47], [461.84, 204.92], [363.76, 154.78], [374.78, 155.59], [468.44, 155.6], [461.84, 154.78]]}, {\"category\": \"car\", \"corners_3d\": [[454.74, 246.72], [384.3, 247.43], [411.23, 234.24], [469.1, 233.76], [454.74, 185.08], [384.3, 185.2], [411.23, 183.01], [469.1, 182.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[748.82, 222.35], [726.33, 222.61], [717.62, 220.75], [739.32, 220.51], [748.82, 165.04], [726.33, 165.0], [717.62, 165.29], [739.32, 165.33]]}, {\"category\": \"car\", \"corners_3d\": [[518.78, 203.43], [503.72, 202.41], [565.9, 200.69], [582.09, 201.6], [518.78, 176.67], [503.72, 176.54], [565.9, 176.33], [582.09, 176.44]]}, {\"category\": \"car\", \"corners_3d\": [[434.51, 221.4], [392.46, 223.59], [344.16, 218.37], [383.93, 216.6], [434.51, 183.93], [392.46, 184.43], [344.16, 183.24], [383.93, 182.83]]}, {\"category\": \"car\", \"corners_3d\": [[632.95, 198.63], [620.64, 197.64], [687.6, 196.69], [701.98, 197.6], [632.95, 174.01], [620.64, 173.97], [687.6, 173.93], [701.98, 173.97]]}]\n```" - } - ] - }, - { - "id": 1409, - "data_source": "Mono3DRefer", - "scene_name": "001024", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-28.95, 235.14], [-79.53, 235.5], [-7.56, 228.27], [36.83, 227.98], [-28.95, 195.65], [-79.53, 195.78], [-7.56, 193.14], [36.83, 193.03]]}, {\"category\": \"car\", \"corners_3d\": [[55.88, 226.05], [19.5, 226.11], [85.93, 219.95], [118.06, 219.91], [55.88, 193.56], [19.5, 193.58], [85.93, 191.19], [118.06, 191.17]]}, {\"category\": \"car\", \"corners_3d\": [[138.47, 218.06], [108.41, 218.17], [146.67, 214.38], [174.14, 214.28], [138.47, 187.2], [108.41, 187.23], [146.67, 186.03], [174.14, 186.0]]}, {\"category\": \"car\", \"corners_3d\": [[188.37, 213.75], [161.38, 213.88], [189.68, 210.85], [214.6, 210.73], [188.37, 188.97], [161.38, 189.02], [189.68, 187.82], [214.6, 187.78]]}, {\"category\": \"car\", \"corners_3d\": [[230.62, 212.19], [205.46, 212.26], [235.4, 209.05], [258.47, 208.99], [230.62, 187.39], [205.46, 187.42], [235.4, 186.23], [258.47, 186.21]]}, {\"category\": \"car\", \"corners_3d\": [[268.11, 203.68], [250.18, 203.67], [274.97, 201.58], [291.7, 201.59], [268.11, 183.93], [250.18, 183.92], [274.97, 183.17], [291.7, 183.18]]}]\n```" - } - ] - }, - { - "id": 1410, - "data_source": "Mono3DRefer", - "scene_name": "001024", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[557.09, 200.81], [605.12, 200.7], [617.76, 209.88], [553.88, 210.07], [557.09, 142.22], [605.12, 142.34], [617.76, 132.29], [553.88, 132.07]]}]\n```" - } - ] - }, - { - "id": 1411, - "data_source": "Mono3DRefer", - "scene_name": "001024", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[557.09, 200.81], [605.12, 200.7], [617.76, 209.88], [553.88, 210.07], [557.09, 142.22], [605.12, 142.34], [617.76, 132.29], [553.88, 132.07]]}, {\"category\": \"car\", \"corners_3d\": [[-28.95, 235.14], [-79.53, 235.5], [-7.56, 228.27], [36.83, 227.98], [-28.95, 195.65], [-79.53, 195.78], [-7.56, 193.14], [36.83, 193.03]]}, {\"category\": \"car\", \"corners_3d\": [[55.88, 226.05], [19.5, 226.11], [85.93, 219.95], [118.06, 219.91], [55.88, 193.56], [19.5, 193.58], [85.93, 191.19], [118.06, 191.17]]}, {\"category\": \"car\", \"corners_3d\": [[138.47, 218.06], [108.41, 218.17], [146.67, 214.38], [174.14, 214.28], [138.47, 187.2], [108.41, 187.23], [146.67, 186.03], [174.14, 186.0]]}, {\"category\": \"car\", \"corners_3d\": [[188.37, 213.75], [161.38, 213.88], [189.68, 210.85], [214.6, 210.73], [188.37, 188.97], [161.38, 189.02], [189.68, 187.82], [214.6, 187.78]]}, {\"category\": \"car\", \"corners_3d\": [[230.62, 212.19], [205.46, 212.26], [235.4, 209.05], [258.47, 208.99], [230.62, 187.39], [205.46, 187.42], [235.4, 186.23], [258.47, 186.21]]}, {\"category\": \"car\", \"corners_3d\": [[268.11, 203.68], [250.18, 203.67], [274.97, 201.58], [291.7, 201.59], [268.11, 183.93], [250.18, 183.92], [274.97, 183.17], [291.7, 183.18]]}]\n```" - } - ] - }, - { - "id": 1412, - "data_source": "Mono3DRefer", - "scene_name": "001025", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[641.21, 204.89], [661.69, 204.88], [666.68, 207.6], [644.46, 207.61], [641.21, 186.6], [661.69, 186.6], [666.68, 187.76], [644.46, 187.77]]}, {\"category\": \"car\", \"corners_3d\": [[348.92, 317.65], [191.04, 318.93], [330.34, 266.63], [431.23, 266.1], [348.92, 174.73], [191.04, 174.74], [330.34, 174.07], [431.23, 174.06]]}]\n```" - } - ] - }, - { - "id": 1413, - "data_source": "Mono3DRefer", - "scene_name": "001026", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[561.43, 228.48], [508.11, 228.2], [530.73, 219.93], [576.15, 220.14], [561.43, 180.17], [508.11, 180.13], [530.73, 179.05], [576.15, 179.07]]}, {\"category\": \"car\", \"corners_3d\": [[617.29, 202.57], [585.71, 202.53], [589.67, 199.91], [618.47, 199.94], [617.29, 175.32], [585.71, 175.32], [589.67, 175.1], [618.47, 175.11]]}]\n```" - } - ] - }, - { - "id": 1414, - "data_source": "Mono3DRefer", - "scene_name": "001026", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[822.84, 202.66], [814.28, 202.76], [808.51, 202.32], [816.97, 202.22], [822.84, 167.37], [814.28, 167.35], [808.51, 167.43], [816.97, 167.45]]}]\n```" - } - ] - }, - { - "id": 1415, - "data_source": "Mono3DRefer", - "scene_name": "001026", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[561.43, 228.48], [508.11, 228.2], [530.73, 219.93], [576.15, 220.14], [561.43, 180.17], [508.11, 180.13], [530.73, 179.05], [576.15, 179.07]]}, {\"category\": \"car\", \"corners_3d\": [[617.29, 202.57], [585.71, 202.53], [589.67, 199.91], [618.47, 199.94], [617.29, 175.32], [585.71, 175.32], [589.67, 175.1], [618.47, 175.11]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[822.84, 202.66], [814.28, 202.76], [808.51, 202.32], [816.97, 202.22], [822.84, 167.37], [814.28, 167.35], [808.51, 167.43], [816.97, 167.45]]}]\n```" - } - ] - }, - { - "id": 1416, - "data_source": "Mono3DRefer", - "scene_name": "001028", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[845.95, 318.15], [980.07, 318.43], [1304.98, 440.09], [1051.78, 439.1], [845.95, 193.69], [980.07, 193.71], [1304.98, 205.36], [1051.78, 205.27]]}, {\"category\": \"car\", \"corners_3d\": [[770.15, 255.9], [851.89, 255.99], [959.84, 289.92], [841.23, 289.72], [770.15, 180.42], [851.89, 180.42], [959.84, 180.38], [841.23, 180.38]]}, {\"category\": \"car\", \"corners_3d\": [[334.12, 218.61], [293.08, 218.31], [345.45, 213.39], [381.49, 213.62], [334.12, 170.27], [293.08, 170.35], [345.45, 171.67], [381.49, 171.61]]}, {\"category\": \"car\", \"corners_3d\": [[535.08, 196.1], [516.83, 196.03], [529.76, 195.09], [546.97, 195.16], [535.08, 178.2], [516.83, 178.21], [529.76, 178.35], [546.97, 178.34]]}, {\"category\": \"car\", \"corners_3d\": [[702.84, 220.05], [746.53, 220.43], [750.89, 226.69], [700.37, 226.17], [702.84, 178.0], [746.53, 177.98], [750.89, 177.58], [700.37, 177.61]]}, {\"category\": \"car\", \"corners_3d\": [[693.63, 215.32], [731.34, 215.43], [742.39, 220.01], [699.71, 219.87], [693.63, 180.92], [731.34, 180.93], [742.39, 180.98], [699.71, 180.98]]}, {\"category\": \"car\", \"corners_3d\": [[684.68, 209.05], [715.51, 209.17], [719.88, 212.31], [685.67, 212.16], [684.68, 180.86], [715.51, 180.86], [719.88, 180.9], [685.67, 180.89]]}, {\"category\": \"car\", \"corners_3d\": [[669.86, 199.6], [690.98, 199.64], [692.92, 200.9], [670.42, 200.85], [669.86, 180.48], [690.98, 180.48], [692.92, 180.48], [670.42, 180.48]]}]\n```" - } - ] - }, - { - "id": 1417, - "data_source": "Mono3DRefer", - "scene_name": "001030", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[416.41, 246.78], [343.41, 246.63], [395.52, 232.9], [455.0, 233.0], [416.41, 185.43], [343.41, 185.41], [395.52, 183.07], [455.0, 183.09]]}]\n```" - } - ] - }, - { - "id": 1418, - "data_source": "Mono3DRefer", - "scene_name": "001032", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[642.39, 202.59], [670.17, 202.52], [681.22, 205.69], [650.48, 205.77], [642.39, 172.67], [670.17, 172.67], [681.22, 172.65], [650.48, 172.65]]}, {\"category\": \"car\", \"corners_3d\": [[666.9, 211.56], [703.64, 211.52], [716.29, 216.05], [675.26, 216.1], [666.9, 178.48], [703.64, 178.47], [716.29, 179.13], [675.26, 179.14]]}, {\"category\": \"car\", \"corners_3d\": [[760.81, 211.0], [798.78, 210.92], [833.72, 217.03], [789.72, 217.14], [760.81, 176.42], [798.78, 176.41], [833.72, 176.99], [789.72, 177.0]]}, {\"category\": \"car\", \"corners_3d\": [[801.44, 200.8], [817.28, 200.31], [867.33, 202.33], [851.25, 202.89], [801.44, 175.42], [817.28, 175.38], [867.33, 175.56], [851.25, 175.61]]}]\n```" - } - ] - }, - { - "id": 1419, - "data_source": "Mono3DRefer", - "scene_name": "001032", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[651.83, 188.97], [678.77, 188.93], [693.56, 191.06], [663.09, 191.1], [651.83, 151.32], [678.77, 151.37], [693.56, 148.54], [663.09, 148.47]]}]\n```" - } - ] - }, - { - "id": 1420, - "data_source": "Mono3DRefer", - "scene_name": "001032", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[642.39, 202.59], [670.17, 202.52], [681.22, 205.69], [650.48, 205.77], [642.39, 172.67], [670.17, 172.67], [681.22, 172.65], [650.48, 172.65]]}, {\"category\": \"car\", \"corners_3d\": [[666.9, 211.56], [703.64, 211.52], [716.29, 216.05], [675.26, 216.1], [666.9, 178.48], [703.64, 178.47], [716.29, 179.13], [675.26, 179.14]]}, {\"category\": \"truck\", \"corners_3d\": [[651.83, 188.97], [678.77, 188.93], [693.56, 191.06], [663.09, 191.1], [651.83, 151.32], [678.77, 151.37], [693.56, 148.54], [663.09, 148.47]]}, {\"category\": \"car\", \"corners_3d\": [[760.81, 211.0], [798.78, 210.92], [833.72, 217.03], [789.72, 217.14], [760.81, 176.42], [798.78, 176.41], [833.72, 176.99], [789.72, 177.0]]}, {\"category\": \"car\", \"corners_3d\": [[801.44, 200.8], [817.28, 200.31], [867.33, 202.33], [851.25, 202.89], [801.44, 175.42], [817.28, 175.38], [867.33, 175.56], [851.25, 175.61]]}]\n```" - } - ] - }, - { - "id": 1421, - "data_source": "Mono3DRefer", - "scene_name": "001035", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[830.83, 304.97], [961.45, 304.99], [1209.15, 398.16], [986.4, 398.1], [830.83, 178.36], [961.45, 178.37], [1209.15, 182.25], [986.4, 182.25]]}, {\"category\": \"car\", \"corners_3d\": [[530.03, 259.44], [445.49, 259.45], [483.4, 239.37], [548.34, 239.36], [530.03, 162.98], [445.49, 162.98], [483.4, 165.27], [548.34, 165.27]]}, {\"category\": \"car\", \"corners_3d\": [[92.47, 588.22], [-242.39, 584.47], [210.94, 368.95], [371.67, 369.8], [92.47, 247.88], [-242.39, 247.2], [210.94, 208.27], [371.67, 208.43]]}, {\"category\": \"car\", \"corners_3d\": [[715.56, 228.15], [768.48, 228.23], [803.0, 241.51], [737.35, 241.38], [715.56, 183.31], [768.48, 183.33], [803.0, 185.84], [737.35, 185.82]]}, {\"category\": \"car\", \"corners_3d\": [[895.82, 275.28], [805.53, 274.78], [758.69, 246.31], [823.51, 246.57], [895.82, 184.63], [805.53, 184.57], [758.69, 181.3], [823.51, 181.33]]}, {\"category\": \"car\", \"corners_3d\": [[589.87, 204.99], [558.51, 205.01], [562.79, 201.82], [591.05, 201.81], [589.87, 174.46], [558.51, 174.46], [562.79, 174.31], [591.05, 174.3]]}, {\"category\": \"car\", \"corners_3d\": [[550.79, 234.16], [494.84, 234.02], [516.14, 224.43], [563.34, 224.52], [550.79, 179.37], [494.84, 179.36], [516.14, 178.34], [563.34, 178.35]]}, {\"category\": \"car\", \"corners_3d\": [[707.64, 219.82], [749.03, 219.85], [768.05, 226.63], [720.67, 226.6], [707.64, 179.32], [749.03, 179.33], [768.05, 180.26], [720.67, 180.26]]}, {\"category\": \"car\", \"corners_3d\": [[691.54, 210.25], [725.36, 210.22], [738.99, 214.15], [701.63, 214.19], [691.54, 178.28], [725.36, 178.28], [738.99, 178.85], [701.63, 178.85]]}, {\"category\": \"car\", \"corners_3d\": [[679.72, 204.86], [709.23, 204.85], [720.39, 208.21], [687.78, 208.23], [679.72, 176.83], [709.23, 176.82], [720.39, 177.24], [687.78, 177.24]]}]\n```" - } - ] - }, - { - "id": 1422, - "data_source": "Mono3DRefer", - "scene_name": "001035", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[920.74, 354.76], [838.49, 349.47], [836.26, 331.64], [909.92, 335.9], [920.74, 183.27], [838.49, 182.97], [836.26, 181.95], [909.92, 182.19]]}]\n```" - } - ] - }, - { - "id": 1423, - "data_source": "Mono3DRefer", - "scene_name": "001035", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[830.83, 304.97], [961.45, 304.99], [1209.15, 398.16], [986.4, 398.1], [830.83, 178.36], [961.45, 178.37], [1209.15, 182.25], [986.4, 182.25]]}, {\"category\": \"car\", \"corners_3d\": [[530.03, 259.44], [445.49, 259.45], [483.4, 239.37], [548.34, 239.36], [530.03, 162.98], [445.49, 162.98], [483.4, 165.27], [548.34, 165.27]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[920.74, 354.76], [838.49, 349.47], [836.26, 331.64], [909.92, 335.9], [920.74, 183.27], [838.49, 182.97], [836.26, 181.95], [909.92, 182.19]]}, {\"category\": \"car\", \"corners_3d\": [[92.47, 588.22], [-242.39, 584.47], [210.94, 368.95], [371.67, 369.8], [92.47, 247.88], [-242.39, 247.2], [210.94, 208.27], [371.67, 208.43]]}, {\"category\": \"car\", \"corners_3d\": [[715.56, 228.15], [768.48, 228.23], [803.0, 241.51], [737.35, 241.38], [715.56, 183.31], [768.48, 183.33], [803.0, 185.84], [737.35, 185.82]]}, {\"category\": \"car\", \"corners_3d\": [[895.82, 275.28], [805.53, 274.78], [758.69, 246.31], [823.51, 246.57], [895.82, 184.63], [805.53, 184.57], [758.69, 181.3], [823.51, 181.33]]}, {\"category\": \"car\", \"corners_3d\": [[589.87, 204.99], [558.51, 205.01], [562.79, 201.82], [591.05, 201.81], [589.87, 174.46], [558.51, 174.46], [562.79, 174.31], [591.05, 174.3]]}, {\"category\": \"car\", \"corners_3d\": [[550.79, 234.16], [494.84, 234.02], [516.14, 224.43], [563.34, 224.52], [550.79, 179.37], [494.84, 179.36], [516.14, 178.34], [563.34, 178.35]]}, {\"category\": \"car\", \"corners_3d\": [[707.64, 219.82], [749.03, 219.85], [768.05, 226.63], [720.67, 226.6], [707.64, 179.32], [749.03, 179.33], [768.05, 180.26], [720.67, 180.26]]}, {\"category\": \"car\", \"corners_3d\": [[691.54, 210.25], [725.36, 210.22], [738.99, 214.15], [701.63, 214.19], [691.54, 178.28], [725.36, 178.28], [738.99, 178.85], [701.63, 178.85]]}, {\"category\": \"car\", \"corners_3d\": [[679.72, 204.86], [709.23, 204.85], [720.39, 208.21], [687.78, 208.23], [679.72, 176.83], [709.23, 176.82], [720.39, 177.24], [687.78, 177.24]]}]\n```" - } - ] - }, - { - "id": 1424, - "data_source": "Mono3DRefer", - "scene_name": "001040", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[70.15, 306.23], [-38.03, 306.04], [136.8, 270.45], [216.22, 270.55], [70.15, 205.69], [-38.03, 205.64], [136.8, 196.88], [216.22, 196.9]]}, {\"category\": \"car\", \"corners_3d\": [[244.61, 261.43], [174.22, 261.26], [261.2, 244.14], [318.07, 244.25], [244.61, 195.54], [174.22, 195.5], [261.2, 191.11], [318.07, 191.14]]}, {\"category\": \"car\", \"corners_3d\": [[446.01, 213.89], [412.43, 213.85], [436.31, 209.21], [466.11, 209.24], [446.01, 181.34], [412.43, 181.33], [436.31, 180.37], [466.11, 180.38]]}]\n```" - } - ] - }, - { - "id": 1425, - "data_source": "Mono3DRefer", - "scene_name": "001040", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[863.55, 196.98], [873.66, 196.98], [888.75, 198.36], [878.05, 198.36], [863.55, 156.07], [873.66, 156.07], [888.75, 155.11], [878.05, 155.11]]}]\n```" - } - ] - }, - { - "id": 1426, - "data_source": "Mono3DRefer", - "scene_name": "001040", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[650.1, 183.19], [669.8, 183.17], [715.86, 186.97], [689.02, 187.02], [650.1, 155.2], [669.8, 155.24], [715.86, 148.74], [689.02, 148.66]]}]\n```" - } - ] - }, - { - "id": 1427, - "data_source": "Mono3DRefer", - "scene_name": "001040", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[70.15, 306.23], [-38.03, 306.04], [136.8, 270.45], [216.22, 270.55], [70.15, 205.69], [-38.03, 205.64], [136.8, 196.88], [216.22, 196.9]]}, {\"category\": \"car\", \"corners_3d\": [[244.61, 261.43], [174.22, 261.26], [261.2, 244.14], [318.07, 244.25], [244.61, 195.54], [174.22, 195.5], [261.2, 191.11], [318.07, 191.14]]}, {\"category\": \"car\", \"corners_3d\": [[446.01, 213.89], [412.43, 213.85], [436.31, 209.21], [466.11, 209.24], [446.01, 181.34], [412.43, 181.33], [436.31, 180.37], [466.11, 180.38]]}, {\"category\": \"cyclist\", \"corners_3d\": [[863.55, 196.98], [873.66, 196.98], [888.75, 198.36], [878.05, 198.36], [863.55, 156.07], [873.66, 156.07], [888.75, 155.11], [878.05, 155.11]]}]\n```" - } - ] - }, - { - "id": 1428, - "data_source": "Mono3DRefer", - "scene_name": "001040", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[650.1, 183.19], [669.8, 183.17], [715.86, 186.97], [689.02, 187.02], [650.1, 155.2], [669.8, 155.24], [715.86, 148.74], [689.02, 148.66]]}, {\"category\": \"car\", \"corners_3d\": [[70.15, 306.23], [-38.03, 306.04], [136.8, 270.45], [216.22, 270.55], [70.15, 205.69], [-38.03, 205.64], [136.8, 196.88], [216.22, 196.9]]}, {\"category\": \"car\", \"corners_3d\": [[244.61, 261.43], [174.22, 261.26], [261.2, 244.14], [318.07, 244.25], [244.61, 195.54], [174.22, 195.5], [261.2, 191.11], [318.07, 191.14]]}, {\"category\": \"car\", \"corners_3d\": [[446.01, 213.89], [412.43, 213.85], [436.31, 209.21], [466.11, 209.24], [446.01, 181.34], [412.43, 181.33], [436.31, 180.37], [466.11, 180.38]]}]\n```" - } - ] - }, - { - "id": 1429, - "data_source": "Mono3DRefer", - "scene_name": "001040", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[650.1, 183.19], [669.8, 183.17], [715.86, 186.97], [689.02, 187.02], [650.1, 155.2], [669.8, 155.24], [715.86, 148.74], [689.02, 148.66]]}, {\"category\": \"cyclist\", \"corners_3d\": [[863.55, 196.98], [873.66, 196.98], [888.75, 198.36], [878.05, 198.36], [863.55, 156.07], [873.66, 156.07], [888.75, 155.11], [878.05, 155.11]]}]\n```" - } - ] - }, - { - "id": 1430, - "data_source": "Mono3DRefer", - "scene_name": "001040", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[650.1, 183.19], [669.8, 183.17], [715.86, 186.97], [689.02, 187.02], [650.1, 155.2], [669.8, 155.24], [715.86, 148.74], [689.02, 148.66]]}, {\"category\": \"car\", \"corners_3d\": [[70.15, 306.23], [-38.03, 306.04], [136.8, 270.45], [216.22, 270.55], [70.15, 205.69], [-38.03, 205.64], [136.8, 196.88], [216.22, 196.9]]}, {\"category\": \"car\", \"corners_3d\": [[244.61, 261.43], [174.22, 261.26], [261.2, 244.14], [318.07, 244.25], [244.61, 195.54], [174.22, 195.5], [261.2, 191.11], [318.07, 191.14]]}, {\"category\": \"car\", \"corners_3d\": [[446.01, 213.89], [412.43, 213.85], [436.31, 209.21], [466.11, 209.24], [446.01, 181.34], [412.43, 181.33], [436.31, 180.37], [466.11, 180.38]]}, {\"category\": \"cyclist\", \"corners_3d\": [[863.55, 196.98], [873.66, 196.98], [888.75, 198.36], [878.05, 198.36], [863.55, 156.07], [873.66, 156.07], [888.75, 155.11], [878.05, 155.11]]}]\n```" - } - ] - }, - { - "id": 1431, - "data_source": "Mono3DRefer", - "scene_name": "001041", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[322.11, 407.64], [93.47, 409.18], [312.75, 305.76], [440.89, 305.27], [322.11, 214.89], [93.47, 215.16], [312.75, 196.65], [440.89, 196.56]]}, {\"category\": \"car\", \"corners_3d\": [[574.27, 189.95], [547.55, 189.97], [550.82, 188.57], [575.34, 188.55], [574.27, 161.87], [547.55, 161.86], [550.82, 162.76], [575.34, 162.77]]}]\n```" - } - ] - }, - { - "id": 1432, - "data_source": "Mono3DRefer", - "scene_name": "001042", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[392.83, 276.17], [289.47, 275.28], [386.18, 247.98], [462.44, 248.46], [392.83, 170.43], [289.47, 170.45], [386.18, 171.09], [462.44, 171.08]]}, {\"category\": \"car\", \"corners_3d\": [[502.88, 240.05], [443.72, 239.38], [489.65, 227.27], [538.33, 227.72], [502.88, 185.1], [443.72, 184.97], [489.65, 182.77], [538.33, 182.85]]}, {\"category\": \"car\", \"corners_3d\": [[606.3, 209.54], [573.45, 209.34], [584.58, 206.02], [614.49, 206.19], [606.3, 181.07], [573.45, 181.02], [584.58, 180.28], [614.49, 180.32]]}, {\"category\": \"car\", \"corners_3d\": [[674.79, 196.01], [655.47, 195.79], [667.7, 194.5], [686.03, 194.69], [674.79, 178.36], [655.47, 178.31], [667.7, 178.0], [686.03, 178.05]]}]\n```" - } - ] - }, - { - "id": 1433, - "data_source": "Mono3DRefer", - "scene_name": "001042", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[686.41, 227.64], [708.07, 227.79], [710.84, 232.43], [687.35, 232.26], [686.41, 172.22], [708.07, 172.22], [710.84, 172.16], [687.35, 172.17]]}]\n```" - } - ] - }, - { - "id": 1434, - "data_source": "Mono3DRefer", - "scene_name": "001042", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[392.83, 276.17], [289.47, 275.28], [386.18, 247.98], [462.44, 248.46], [392.83, 170.43], [289.47, 170.45], [386.18, 171.09], [462.44, 171.08]]}, {\"category\": \"car\", \"corners_3d\": [[502.88, 240.05], [443.72, 239.38], [489.65, 227.27], [538.33, 227.72], [502.88, 185.1], [443.72, 184.97], [489.65, 182.77], [538.33, 182.85]]}, {\"category\": \"car\", \"corners_3d\": [[606.3, 209.54], [573.45, 209.34], [584.58, 206.02], [614.49, 206.19], [606.3, 181.07], [573.45, 181.02], [584.58, 180.28], [614.49, 180.32]]}, {\"category\": \"car\", \"corners_3d\": [[674.79, 196.01], [655.47, 195.79], [667.7, 194.5], [686.03, 194.69], [674.79, 178.36], [655.47, 178.31], [667.7, 178.0], [686.03, 178.05]]}, {\"category\": \"cyclist\", \"corners_3d\": [[686.41, 227.64], [708.07, 227.79], [710.84, 232.43], [687.35, 232.26], [686.41, 172.22], [708.07, 172.22], [710.84, 172.16], [687.35, 172.17]]}]\n```" - } - ] - }, - { - "id": 1435, - "data_source": "Mono3DRefer", - "scene_name": "001045", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[660.74, 224.09], [726.12, 223.91], [773.31, 240.54], [686.76, 240.86], [660.74, 145.6], [726.12, 145.69], [773.31, 136.85], [686.76, 136.68]]}, {\"category\": \"van\", \"corners_3d\": [[358.76, 220.0], [302.04, 220.01], [350.1, 212.62], [397.94, 212.62], [358.76, 159.25], [302.04, 159.24], [350.1, 161.38], [397.94, 161.38]]}]\n```" - } - ] - }, - { - "id": 1436, - "data_source": "Mono3DRefer", - "scene_name": "001045", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[176.19, 210.52], [161.78, 210.52], [184.24, 208.63], [197.92, 208.63], [176.19, 175.01], [161.78, 175.01], [184.24, 174.9], [197.92, 174.9]]}]\n```" - } - ] - }, - { - "id": 1437, - "data_source": "Mono3DRefer", - "scene_name": "001045", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[569.42, 218.16], [617.88, 218.22], [616.84, 225.27], [560.86, 225.18], [569.42, 179.15], [617.88, 179.16], [616.84, 180.14], [560.86, 180.13]]}, {\"category\": \"car\", \"corners_3d\": [[91.41, 284.16], [182.13, 270.42], [370.67, 279.72], [299.43, 296.44], [91.41, 176.09], [182.13, 175.69], [370.67, 175.96], [299.43, 176.44]]}, {\"category\": \"car\", \"corners_3d\": [[352.59, 224.01], [392.79, 221.2], [478.72, 225.91], [439.86, 229.32], [352.59, 178.77], [392.79, 178.45], [478.72, 178.99], [439.86, 179.38]]}, {\"category\": \"car\", \"corners_3d\": [[393.76, 217.63], [428.26, 215.58], [497.52, 219.36], [463.44, 221.79], [393.76, 180.26], [428.26, 179.92], [497.52, 180.55], [463.44, 180.95]]}, {\"category\": \"car\", \"corners_3d\": [[526.03, 206.81], [563.93, 206.79], [559.09, 211.0], [516.49, 211.02], [526.03, 179.59], [563.93, 179.59], [559.09, 180.43], [516.49, 180.43]]}, {\"category\": \"car\", \"corners_3d\": [[583.3, 206.77], [618.05, 206.75], [619.81, 210.69], [581.02, 210.71], [583.3, 175.14], [618.05, 175.13], [619.81, 175.4], [581.02, 175.4]]}]\n```" - } - ] - }, - { - "id": 1438, - "data_source": "Mono3DRefer", - "scene_name": "001045", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[660.74, 224.09], [726.12, 223.91], [773.31, 240.54], [686.76, 240.86], [660.74, 145.6], [726.12, 145.69], [773.31, 136.85], [686.76, 136.68]]}, {\"category\": \"van\", \"corners_3d\": [[358.76, 220.0], [302.04, 220.01], [350.1, 212.62], [397.94, 212.62], [358.76, 159.25], [302.04, 159.24], [350.1, 161.38], [397.94, 161.38]]}, {\"category\": \"cyclist\", \"corners_3d\": [[176.19, 210.52], [161.78, 210.52], [184.24, 208.63], [197.92, 208.63], [176.19, 175.01], [161.78, 175.01], [184.24, 174.9], [197.92, 174.9]]}]\n```" - } - ] - }, - { - "id": 1439, - "data_source": "Mono3DRefer", - "scene_name": "001045", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[569.42, 218.16], [617.88, 218.22], [616.84, 225.27], [560.86, 225.18], [569.42, 179.15], [617.88, 179.16], [616.84, 180.14], [560.86, 180.13]]}, {\"category\": \"car\", \"corners_3d\": [[91.41, 284.16], [182.13, 270.42], [370.67, 279.72], [299.43, 296.44], [91.41, 176.09], [182.13, 175.69], [370.67, 175.96], [299.43, 176.44]]}, {\"category\": \"car\", \"corners_3d\": [[352.59, 224.01], [392.79, 221.2], [478.72, 225.91], [439.86, 229.32], [352.59, 178.77], [392.79, 178.45], [478.72, 178.99], [439.86, 179.38]]}, {\"category\": \"car\", \"corners_3d\": [[393.76, 217.63], [428.26, 215.58], [497.52, 219.36], [463.44, 221.79], [393.76, 180.26], [428.26, 179.92], [497.52, 180.55], [463.44, 180.95]]}, {\"category\": \"car\", \"corners_3d\": [[526.03, 206.81], [563.93, 206.79], [559.09, 211.0], [516.49, 211.02], [526.03, 179.59], [563.93, 179.59], [559.09, 180.43], [516.49, 180.43]]}, {\"category\": \"van\", \"corners_3d\": [[660.74, 224.09], [726.12, 223.91], [773.31, 240.54], [686.76, 240.86], [660.74, 145.6], [726.12, 145.69], [773.31, 136.85], [686.76, 136.68]]}, {\"category\": \"van\", \"corners_3d\": [[358.76, 220.0], [302.04, 220.01], [350.1, 212.62], [397.94, 212.62], [358.76, 159.25], [302.04, 159.24], [350.1, 161.38], [397.94, 161.38]]}, {\"category\": \"car\", \"corners_3d\": [[583.3, 206.77], [618.05, 206.75], [619.81, 210.69], [581.02, 210.71], [583.3, 175.14], [618.05, 175.13], [619.81, 175.4], [581.02, 175.4]]}]\n```" - } - ] - }, - { - "id": 1440, - "data_source": "Mono3DRefer", - "scene_name": "001045", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[569.42, 218.16], [617.88, 218.22], [616.84, 225.27], [560.86, 225.18], [569.42, 179.15], [617.88, 179.16], [616.84, 180.14], [560.86, 180.13]]}, {\"category\": \"car\", \"corners_3d\": [[91.41, 284.16], [182.13, 270.42], [370.67, 279.72], [299.43, 296.44], [91.41, 176.09], [182.13, 175.69], [370.67, 175.96], [299.43, 176.44]]}, {\"category\": \"car\", \"corners_3d\": [[352.59, 224.01], [392.79, 221.2], [478.72, 225.91], [439.86, 229.32], [352.59, 178.77], [392.79, 178.45], [478.72, 178.99], [439.86, 179.38]]}, {\"category\": \"car\", \"corners_3d\": [[393.76, 217.63], [428.26, 215.58], [497.52, 219.36], [463.44, 221.79], [393.76, 180.26], [428.26, 179.92], [497.52, 180.55], [463.44, 180.95]]}, {\"category\": \"car\", \"corners_3d\": [[526.03, 206.81], [563.93, 206.79], [559.09, 211.0], [516.49, 211.02], [526.03, 179.59], [563.93, 179.59], [559.09, 180.43], [516.49, 180.43]]}, {\"category\": \"car\", \"corners_3d\": [[583.3, 206.77], [618.05, 206.75], [619.81, 210.69], [581.02, 210.71], [583.3, 175.14], [618.05, 175.13], [619.81, 175.4], [581.02, 175.4]]}, {\"category\": \"cyclist\", \"corners_3d\": [[176.19, 210.52], [161.78, 210.52], [184.24, 208.63], [197.92, 208.63], [176.19, 175.01], [161.78, 175.01], [184.24, 174.9], [197.92, 174.9]]}]\n```" - } - ] - }, - { - "id": 1441, - "data_source": "Mono3DRefer", - "scene_name": "001045", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[569.42, 218.16], [617.88, 218.22], [616.84, 225.27], [560.86, 225.18], [569.42, 179.15], [617.88, 179.16], [616.84, 180.14], [560.86, 180.13]]}, {\"category\": \"car\", \"corners_3d\": [[91.41, 284.16], [182.13, 270.42], [370.67, 279.72], [299.43, 296.44], [91.41, 176.09], [182.13, 175.69], [370.67, 175.96], [299.43, 176.44]]}, {\"category\": \"car\", \"corners_3d\": [[352.59, 224.01], [392.79, 221.2], [478.72, 225.91], [439.86, 229.32], [352.59, 178.77], [392.79, 178.45], [478.72, 178.99], [439.86, 179.38]]}, {\"category\": \"car\", \"corners_3d\": [[393.76, 217.63], [428.26, 215.58], [497.52, 219.36], [463.44, 221.79], [393.76, 180.26], [428.26, 179.92], [497.52, 180.55], [463.44, 180.95]]}, {\"category\": \"car\", \"corners_3d\": [[526.03, 206.81], [563.93, 206.79], [559.09, 211.0], [516.49, 211.02], [526.03, 179.59], [563.93, 179.59], [559.09, 180.43], [516.49, 180.43]]}, {\"category\": \"van\", \"corners_3d\": [[660.74, 224.09], [726.12, 223.91], [773.31, 240.54], [686.76, 240.86], [660.74, 145.6], [726.12, 145.69], [773.31, 136.85], [686.76, 136.68]]}, {\"category\": \"van\", \"corners_3d\": [[358.76, 220.0], [302.04, 220.01], [350.1, 212.62], [397.94, 212.62], [358.76, 159.25], [302.04, 159.24], [350.1, 161.38], [397.94, 161.38]]}, {\"category\": \"car\", \"corners_3d\": [[583.3, 206.77], [618.05, 206.75], [619.81, 210.69], [581.02, 210.71], [583.3, 175.14], [618.05, 175.13], [619.81, 175.4], [581.02, 175.4]]}, {\"category\": \"cyclist\", \"corners_3d\": [[176.19, 210.52], [161.78, 210.52], [184.24, 208.63], [197.92, 208.63], [176.19, 175.01], [161.78, 175.01], [184.24, 174.9], [197.92, 174.9]]}]\n```" - } - ] - }, - { - "id": 1442, - "data_source": "Mono3DRefer", - "scene_name": "001047", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[690.84, 195.04], [711.48, 194.86], [735.12, 196.64], [713.0, 196.84], [690.84, 174.14], [711.48, 174.13], [735.12, 174.23], [713.0, 174.24]]}]\n```" - } - ] - }, - { - "id": 1443, - "data_source": "Mono3DRefer", - "scene_name": "001051", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[574.54, 237.02], [537.08, 237.09], [543.88, 229.49], [576.91, 229.44], [574.54, 202.4], [537.08, 202.43], [543.88, 198.93], [576.91, 198.9]]}]\n```" - } - ] - }, - { - "id": 1444, - "data_source": "Mono3DRefer", - "scene_name": "001052", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[563.76, 222.42], [611.63, 222.45], [610.75, 230.05], [555.56, 230.0], [563.76, 183.89], [611.63, 183.9], [610.75, 185.59], [555.56, 185.58]]}, {\"category\": \"car\", \"corners_3d\": [[271.42, 247.04], [317.02, 240.85], [449.31, 244.94], [413.53, 251.93], [271.42, 177.45], [317.02, 177.07], [449.31, 177.32], [413.53, 177.75]]}, {\"category\": \"car\", \"corners_3d\": [[402.34, 214.97], [432.01, 213.11], [502.98, 216.14], [474.52, 218.3], [402.34, 179.25], [432.01, 178.97], [502.98, 179.43], [474.52, 179.76]]}, {\"category\": \"car\", \"corners_3d\": [[430.64, 212.54], [457.51, 211.05], [515.39, 213.77], [488.95, 215.48], [430.64, 182.21], [457.51, 181.86], [515.39, 182.5], [488.95, 182.9]]}, {\"category\": \"car\", \"corners_3d\": [[535.47, 203.86], [567.42, 203.85], [563.78, 207.03], [528.55, 207.05], [535.47, 180.92], [567.42, 180.92], [563.78, 181.75], [528.55, 181.75]]}, {\"category\": \"car\", \"corners_3d\": [[854.72, 348.33], [1024.15, 347.11], [1633.08, 590.68], [1233.13, 597.73], [854.72, 202.93], [1024.15, 202.73], [1633.08, 244.48], [1233.13, 245.69]]}, {\"category\": \"car\", \"corners_3d\": [[581.72, 210.63], [616.58, 210.63], [617.33, 215.03], [578.41, 215.03], [581.72, 178.9], [616.58, 178.9], [617.33, 179.6], [578.41, 179.6]]}]\n```" - } - ] - }, - { - "id": 1445, - "data_source": "Mono3DRefer", - "scene_name": "001052", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[644.93, 215.97], [695.97, 215.79], [726.64, 225.98], [663.58, 226.26], [644.93, 154.58], [695.97, 154.66], [726.64, 150.34], [663.58, 150.22]]}, {\"category\": \"van\", \"corners_3d\": [[404.31, 211.29], [359.49, 211.25], [393.43, 206.31], [432.51, 206.35], [404.31, 162.95], [359.49, 162.96], [393.43, 164.23], [432.51, 164.22]]}]\n```" - } - ] - }, - { - "id": 1446, - "data_source": "Mono3DRefer", - "scene_name": "001052", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[563.76, 222.42], [611.63, 222.45], [610.75, 230.05], [555.56, 230.0], [563.76, 183.89], [611.63, 183.9], [610.75, 185.59], [555.56, 185.58]]}, {\"category\": \"car\", \"corners_3d\": [[271.42, 247.04], [317.02, 240.85], [449.31, 244.94], [413.53, 251.93], [271.42, 177.45], [317.02, 177.07], [449.31, 177.32], [413.53, 177.75]]}, {\"category\": \"car\", \"corners_3d\": [[402.34, 214.97], [432.01, 213.11], [502.98, 216.14], [474.52, 218.3], [402.34, 179.25], [432.01, 178.97], [502.98, 179.43], [474.52, 179.76]]}, {\"category\": \"car\", \"corners_3d\": [[430.64, 212.54], [457.51, 211.05], [515.39, 213.77], [488.95, 215.48], [430.64, 182.21], [457.51, 181.86], [515.39, 182.5], [488.95, 182.9]]}, {\"category\": \"car\", \"corners_3d\": [[535.47, 203.86], [567.42, 203.85], [563.78, 207.03], [528.55, 207.05], [535.47, 180.92], [567.42, 180.92], [563.78, 181.75], [528.55, 181.75]]}, {\"category\": \"car\", \"corners_3d\": [[854.72, 348.33], [1024.15, 347.11], [1633.08, 590.68], [1233.13, 597.73], [854.72, 202.93], [1024.15, 202.73], [1633.08, 244.48], [1233.13, 245.69]]}, {\"category\": \"van\", \"corners_3d\": [[644.93, 215.97], [695.97, 215.79], [726.64, 225.98], [663.58, 226.26], [644.93, 154.58], [695.97, 154.66], [726.64, 150.34], [663.58, 150.22]]}, {\"category\": \"van\", \"corners_3d\": [[404.31, 211.29], [359.49, 211.25], [393.43, 206.31], [432.51, 206.35], [404.31, 162.95], [359.49, 162.96], [393.43, 164.23], [432.51, 164.22]]}, {\"category\": \"car\", \"corners_3d\": [[581.72, 210.63], [616.58, 210.63], [617.33, 215.03], [578.41, 215.03], [581.72, 178.9], [616.58, 178.9], [617.33, 179.6], [578.41, 179.6]]}]\n```" - } - ] - }, - { - "id": 1447, - "data_source": "Mono3DRefer", - "scene_name": "001053", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[595.98, 190.34], [619.82, 190.35], [620.27, 193.56], [592.06, 193.55], [595.98, 164.51], [619.82, 164.51], [620.27, 162.98], [592.06, 162.98]]}]\n```" - } - ] - }, - { - "id": 1448, - "data_source": "Mono3DRefer", - "scene_name": "001053", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[425.82, 205.0], [401.51, 204.95], [417.66, 202.76], [440.33, 202.8], [425.82, 185.68], [401.51, 185.66], [417.66, 184.79], [440.33, 184.8]]}]\n```" - } - ] - }, - { - "id": 1449, - "data_source": "Mono3DRefer", - "scene_name": "001053", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[595.98, 190.34], [619.82, 190.35], [620.27, 193.56], [592.06, 193.55], [595.98, 164.51], [619.82, 164.51], [620.27, 162.98], [592.06, 162.98]]}, {\"category\": \"car\", \"corners_3d\": [[425.82, 205.0], [401.51, 204.95], [417.66, 202.76], [440.33, 202.8], [425.82, 185.68], [401.51, 185.66], [417.66, 184.79], [440.33, 184.8]]}]\n```" - } - ] - }, - { - "id": 1450, - "data_source": "Mono3DRefer", - "scene_name": "001054", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[518.56, 199.26], [537.1, 199.19], [540.37, 200.06], [520.69, 200.13], [518.56, 182.8], [537.1, 182.81], [540.37, 182.66], [520.69, 182.65]]}, {\"category\": \"car\", \"corners_3d\": [[557.88, 235.96], [618.79, 235.37], [643.97, 246.61], [569.57, 247.51], [557.88, 179.92], [618.79, 179.98], [643.97, 178.81], [569.57, 178.72]]}, {\"category\": \"car\", \"corners_3d\": [[79.26, 300.31], [203.84, 297.77], [28.41, 363.02], [-177.3, 369.43], [79.26, 201.79], [203.84, 201.43], [28.41, 210.83], [-177.3, 211.75]]}, {\"category\": \"car\", \"corners_3d\": [[257.97, 252.72], [331.06, 251.88], [275.24, 272.25], [178.52, 273.69], [257.97, 192.23], [331.06, 192.14], [275.24, 194.26], [178.52, 194.41]]}, {\"category\": \"car\", \"corners_3d\": [[466.03, 222.84], [510.64, 222.46], [514.96, 228.58], [462.98, 229.1], [466.03, 181.24], [510.64, 181.28], [514.96, 180.63], [462.98, 180.57]]}, {\"category\": \"car\", \"corners_3d\": [[327.98, 225.84], [378.12, 225.44], [352.15, 234.07], [290.81, 234.66], [327.98, 185.18], [378.12, 185.18], [352.15, 185.18], [290.81, 185.18]]}, {\"category\": \"car\", \"corners_3d\": [[237.13, 225.43], [287.61, 225.04], [247.81, 232.53], [187.26, 233.09], [237.13, 185.47], [287.61, 185.46], [247.81, 185.51], [187.26, 185.51]]}, {\"category\": \"car\", \"corners_3d\": [[386.8, 211.4], [417.49, 211.26], [405.26, 214.57], [370.58, 214.75], [386.8, 183.63], [417.49, 183.64], [405.26, 183.44], [370.58, 183.43]]}, {\"category\": \"car\", \"corners_3d\": [[406.66, 206.1], [432.92, 205.99], [425.78, 208.01], [396.91, 208.14], [406.66, 183.46], [432.92, 183.46], [425.78, 183.29], [396.91, 183.28]]}]\n```" - } - ] - }, - { - "id": 1451, - "data_source": "Mono3DRefer", - "scene_name": "001054", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[60.36, 250.26], [147.46, 249.07], [5.97, 275.89], [-122.45, 278.31], [60.36, 169.75], [147.46, 170.03], [5.97, 163.66], [-122.45, 163.08]]}, {\"category\": \"van\", \"corners_3d\": [[284.93, 212.62], [325.43, 212.4], [299.89, 216.4], [253.14, 216.69], [284.93, 171.89], [325.43, 171.99], [299.89, 170.05], [253.14, 169.91]]}, {\"category\": \"van\", \"corners_3d\": [[428.74, 199.65], [451.14, 199.57], [448.21, 201.2], [423.23, 201.3], [428.74, 170.07], [451.14, 170.15], [448.21, 168.44], [423.23, 168.34]]}]\n```" - } - ] - }, - { - "id": 1452, - "data_source": "Mono3DRefer", - "scene_name": "001054", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[518.56, 199.26], [537.1, 199.19], [540.37, 200.06], [520.69, 200.13], [518.56, 182.8], [537.1, 182.81], [540.37, 182.66], [520.69, 182.65]]}, {\"category\": \"car\", \"corners_3d\": [[557.88, 235.96], [618.79, 235.37], [643.97, 246.61], [569.57, 247.51], [557.88, 179.92], [618.79, 179.98], [643.97, 178.81], [569.57, 178.72]]}, {\"category\": \"car\", \"corners_3d\": [[79.26, 300.31], [203.84, 297.77], [28.41, 363.02], [-177.3, 369.43], [79.26, 201.79], [203.84, 201.43], [28.41, 210.83], [-177.3, 211.75]]}, {\"category\": \"van\", \"corners_3d\": [[60.36, 250.26], [147.46, 249.07], [5.97, 275.89], [-122.45, 278.31], [60.36, 169.75], [147.46, 170.03], [5.97, 163.66], [-122.45, 163.08]]}, {\"category\": \"car\", \"corners_3d\": [[257.97, 252.72], [331.06, 251.88], [275.24, 272.25], [178.52, 273.69], [257.97, 192.23], [331.06, 192.14], [275.24, 194.26], [178.52, 194.41]]}, {\"category\": \"car\", \"corners_3d\": [[466.03, 222.84], [510.64, 222.46], [514.96, 228.58], [462.98, 229.1], [466.03, 181.24], [510.64, 181.28], [514.96, 180.63], [462.98, 180.57]]}, {\"category\": \"car\", \"corners_3d\": [[327.98, 225.84], [378.12, 225.44], [352.15, 234.07], [290.81, 234.66], [327.98, 185.18], [378.12, 185.18], [352.15, 185.18], [290.81, 185.18]]}, {\"category\": \"car\", \"corners_3d\": [[237.13, 225.43], [287.61, 225.04], [247.81, 232.53], [187.26, 233.09], [237.13, 185.47], [287.61, 185.46], [247.81, 185.51], [187.26, 185.51]]}, {\"category\": \"car\", \"corners_3d\": [[386.8, 211.4], [417.49, 211.26], [405.26, 214.57], [370.58, 214.75], [386.8, 183.63], [417.49, 183.64], [405.26, 183.44], [370.58, 183.43]]}, {\"category\": \"van\", \"corners_3d\": [[284.93, 212.62], [325.43, 212.4], [299.89, 216.4], [253.14, 216.69], [284.93, 171.89], [325.43, 171.99], [299.89, 170.05], [253.14, 169.91]]}, {\"category\": \"car\", \"corners_3d\": [[406.66, 206.1], [432.92, 205.99], [425.78, 208.01], [396.91, 208.14], [406.66, 183.46], [432.92, 183.46], [425.78, 183.29], [396.91, 183.28]]}, {\"category\": \"van\", \"corners_3d\": [[428.74, 199.65], [451.14, 199.57], [448.21, 201.2], [423.23, 201.3], [428.74, 170.07], [451.14, 170.15], [448.21, 168.44], [423.23, 168.34]]}]\n```" - } - ] - }, - { - "id": 1453, - "data_source": "Mono3DRefer", - "scene_name": "001055", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[141.39, 449.52], [33.57, 447.85], [194.11, 374.8], [273.86, 375.69], [141.39, 162.35], [33.57, 162.41], [194.11, 165.19], [273.86, 165.15]]}]\n```" - } - ] - }, - { - "id": 1454, - "data_source": "Mono3DRefer", - "scene_name": "001055", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[737.35, 276.72], [841.52, 274.73], [999.39, 324.11], [847.88, 328.54], [737.35, 168.53], [841.52, 168.61], [999.39, 166.56], [847.88, 166.37]]}, {\"category\": \"car\", \"corners_3d\": [[600.01, 208.32], [643.37, 207.87], [672.08, 213.24], [622.39, 213.84], [600.01, 167.9], [643.37, 167.96], [672.08, 167.21], [622.39, 167.13]]}, {\"category\": \"car\", \"corners_3d\": [[557.12, 189.95], [542.12, 189.46], [598.54, 188.5], [614.22, 188.93], [557.12, 159.28], [542.12, 159.67], [598.54, 160.44], [614.22, 160.09]]}, {\"category\": \"car\", \"corners_3d\": [[714.8, 190.83], [741.21, 191.46], [690.54, 192.89], [663.98, 192.17], [714.8, 160.33], [741.21, 159.9], [690.54, 158.9], [663.98, 159.4]]}, {\"category\": \"car\", \"corners_3d\": [[768.95, 186.96], [793.13, 187.37], [748.6, 188.49], [723.93, 188.02], [768.95, 159.82], [793.13, 159.44], [748.6, 158.41], [723.93, 158.84]]}, {\"category\": \"car\", \"corners_3d\": [[819.52, 184.11], [838.9, 184.38], [806.38, 184.97], [786.82, 184.68], [819.52, 160.56], [838.9, 160.27], [806.38, 159.62], [786.82, 159.94]]}, {\"category\": \"car\", \"corners_3d\": [[860.02, 184.59], [885.74, 184.91], [850.23, 185.8], [823.69, 185.43], [860.02, 157.98], [885.74, 157.57], [850.23, 156.45], [823.69, 156.92]]}]\n```" - } - ] - }, - { - "id": 1455, - "data_source": "Mono3DRefer", - "scene_name": "001055", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[737.35, 276.72], [841.52, 274.73], [999.39, 324.11], [847.88, 328.54], [737.35, 168.53], [841.52, 168.61], [999.39, 166.56], [847.88, 166.37]]}, {\"category\": \"car\", \"corners_3d\": [[600.01, 208.32], [643.37, 207.87], [672.08, 213.24], [622.39, 213.84], [600.01, 167.9], [643.37, 167.96], [672.08, 167.21], [622.39, 167.13]]}, {\"category\": \"car\", \"corners_3d\": [[557.12, 189.95], [542.12, 189.46], [598.54, 188.5], [614.22, 188.93], [557.12, 159.28], [542.12, 159.67], [598.54, 160.44], [614.22, 160.09]]}, {\"category\": \"cyclist\", \"corners_3d\": [[141.39, 449.52], [33.57, 447.85], [194.11, 374.8], [273.86, 375.69], [141.39, 162.35], [33.57, 162.41], [194.11, 165.19], [273.86, 165.15]]}, {\"category\": \"car\", \"corners_3d\": [[714.8, 190.83], [741.21, 191.46], [690.54, 192.89], [663.98, 192.17], [714.8, 160.33], [741.21, 159.9], [690.54, 158.9], [663.98, 159.4]]}, {\"category\": \"car\", \"corners_3d\": [[768.95, 186.96], [793.13, 187.37], [748.6, 188.49], [723.93, 188.02], [768.95, 159.82], [793.13, 159.44], [748.6, 158.41], [723.93, 158.84]]}, {\"category\": \"car\", \"corners_3d\": [[819.52, 184.11], [838.9, 184.38], [806.38, 184.97], [786.82, 184.68], [819.52, 160.56], [838.9, 160.27], [806.38, 159.62], [786.82, 159.94]]}, {\"category\": \"car\", \"corners_3d\": [[860.02, 184.59], [885.74, 184.91], [850.23, 185.8], [823.69, 185.43], [860.02, 157.98], [885.74, 157.57], [850.23, 156.45], [823.69, 156.92]]}]\n```" - } - ] - }, - { - "id": 1456, - "data_source": "Mono3DRefer", - "scene_name": "001056", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[871.7, 244.09], [981.19, 243.77], [1340.11, 308.58], [1132.78, 309.75], [871.7, 89.86], [981.19, 90.23], [1340.11, 14.72], [1132.78, 13.36]]}]\n```" - } - ] - }, - { - "id": 1457, - "data_source": "Mono3DRefer", - "scene_name": "001056", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[632.23, 194.68], [651.59, 194.66], [656.02, 196.03], [635.45, 196.06], [632.23, 175.66], [651.59, 175.66], [656.02, 175.84], [635.45, 175.84]]}, {\"category\": \"car\", \"corners_3d\": [[671.97, 222.13], [716.53, 222.04], [737.02, 229.91], [685.36, 230.03], [671.97, 180.66], [716.53, 180.64], [737.02, 181.89], [685.36, 181.91]]}, {\"category\": \"car\", \"corners_3d\": [[777.96, 275.61], [877.73, 275.19], [1023.0, 326.59], [873.7, 327.54], [777.96, 182.18], [877.73, 182.14], [1023.0, 186.8], [873.7, 186.89]]}, {\"category\": \"car\", \"corners_3d\": [[734.49, 205.36], [765.36, 205.25], [791.73, 209.26], [757.13, 209.4], [734.49, 176.28], [765.36, 176.27], [791.73, 176.69], [757.13, 176.71]]}]\n```" - } - ] - }, - { - "id": 1458, - "data_source": "Mono3DRefer", - "scene_name": "001056", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[871.7, 244.09], [981.19, 243.77], [1340.11, 308.58], [1132.78, 309.75], [871.7, 89.86], [981.19, 90.23], [1340.11, 14.72], [1132.78, 13.36]]}, {\"category\": \"car\", \"corners_3d\": [[632.23, 194.68], [651.59, 194.66], [656.02, 196.03], [635.45, 196.06], [632.23, 175.66], [651.59, 175.66], [656.02, 175.84], [635.45, 175.84]]}, {\"category\": \"car\", \"corners_3d\": [[671.97, 222.13], [716.53, 222.04], [737.02, 229.91], [685.36, 230.03], [671.97, 180.66], [716.53, 180.64], [737.02, 181.89], [685.36, 181.91]]}, {\"category\": \"car\", \"corners_3d\": [[777.96, 275.61], [877.73, 275.19], [1023.0, 326.59], [873.7, 327.54], [777.96, 182.18], [877.73, 182.14], [1023.0, 186.8], [873.7, 186.89]]}, {\"category\": \"car\", \"corners_3d\": [[734.49, 205.36], [765.36, 205.25], [791.73, 209.26], [757.13, 209.4], [734.49, 176.28], [765.36, 176.27], [791.73, 176.69], [757.13, 176.71]]}]\n```" - } - ] - }, - { - "id": 1459, - "data_source": "Mono3DRefer", - "scene_name": "001057", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[383.3, 296.28], [370.58, 303.41], [271.75, 303.75], [290.24, 296.59], [383.3, 144.1], [370.58, 141.85], [271.75, 141.75], [290.24, 144.0]]}]\n```" - } - ] - }, - { - "id": 1460, - "data_source": "Mono3DRefer", - "scene_name": "001059", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[523.71, 204.99], [573.15, 204.8], [582.7, 215.65], [516.4, 216.0], [523.71, 144.81], [573.15, 144.98], [582.7, 135.5], [516.4, 135.2]]}]\n```" - } - ] - }, - { - "id": 1461, - "data_source": "Mono3DRefer", - "scene_name": "001059", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[74.36, 224.55], [40.94, 224.82], [82.92, 220.28], [113.25, 220.06], [74.36, 194.86], [40.94, 194.97], [82.92, 193.04], [113.25, 192.95]]}, {\"category\": \"car\", \"corners_3d\": [[130.36, 217.98], [101.98, 217.95], [151.13, 213.7], [176.87, 213.72], [130.36, 188.85], [101.98, 188.84], [151.13, 187.33], [176.87, 187.34]]}, {\"category\": \"car\", \"corners_3d\": [[190.45, 212.78], [169.54, 212.8], [202.24, 209.73], [221.53, 209.71], [190.45, 190.14], [169.54, 190.15], [202.24, 188.82], [221.53, 188.81]]}]\n```" - } - ] - }, - { - "id": 1462, - "data_source": "Mono3DRefer", - "scene_name": "001059", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[523.71, 204.99], [573.15, 204.8], [582.7, 215.65], [516.4, 216.0], [523.71, 144.81], [573.15, 144.98], [582.7, 135.5], [516.4, 135.2]]}, {\"category\": \"car\", \"corners_3d\": [[74.36, 224.55], [40.94, 224.82], [82.92, 220.28], [113.25, 220.06], [74.36, 194.86], [40.94, 194.97], [82.92, 193.04], [113.25, 192.95]]}, {\"category\": \"car\", \"corners_3d\": [[130.36, 217.98], [101.98, 217.95], [151.13, 213.7], [176.87, 213.72], [130.36, 188.85], [101.98, 188.84], [151.13, 187.33], [176.87, 187.34]]}, {\"category\": \"car\", \"corners_3d\": [[190.45, 212.78], [169.54, 212.8], [202.24, 209.73], [221.53, 209.71], [190.45, 190.14], [169.54, 190.15], [202.24, 188.82], [221.53, 188.81]]}]\n```" - } - ] - }, - { - "id": 1463, - "data_source": "Mono3DRefer", - "scene_name": "001061", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[748.46, 203.67], [775.94, 203.98], [772.73, 206.93], [742.71, 206.56], [748.46, 181.63], [775.94, 181.72], [772.73, 182.56], [742.71, 182.46]]}, {\"category\": \"car\", \"corners_3d\": [[684.57, 207.09], [655.28, 206.86], [662.04, 203.89], [688.8, 204.09], [684.57, 180.67], [655.28, 180.61], [662.04, 179.94], [688.8, 179.98]]}]\n```" - } - ] - }, - { - "id": 1464, - "data_source": "Mono3DRefer", - "scene_name": "001061", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1200.12, 247.33], [1228.43, 250.62], [1139.23, 251.33], [1114.48, 247.99], [1200.12, 161.98], [1228.43, 161.5], [1139.23, 161.4], [1114.48, 161.89]]}]\n```" - } - ] - }, - { - "id": 1465, - "data_source": "Mono3DRefer", - "scene_name": "001061", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[544.41, 211.58], [545.4, 211.18], [561.05, 211.21], [560.23, 211.61], [544.41, 180.78], [545.4, 180.7], [561.05, 180.71], [560.23, 180.79]]}]\n```" - } - ] - }, - { - "id": 1466, - "data_source": "Mono3DRefer", - "scene_name": "001061", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1200.12, 247.33], [1228.43, 250.62], [1139.23, 251.33], [1114.48, 247.99], [1200.12, 161.98], [1228.43, 161.5], [1139.23, 161.4], [1114.48, 161.89]]}, {\"category\": \"car\", \"corners_3d\": [[748.46, 203.67], [775.94, 203.98], [772.73, 206.93], [742.71, 206.56], [748.46, 181.63], [775.94, 181.72], [772.73, 182.56], [742.71, 182.46]]}, {\"category\": \"car\", \"corners_3d\": [[684.57, 207.09], [655.28, 206.86], [662.04, 203.89], [688.8, 204.09], [684.57, 180.67], [655.28, 180.61], [662.04, 179.94], [688.8, 179.98]]}]\n```" - } - ] - }, - { - "id": 1467, - "data_source": "Mono3DRefer", - "scene_name": "001061", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[748.46, 203.67], [775.94, 203.98], [772.73, 206.93], [742.71, 206.56], [748.46, 181.63], [775.94, 181.72], [772.73, 182.56], [742.71, 182.46]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[544.41, 211.58], [545.4, 211.18], [561.05, 211.21], [560.23, 211.61], [544.41, 180.78], [545.4, 180.7], [561.05, 180.71], [560.23, 180.79]]}, {\"category\": \"car\", \"corners_3d\": [[684.57, 207.09], [655.28, 206.86], [662.04, 203.89], [688.8, 204.09], [684.57, 180.67], [655.28, 180.61], [662.04, 179.94], [688.8, 179.98]]}]\n```" - } - ] - }, - { - "id": 1468, - "data_source": "Mono3DRefer", - "scene_name": "001061", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1200.12, 247.33], [1228.43, 250.62], [1139.23, 251.33], [1114.48, 247.99], [1200.12, 161.98], [1228.43, 161.5], [1139.23, 161.4], [1114.48, 161.89]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[544.41, 211.58], [545.4, 211.18], [561.05, 211.21], [560.23, 211.61], [544.41, 180.78], [545.4, 180.7], [561.05, 180.71], [560.23, 180.79]]}]\n```" - } - ] - }, - { - "id": 1469, - "data_source": "Mono3DRefer", - "scene_name": "001061", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1200.12, 247.33], [1228.43, 250.62], [1139.23, 251.33], [1114.48, 247.99], [1200.12, 161.98], [1228.43, 161.5], [1139.23, 161.4], [1114.48, 161.89]]}, {\"category\": \"car\", \"corners_3d\": [[748.46, 203.67], [775.94, 203.98], [772.73, 206.93], [742.71, 206.56], [748.46, 181.63], [775.94, 181.72], [772.73, 182.56], [742.71, 182.46]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[544.41, 211.58], [545.4, 211.18], [561.05, 211.21], [560.23, 211.61], [544.41, 180.78], [545.4, 180.7], [561.05, 180.71], [560.23, 180.79]]}, {\"category\": \"car\", \"corners_3d\": [[684.57, 207.09], [655.28, 206.86], [662.04, 203.89], [688.8, 204.09], [684.57, 180.67], [655.28, 180.61], [662.04, 179.94], [688.8, 179.98]]}]\n```" - } - ] - }, - { - "id": 1470, - "data_source": "Mono3DRefer", - "scene_name": "001063", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[200.96, 245.24], [136.18, 245.25], [217.57, 232.79], [271.19, 232.78], [200.96, 188.85], [136.18, 188.85], [217.57, 186.1], [271.19, 186.1]]}, {\"category\": \"car\", \"corners_3d\": [[324.35, 209.41], [296.66, 209.36], [326.17, 206.2], [351.51, 206.25], [324.35, 183.54], [296.66, 183.52], [326.17, 182.6], [351.51, 182.61]]}, {\"category\": \"car\", \"corners_3d\": [[429.68, 200.35], [407.32, 200.32], [419.55, 198.82], [440.69, 198.84], [429.68, 180.69], [407.32, 180.68], [419.55, 180.25], [440.69, 180.26]]}]\n```" - } - ] - }, - { - "id": 1471, - "data_source": "Mono3DRefer", - "scene_name": "001064", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[746.45, 198.15], [763.11, 197.82], [800.53, 199.58], [783.21, 199.97], [746.45, 175.18], [763.11, 175.15], [800.53, 175.31], [783.21, 175.34]]}]\n```" - } - ] - }, - { - "id": 1472, - "data_source": "Mono3DRefer", - "scene_name": "001064", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[660.23, 191.13], [680.28, 191.1], [688.08, 192.44], [666.57, 192.47], [660.23, 168.17], [680.28, 168.18], [688.08, 167.83], [666.57, 167.83]]}]\n```" - } - ] - }, - { - "id": 1473, - "data_source": "Mono3DRefer", - "scene_name": "001064", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[660.23, 191.13], [680.28, 191.1], [688.08, 192.44], [666.57, 192.47], [660.23, 168.17], [680.28, 168.18], [688.08, 167.83], [666.57, 167.83]]}, {\"category\": \"car\", \"corners_3d\": [[746.45, 198.15], [763.11, 197.82], [800.53, 199.58], [783.21, 199.97], [746.45, 175.18], [763.11, 175.15], [800.53, 175.31], [783.21, 175.34]]}]\n```" - } - ] - }, - { - "id": 1474, - "data_source": "Mono3DRefer", - "scene_name": "001072", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[656.42, 206.05], [685.28, 206.0], [695.38, 209.22], [663.73, 209.28], [656.42, 177.65], [685.28, 177.64], [695.38, 178.1], [663.73, 178.11]]}, {\"category\": \"car\", \"corners_3d\": [[751.59, 270.69], [839.32, 270.22], [936.24, 306.79], [815.99, 307.68], [751.59, 188.35], [839.32, 188.27], [936.24, 194.06], [815.99, 194.2]]}, {\"category\": \"car\", \"corners_3d\": [[748.59, 208.27], [780.43, 208.21], [801.99, 212.04], [766.73, 212.11], [748.59, 177.19], [780.43, 177.19], [801.99, 177.65], [766.73, 177.66]]}, {\"category\": \"car\", \"corners_3d\": [[838.71, 229.3], [891.23, 228.96], [970.49, 242.08], [906.19, 242.59], [838.71, 179.11], [891.23, 179.08], [970.49, 180.53], [906.19, 180.59]]}]\n```" - } - ] - }, - { - "id": 1475, - "data_source": "Mono3DRefer", - "scene_name": "001073", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[424.37, 206.76], [431.63, 205.36], [498.58, 205.33], [494.19, 206.72], [424.37, 178.69], [431.63, 178.45], [498.58, 178.44], [494.19, 178.68]]}]\n```" - } - ] - }, - { - "id": 1476, - "data_source": "Mono3DRefer", - "scene_name": "001074", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[746.9, 261.16], [844.48, 260.93], [934.19, 292.7], [801.62, 293.12], [746.9, 165.86], [844.48, 165.88], [934.19, 163.36], [801.62, 163.33]]}]\n```" - } - ] - }, - { - "id": 1477, - "data_source": "Mono3DRefer", - "scene_name": "001081", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[726.77, 205.98], [763.55, 205.93], [788.7, 210.69], [746.67, 210.75], [726.77, 163.85], [763.55, 163.86], [788.7, 162.57], [746.67, 162.55]]}]\n```" - } - ] - }, - { - "id": 1478, - "data_source": "Mono3DRefer", - "scene_name": "001081", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[668.76, 191.6], [687.48, 191.57], [696.28, 192.94], [676.21, 192.97], [668.76, 174.61], [687.48, 174.6], [696.28, 174.73], [676.21, 174.74]]}, {\"category\": \"car\", \"corners_3d\": [[799.8, 200.76], [815.47, 200.27], [865.72, 202.27], [849.85, 202.84], [799.8, 175.42], [815.47, 175.37], [865.72, 175.56], [849.85, 175.61]]}]\n```" - } - ] - }, - { - "id": 1479, - "data_source": "Mono3DRefer", - "scene_name": "001081", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[668.76, 191.6], [687.48, 191.57], [696.28, 192.94], [676.21, 192.97], [668.76, 174.61], [687.48, 174.6], [696.28, 174.73], [676.21, 174.74]]}, {\"category\": \"van\", \"corners_3d\": [[726.77, 205.98], [763.55, 205.93], [788.7, 210.69], [746.67, 210.75], [726.77, 163.85], [763.55, 163.86], [788.7, 162.57], [746.67, 162.55]]}, {\"category\": \"car\", \"corners_3d\": [[799.8, 200.76], [815.47, 200.27], [865.72, 202.27], [849.85, 202.84], [799.8, 175.42], [815.47, 175.37], [865.72, 175.56], [849.85, 175.61]]}]\n```" - } - ] - }, - { - "id": 1480, - "data_source": "Mono3DRefer", - "scene_name": "001086", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[567.78, 214.87], [537.19, 214.89], [543.67, 210.67], [571.19, 210.66], [567.78, 186.67], [537.19, 186.68], [543.67, 185.29], [571.19, 185.29]]}, {\"category\": \"car\", \"corners_3d\": [[2170.53, 887.28], [1563.81, 907.1], [906.52, 412.87], [1114.43, 410.71], [2170.53, 220.05], [1563.81, 221.36], [906.52, 188.71], [1114.43, 188.57]]}, {\"category\": \"car\", \"corners_3d\": [[519.93, 263.71], [445.51, 264.09], [469.2, 248.03], [530.49, 247.77], [519.93, 196.36], [445.51, 196.46], [469.2, 192.31], [530.49, 192.24]]}, {\"category\": \"car\", \"corners_3d\": [[671.0, 239.65], [733.3, 239.6], [761.92, 254.25], [685.95, 254.33], [671.0, 179.45], [733.3, 179.44], [761.92, 180.89], [685.95, 180.9]]}, {\"category\": \"car\", \"corners_3d\": [[660.46, 229.38], [705.74, 229.31], [723.82, 238.59], [671.11, 238.68], [660.46, 188.18], [705.74, 188.16], [723.82, 190.68], [671.11, 190.7]]}]\n```" - } - ] - }, - { - "id": 1481, - "data_source": "Mono3DRefer", - "scene_name": "001086", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[795.54, 253.91], [812.03, 253.91], [829.58, 260.96], [811.66, 260.96], [795.54, 187.95], [812.03, 187.95], [829.58, 189.27], [811.66, 189.27]]}]\n```" - } - ] - }, - { - "id": 1482, - "data_source": "Mono3DRefer", - "scene_name": "001086", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[639.15, 209.73], [667.79, 209.69], [677.01, 213.97], [645.04, 214.02], [639.15, 177.09], [667.79, 177.09], [677.01, 177.58], [645.04, 177.58]]}, {\"category\": \"van\", \"corners_3d\": [[647.98, 220.1], [688.21, 219.99], [705.97, 227.82], [659.09, 227.96], [647.98, 177.72], [688.21, 177.71], [705.97, 178.51], [659.09, 178.53]]}]\n```" - } - ] - }, - { - "id": 1483, - "data_source": "Mono3DRefer", - "scene_name": "001086", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[567.78, 214.87], [537.19, 214.89], [543.67, 210.67], [571.19, 210.66], [567.78, 186.67], [537.19, 186.68], [543.67, 185.29], [571.19, 185.29]]}, {\"category\": \"car\", \"corners_3d\": [[2170.53, 887.28], [1563.81, 907.1], [906.52, 412.87], [1114.43, 410.71], [2170.53, 220.05], [1563.81, 221.36], [906.52, 188.71], [1114.43, 188.57]]}, {\"category\": \"cyclist\", \"corners_3d\": [[795.54, 253.91], [812.03, 253.91], [829.58, 260.96], [811.66, 260.96], [795.54, 187.95], [812.03, 187.95], [829.58, 189.27], [811.66, 189.27]]}, {\"category\": \"car\", \"corners_3d\": [[519.93, 263.71], [445.51, 264.09], [469.2, 248.03], [530.49, 247.77], [519.93, 196.36], [445.51, 196.46], [469.2, 192.31], [530.49, 192.24]]}, {\"category\": \"car\", \"corners_3d\": [[671.0, 239.65], [733.3, 239.6], [761.92, 254.25], [685.95, 254.33], [671.0, 179.45], [733.3, 179.44], [761.92, 180.89], [685.95, 180.9]]}, {\"category\": \"car\", \"corners_3d\": [[660.46, 229.38], [705.74, 229.31], [723.82, 238.59], [671.11, 238.68], [660.46, 188.18], [705.74, 188.16], [723.82, 190.68], [671.11, 190.7]]}]\n```" - } - ] - }, - { - "id": 1484, - "data_source": "Mono3DRefer", - "scene_name": "001086", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[567.78, 214.87], [537.19, 214.89], [543.67, 210.67], [571.19, 210.66], [567.78, 186.67], [537.19, 186.68], [543.67, 185.29], [571.19, 185.29]]}, {\"category\": \"van\", \"corners_3d\": [[639.15, 209.73], [667.79, 209.69], [677.01, 213.97], [645.04, 214.02], [639.15, 177.09], [667.79, 177.09], [677.01, 177.58], [645.04, 177.58]]}, {\"category\": \"car\", \"corners_3d\": [[2170.53, 887.28], [1563.81, 907.1], [906.52, 412.87], [1114.43, 410.71], [2170.53, 220.05], [1563.81, 221.36], [906.52, 188.71], [1114.43, 188.57]]}, {\"category\": \"car\", \"corners_3d\": [[519.93, 263.71], [445.51, 264.09], [469.2, 248.03], [530.49, 247.77], [519.93, 196.36], [445.51, 196.46], [469.2, 192.31], [530.49, 192.24]]}, {\"category\": \"car\", \"corners_3d\": [[671.0, 239.65], [733.3, 239.6], [761.92, 254.25], [685.95, 254.33], [671.0, 179.45], [733.3, 179.44], [761.92, 180.89], [685.95, 180.9]]}, {\"category\": \"car\", \"corners_3d\": [[660.46, 229.38], [705.74, 229.31], [723.82, 238.59], [671.11, 238.68], [660.46, 188.18], [705.74, 188.16], [723.82, 190.68], [671.11, 190.7]]}, {\"category\": \"van\", \"corners_3d\": [[647.98, 220.1], [688.21, 219.99], [705.97, 227.82], [659.09, 227.96], [647.98, 177.72], [688.21, 177.71], [705.97, 178.51], [659.09, 178.53]]}]\n```" - } - ] - }, - { - "id": 1485, - "data_source": "Mono3DRefer", - "scene_name": "001086", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[639.15, 209.73], [667.79, 209.69], [677.01, 213.97], [645.04, 214.02], [639.15, 177.09], [667.79, 177.09], [677.01, 177.58], [645.04, 177.58]]}, {\"category\": \"cyclist\", \"corners_3d\": [[795.54, 253.91], [812.03, 253.91], [829.58, 260.96], [811.66, 260.96], [795.54, 187.95], [812.03, 187.95], [829.58, 189.27], [811.66, 189.27]]}, {\"category\": \"van\", \"corners_3d\": [[647.98, 220.1], [688.21, 219.99], [705.97, 227.82], [659.09, 227.96], [647.98, 177.72], [688.21, 177.71], [705.97, 178.51], [659.09, 178.53]]}]\n```" - } - ] - }, - { - "id": 1486, - "data_source": "Mono3DRefer", - "scene_name": "001086", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[567.78, 214.87], [537.19, 214.89], [543.67, 210.67], [571.19, 210.66], [567.78, 186.67], [537.19, 186.68], [543.67, 185.29], [571.19, 185.29]]}, {\"category\": \"van\", \"corners_3d\": [[639.15, 209.73], [667.79, 209.69], [677.01, 213.97], [645.04, 214.02], [639.15, 177.09], [667.79, 177.09], [677.01, 177.58], [645.04, 177.58]]}, {\"category\": \"car\", \"corners_3d\": [[2170.53, 887.28], [1563.81, 907.1], [906.52, 412.87], [1114.43, 410.71], [2170.53, 220.05], [1563.81, 221.36], [906.52, 188.71], [1114.43, 188.57]]}, {\"category\": \"cyclist\", \"corners_3d\": [[795.54, 253.91], [812.03, 253.91], [829.58, 260.96], [811.66, 260.96], [795.54, 187.95], [812.03, 187.95], [829.58, 189.27], [811.66, 189.27]]}, {\"category\": \"car\", \"corners_3d\": [[519.93, 263.71], [445.51, 264.09], [469.2, 248.03], [530.49, 247.77], [519.93, 196.36], [445.51, 196.46], [469.2, 192.31], [530.49, 192.24]]}, {\"category\": \"car\", \"corners_3d\": [[671.0, 239.65], [733.3, 239.6], [761.92, 254.25], [685.95, 254.33], [671.0, 179.45], [733.3, 179.44], [761.92, 180.89], [685.95, 180.9]]}, {\"category\": \"car\", \"corners_3d\": [[660.46, 229.38], [705.74, 229.31], [723.82, 238.59], [671.11, 238.68], [660.46, 188.18], [705.74, 188.16], [723.82, 190.68], [671.11, 190.7]]}, {\"category\": \"van\", \"corners_3d\": [[647.98, 220.1], [688.21, 219.99], [705.97, 227.82], [659.09, 227.96], [647.98, 177.72], [688.21, 177.71], [705.97, 178.51], [659.09, 178.53]]}]\n```" - } - ] - }, - { - "id": 1487, - "data_source": "Mono3DRefer", - "scene_name": "001087", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[583.0, 205.09], [609.52, 205.01], [614.61, 208.29], [585.4, 208.39], [583.0, 180.57], [609.52, 180.55], [614.61, 181.33], [585.4, 181.35]]}, {\"category\": \"car\", \"corners_3d\": [[1357.21, 340.34], [1205.5, 340.02], [1006.38, 283.54], [1106.51, 283.68], [1357.21, 195.75], [1205.5, 195.71], [1006.38, 187.99], [1106.51, 188.01]]}, {\"category\": \"car\", \"corners_3d\": [[711.61, 230.22], [701.47, 226.05], [833.28, 225.54], [853.54, 229.63], [711.61, 178.34], [701.47, 177.94], [833.28, 177.89], [853.54, 178.28]]}, {\"category\": \"car\", \"corners_3d\": [[816.09, 220.55], [829.04, 223.54], [702.75, 223.54], [697.26, 220.55], [816.09, 178.27], [829.04, 178.61], [702.75, 178.61], [697.26, 178.27]]}, {\"category\": \"car\", \"corners_3d\": [[774.21, 213.54], [778.41, 215.74], [674.99, 215.02], [676.08, 212.89], [774.21, 180.14], [778.41, 180.54], [674.99, 180.41], [676.08, 180.03]]}]\n```" - } - ] - }, - { - "id": 1488, - "data_source": "Mono3DRefer", - "scene_name": "001091", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[141.68, 325.41], [105.43, 336.12], [-2.37, 335.14], [41.04, 324.57], [141.68, 152.66], [105.43, 150.6], [-2.37, 150.79], [41.04, 152.82]]}]\n```" - } - ] - }, - { - "id": 1489, - "data_source": "Mono3DRefer", - "scene_name": "001092", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[609.24, 195.28], [647.2, 195.35], [645.67, 200.86], [598.45, 200.75], [609.24, 149.11], [647.2, 149.03], [645.67, 143.19], [598.45, 143.31]]}]\n```" - } - ] - }, - { - "id": 1490, - "data_source": "Mono3DRefer", - "scene_name": "001092", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[443.86, 193.08], [427.78, 193.05], [439.17, 192.03], [454.44, 192.05], [443.86, 177.34], [427.78, 177.33], [439.17, 177.1], [454.44, 177.11]]}]\n```" - } - ] - }, - { - "id": 1491, - "data_source": "Mono3DRefer", - "scene_name": "001092", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[609.24, 195.28], [647.2, 195.35], [645.67, 200.86], [598.45, 200.75], [609.24, 149.11], [647.2, 149.03], [645.67, 143.19], [598.45, 143.31]]}, {\"category\": \"car\", \"corners_3d\": [[443.86, 193.08], [427.78, 193.05], [439.17, 192.03], [454.44, 192.05], [443.86, 177.34], [427.78, 177.33], [439.17, 177.1], [454.44, 177.11]]}]\n```" - } - ] - }, - { - "id": 1492, - "data_source": "Mono3DRefer", - "scene_name": "001107", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[19.9, 441.81], [281.61, 443.87], [-1301.33, 1694.79], [-2655.68, 1631.91], [19.9, 220.27], [281.61, 220.63], [-1301.33, 441.17], [-2655.68, 430.08]]}, {\"category\": \"car\", \"corners_3d\": [[727.17, 239.41], [661.19, 239.6], [648.19, 227.86], [702.62, 227.73], [727.17, 178.18], [661.19, 178.2], [648.19, 177.26], [702.62, 177.25]]}, {\"category\": \"car\", \"corners_3d\": [[417.14, 210.76], [426.48, 208.93], [514.8, 208.93], [509.93, 210.77], [417.14, 178.73], [426.48, 178.45], [514.8, 178.45], [509.93, 178.73]]}, {\"category\": \"car\", \"corners_3d\": [[1149.79, 255.33], [1211.09, 265.65], [950.94, 263.44], [919.27, 253.57], [1149.79, 162.78], [1211.09, 161.51], [950.94, 161.79], [919.27, 162.99]]}]\n```" - } - ] - }, - { - "id": 1493, - "data_source": "Mono3DRefer", - "scene_name": "001107", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[559.49, 207.18], [523.01, 207.04], [537.52, 203.59], [570.36, 203.71], [559.49, 170.04], [523.01, 170.05], [537.52, 170.33], [570.36, 170.32]]}, {\"category\": \"van\", \"corners_3d\": [[639.75, 202.62], [669.58, 202.65], [673.56, 205.79], [640.59, 205.75], [639.75, 170.29], [669.58, 170.29], [673.56, 170.02], [640.59, 170.02]]}]\n```" - } - ] - }, - { - "id": 1494, - "data_source": "Mono3DRefer", - "scene_name": "001107", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[19.9, 441.81], [281.61, 443.87], [-1301.33, 1694.79], [-2655.68, 1631.91], [19.9, 220.27], [281.61, 220.63], [-1301.33, 441.17], [-2655.68, 430.08]]}, {\"category\": \"car\", \"corners_3d\": [[727.17, 239.41], [661.19, 239.6], [648.19, 227.86], [702.62, 227.73], [727.17, 178.18], [661.19, 178.2], [648.19, 177.26], [702.62, 177.25]]}, {\"category\": \"van\", \"corners_3d\": [[559.49, 207.18], [523.01, 207.04], [537.52, 203.59], [570.36, 203.71], [559.49, 170.04], [523.01, 170.05], [537.52, 170.33], [570.36, 170.32]]}, {\"category\": \"van\", \"corners_3d\": [[639.75, 202.62], [669.58, 202.65], [673.56, 205.79], [640.59, 205.75], [639.75, 170.29], [669.58, 170.29], [673.56, 170.02], [640.59, 170.02]]}, {\"category\": \"car\", \"corners_3d\": [[417.14, 210.76], [426.48, 208.93], [514.8, 208.93], [509.93, 210.77], [417.14, 178.73], [426.48, 178.45], [514.8, 178.45], [509.93, 178.73]]}, {\"category\": \"car\", \"corners_3d\": [[1149.79, 255.33], [1211.09, 265.65], [950.94, 263.44], [919.27, 253.57], [1149.79, 162.78], [1211.09, 161.51], [950.94, 161.79], [919.27, 162.99]]}]\n```" - } - ] - }, - { - "id": 1495, - "data_source": "Mono3DRefer", - "scene_name": "001112", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1061.6, 236.6], [1084.98, 236.7], [1147.2, 245.75], [1120.39, 245.62], [1061.6, 151.23], [1084.98, 151.2], [1147.2, 148.12], [1120.39, 148.17]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1033.56, 226.91], [1053.63, 227.73], [1041.97, 232.29], [1020.45, 231.32], [1033.56, 163.44], [1053.63, 163.29], [1041.97, 162.5], [1020.45, 162.67]]}]\n```" - } - ] - }, - { - "id": 1496, - "data_source": "Mono3DRefer", - "scene_name": "001112", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[376.85, 239.68], [358.08, 245.63], [210.96, 246.07], [241.91, 240.06], [376.85, 185.06], [358.08, 186.14], [210.96, 186.23], [241.91, 185.13]]}, {\"category\": \"car\", \"corners_3d\": [[292.63, 231.46], [317.11, 227.71], [419.58, 228.46], [401.84, 232.31], [292.63, 185.67], [317.11, 184.85], [419.58, 185.01], [401.84, 185.85]]}, {\"category\": \"car\", \"corners_3d\": [[932.86, 186.03], [904.69, 185.92], [894.64, 184.81], [920.33, 184.91], [932.86, 162.61], [904.69, 162.7], [894.64, 163.56], [920.33, 163.48]]}]\n```" - } - ] - }, - { - "id": 1497, - "data_source": "Mono3DRefer", - "scene_name": "001112", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[376.85, 239.68], [358.08, 245.63], [210.96, 246.07], [241.91, 240.06], [376.85, 185.06], [358.08, 186.14], [210.96, 186.23], [241.91, 185.13]]}, {\"category\": \"car\", \"corners_3d\": [[292.63, 231.46], [317.11, 227.71], [419.58, 228.46], [401.84, 232.31], [292.63, 185.67], [317.11, 184.85], [419.58, 185.01], [401.84, 185.85]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1061.6, 236.6], [1084.98, 236.7], [1147.2, 245.75], [1120.39, 245.62], [1061.6, 151.23], [1084.98, 151.2], [1147.2, 148.12], [1120.39, 148.17]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1033.56, 226.91], [1053.63, 227.73], [1041.97, 232.29], [1020.45, 231.32], [1033.56, 163.44], [1053.63, 163.29], [1041.97, 162.5], [1020.45, 162.67]]}, {\"category\": \"car\", \"corners_3d\": [[932.86, 186.03], [904.69, 185.92], [894.64, 184.81], [920.33, 184.91], [932.86, 162.61], [904.69, 162.7], [894.64, 163.56], [920.33, 163.48]]}]\n```" - } - ] - }, - { - "id": 1498, - "data_source": "Mono3DRefer", - "scene_name": "001113", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[747.08, 310.17], [733.42, 303.31], [813.32, 300.98], [831.08, 307.57], [747.08, 153.67], [733.42, 155.09], [813.32, 155.58], [831.08, 154.21]]}]\n```" - } - ] - }, - { - "id": 1499, - "data_source": "Mono3DRefer", - "scene_name": "001119", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[31.19, 256.21], [81.68, 256.72], [-53.66, 276.07], [-114.72, 275.31], [31.19, 207.88], [81.68, 208.09], [-53.66, 216.22], [-114.72, 215.9]]}, {\"category\": \"car\", \"corners_3d\": [[-422.99, 467.65], [-675.18, 463.37], [-32.43, 321.16], [100.19, 322.27], [-422.99, 237.1], [-675.18, 236.17], [-32.43, 205.17], [100.19, 205.42]]}, {\"category\": \"car\", \"corners_3d\": [[310.28, 251.28], [365.92, 251.6], [313.69, 266.27], [247.97, 265.83], [310.28, 201.72], [365.92, 201.84], [313.69, 207.24], [247.97, 207.08]]}, {\"category\": \"car\", \"corners_3d\": [[254.28, 249.98], [304.09, 250.43], [213.93, 269.65], [152.51, 268.95], [254.28, 201.22], [304.09, 201.39], [213.93, 208.46], [152.51, 208.2]]}]\n```" - } - ] - }, - { - "id": 1500, - "data_source": "Mono3DRefer", - "scene_name": "001121", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[386.94, 215.52], [357.36, 216.25], [351.65, 213.43], [379.43, 212.8], [386.94, 189.05], [357.36, 189.33], [351.65, 188.26], [379.43, 188.02]]}, {\"category\": \"car\", \"corners_3d\": [[454.26, 275.71], [356.64, 277.22], [400.56, 250.67], [473.14, 249.83], [454.26, 193.97], [356.64, 194.28], [400.56, 188.83], [473.14, 188.65]]}]\n```" - } - ] - }, - { - "id": 1501, - "data_source": "Mono3DRefer", - "scene_name": "001122", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[630.58, 197.08], [672.36, 197.28], [661.51, 203.99], [608.47, 203.67], [630.58, 146.31], [672.36, 146.09], [661.51, 138.73], [608.47, 139.09]]}]\n```" - } - ] - }, - { - "id": 1502, - "data_source": "Mono3DRefer", - "scene_name": "001122", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[349.52, 202.46], [325.97, 202.42], [349.23, 200.26], [371.1, 200.3], [349.52, 179.3], [325.97, 179.29], [349.23, 178.82], [371.1, 178.83]]}, {\"category\": \"car\", \"corners_3d\": [[378.08, 199.73], [354.55, 199.69], [375.11, 197.79], [397.0, 197.83], [378.08, 180.8], [354.55, 180.78], [375.11, 180.22], [397.0, 180.23]]}]\n```" - } - ] - }, - { - "id": 1503, - "data_source": "Mono3DRefer", - "scene_name": "001122", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[630.58, 197.08], [672.36, 197.28], [661.51, 203.99], [608.47, 203.67], [630.58, 146.31], [672.36, 146.09], [661.51, 138.73], [608.47, 139.09]]}, {\"category\": \"car\", \"corners_3d\": [[349.52, 202.46], [325.97, 202.42], [349.23, 200.26], [371.1, 200.3], [349.52, 179.3], [325.97, 179.29], [349.23, 178.82], [371.1, 178.83]]}, {\"category\": \"car\", \"corners_3d\": [[378.08, 199.73], [354.55, 199.69], [375.11, 197.79], [397.0, 197.83], [378.08, 180.8], [354.55, 180.78], [375.11, 180.22], [397.0, 180.23]]}]\n```" - } - ] - }, - { - "id": 1504, - "data_source": "Mono3DRefer", - "scene_name": "001125", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1032.19, 334.89], [949.86, 304.45], [1291.53, 303.09], [1450.85, 332.83], [1032.19, 163.06], [949.86, 164.9], [1291.53, 164.98], [1450.85, 163.18]]}, {\"category\": \"car\", \"corners_3d\": [[-337.03, 352.06], [-160.52, 318.6], [137.81, 318.51], [29.87, 351.91], [-337.03, 193.23], [-160.52, 189.43], [137.81, 189.42], [29.87, 193.22]]}, {\"category\": \"car\", \"corners_3d\": [[144.68, 260.27], [194.5, 250.93], [400.88, 250.96], [375.76, 260.31], [144.68, 181.95], [194.5, 180.98], [400.88, 180.98], [375.76, 181.96]]}, {\"category\": \"car\", \"corners_3d\": [[288.58, 241.95], [309.11, 235.86], [471.35, 234.21], [466.65, 239.97], [288.58, 170.46], [309.11, 170.67], [471.35, 170.73], [466.65, 170.53]]}, {\"category\": \"car\", \"corners_3d\": [[423.79, 207.92], [432.58, 206.39], [499.84, 206.46], [494.1, 207.99], [423.79, 174.13], [432.58, 174.07], [499.84, 174.08], [494.1, 174.13]]}, {\"category\": \"car\", \"corners_3d\": [[526.33, 205.29], [522.07, 206.82], [441.4, 206.78], [449.28, 205.26], [526.33, 173.42], [522.07, 173.44], [441.4, 173.44], [449.28, 173.42]]}]\n```" - } - ] - }, - { - "id": 1505, - "data_source": "Mono3DRefer", - "scene_name": "001125", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[675.48, 224.5], [741.67, 224.51], [794.35, 245.19], [701.65, 245.19], [675.48, 145.83], [741.67, 145.83], [794.35, 135.0], [701.65, 135.01]]}, {\"category\": \"van\", \"corners_3d\": [[642.12, 202.79], [678.83, 202.76], [695.84, 208.89], [651.63, 208.93], [642.12, 155.29], [678.83, 155.3], [695.84, 151.71], [651.63, 151.69]]}, {\"category\": \"van\", \"corners_3d\": [[561.17, 196.2], [559.31, 196.98], [484.87, 196.96], [489.13, 196.18], [561.17, 160.68], [559.31, 160.27], [484.87, 160.28], [489.13, 160.69]]}]\n```" - } - ] - }, - { - "id": 1506, - "data_source": "Mono3DRefer", - "scene_name": "001125", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1032.19, 334.89], [949.86, 304.45], [1291.53, 303.09], [1450.85, 332.83], [1032.19, 163.06], [949.86, 164.9], [1291.53, 164.98], [1450.85, 163.18]]}, {\"category\": \"car\", \"corners_3d\": [[-337.03, 352.06], [-160.52, 318.6], [137.81, 318.51], [29.87, 351.91], [-337.03, 193.23], [-160.52, 189.43], [137.81, 189.42], [29.87, 193.22]]}, {\"category\": \"van\", \"corners_3d\": [[675.48, 224.5], [741.67, 224.51], [794.35, 245.19], [701.65, 245.19], [675.48, 145.83], [741.67, 145.83], [794.35, 135.0], [701.65, 135.01]]}, {\"category\": \"car\", \"corners_3d\": [[144.68, 260.27], [194.5, 250.93], [400.88, 250.96], [375.76, 260.31], [144.68, 181.95], [194.5, 180.98], [400.88, 180.98], [375.76, 181.96]]}, {\"category\": \"car\", \"corners_3d\": [[288.58, 241.95], [309.11, 235.86], [471.35, 234.21], [466.65, 239.97], [288.58, 170.46], [309.11, 170.67], [471.35, 170.73], [466.65, 170.53]]}, {\"category\": \"van\", \"corners_3d\": [[642.12, 202.79], [678.83, 202.76], [695.84, 208.89], [651.63, 208.93], [642.12, 155.29], [678.83, 155.3], [695.84, 151.71], [651.63, 151.69]]}, {\"category\": \"car\", \"corners_3d\": [[423.79, 207.92], [432.58, 206.39], [499.84, 206.46], [494.1, 207.99], [423.79, 174.13], [432.58, 174.07], [499.84, 174.08], [494.1, 174.13]]}, {\"category\": \"car\", \"corners_3d\": [[526.33, 205.29], [522.07, 206.82], [441.4, 206.78], [449.28, 205.26], [526.33, 173.42], [522.07, 173.44], [441.4, 173.44], [449.28, 173.42]]}, {\"category\": \"van\", \"corners_3d\": [[561.17, 196.2], [559.31, 196.98], [484.87, 196.96], [489.13, 196.18], [561.17, 160.68], [559.31, 160.27], [484.87, 160.28], [489.13, 160.69]]}]\n```" - } - ] - }, - { - "id": 1507, - "data_source": "Mono3DRefer", - "scene_name": "001126", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[494.0, 227.18], [442.2, 227.76], [450.49, 219.88], [494.85, 219.45], [494.0, 175.02], [442.2, 175.04], [450.49, 174.73], [494.85, 174.71]]}, {\"category\": \"car\", \"corners_3d\": [[487.8, 208.91], [454.58, 209.4], [447.59, 205.69], [477.57, 205.29], [487.8, 180.71], [454.58, 180.82], [447.59, 180.01], [477.57, 179.92]]}]\n```" - } - ] - }, - { - "id": 1508, - "data_source": "Mono3DRefer", - "scene_name": "001128", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[638.86, 203.69], [666.3, 203.65], [673.52, 206.51], [643.54, 206.55], [638.86, 178.2], [666.3, 178.2], [673.52, 178.69], [643.54, 178.7]]}, {\"category\": \"car\", \"corners_3d\": [[677.53, 220.07], [723.11, 219.99], [747.27, 228.44], [693.54, 228.56], [677.53, 177.65], [723.11, 177.64], [747.27, 178.5], [693.54, 178.51]]}, {\"category\": \"car\", \"corners_3d\": [[837.1, 304.93], [962.39, 304.48], [1246.71, 406.55], [1025.4, 407.95], [837.1, 169.62], [962.39, 169.63], [1246.71, 167.14], [1025.4, 167.1]]}, {\"category\": \"car\", \"corners_3d\": [[911.52, 206.89], [924.63, 206.08], [996.54, 208.67], [984.26, 209.61], [911.52, 177.43], [924.63, 177.32], [996.54, 177.67], [984.26, 177.79]]}]\n```" - } - ] - }, - { - "id": 1509, - "data_source": "Mono3DRefer", - "scene_name": "001128", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[743.83, 211.72], [804.57, 211.6], [884.68, 225.49], [802.41, 225.73], [743.83, 126.53], [804.57, 126.68], [884.68, 110.12], [802.41, 109.83]]}]\n```" - } - ] - }, - { - "id": 1510, - "data_source": "Mono3DRefer", - "scene_name": "001128", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[638.86, 203.69], [666.3, 203.65], [673.52, 206.51], [643.54, 206.55], [638.86, 178.2], [666.3, 178.2], [673.52, 178.69], [643.54, 178.7]]}, {\"category\": \"car\", \"corners_3d\": [[677.53, 220.07], [723.11, 219.99], [747.27, 228.44], [693.54, 228.56], [677.53, 177.65], [723.11, 177.64], [747.27, 178.5], [693.54, 178.51]]}, {\"category\": \"car\", \"corners_3d\": [[837.1, 304.93], [962.39, 304.48], [1246.71, 406.55], [1025.4, 407.95], [837.1, 169.62], [962.39, 169.63], [1246.71, 167.14], [1025.4, 167.1]]}, {\"category\": \"truck\", \"corners_3d\": [[743.83, 211.72], [804.57, 211.6], [884.68, 225.49], [802.41, 225.73], [743.83, 126.53], [804.57, 126.68], [884.68, 110.12], [802.41, 109.83]]}, {\"category\": \"car\", \"corners_3d\": [[911.52, 206.89], [924.63, 206.08], [996.54, 208.67], [984.26, 209.61], [911.52, 177.43], [924.63, 177.32], [996.54, 177.67], [984.26, 177.79]]}]\n```" - } - ] - }, - { - "id": 1511, - "data_source": "Mono3DRefer", - "scene_name": "001134", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[738.28, 245.16], [760.93, 245.39], [770.61, 254.35], [745.14, 254.06], [738.28, 178.31], [760.93, 178.32], [770.61, 179.0], [745.14, 178.98]]}, {\"category\": \"cyclist\", \"corners_3d\": [[747.84, 262.04], [784.41, 262.57], [797.44, 275.97], [755.36, 275.26], [747.84, 169.72], [784.41, 169.7], [797.44, 169.23], [755.36, 169.26]]}]\n```" - } - ] - }, - { - "id": 1512, - "data_source": "Mono3DRefer", - "scene_name": "001134", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[925.29, 194.45], [933.79, 195.49], [835.7, 194.94], [831.79, 193.95], [925.29, 151.41], [933.79, 150.38], [835.7, 150.92], [831.79, 151.9]]}]\n```" - } - ] - }, - { - "id": 1513, - "data_source": "Mono3DRefer", - "scene_name": "001134", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[503.08, 267.08], [422.49, 265.68], [478.6, 247.13], [543.57, 248.02], [503.08, 195.13], [422.49, 194.8], [478.6, 190.41], [543.57, 190.62]]}, {\"category\": \"car\", \"corners_3d\": [[647.73, 208.99], [617.68, 208.5], [637.54, 205.43], [665.21, 205.84], [647.73, 181.31], [617.68, 181.19], [637.54, 180.47], [665.21, 180.57]]}, {\"category\": \"car\", \"corners_3d\": [[771.67, 197.61], [795.84, 197.81], [795.95, 199.77], [769.91, 199.54], [771.67, 177.77], [795.84, 177.81], [795.95, 178.2], [769.91, 178.16]]}]\n```" - } - ] - }, - { - "id": 1514, - "data_source": "Mono3DRefer", - "scene_name": "001134", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[738.28, 245.16], [760.93, 245.39], [770.61, 254.35], [745.14, 254.06], [738.28, 178.31], [760.93, 178.32], [770.61, 179.0], [745.14, 178.98]]}, {\"category\": \"van\", \"corners_3d\": [[925.29, 194.45], [933.79, 195.49], [835.7, 194.94], [831.79, 193.95], [925.29, 151.41], [933.79, 150.38], [835.7, 150.92], [831.79, 151.9]]}, {\"category\": \"cyclist\", \"corners_3d\": [[747.84, 262.04], [784.41, 262.57], [797.44, 275.97], [755.36, 275.26], [747.84, 169.72], [784.41, 169.7], [797.44, 169.23], [755.36, 169.26]]}]\n```" - } - ] - }, - { - "id": 1515, - "data_source": "Mono3DRefer", - "scene_name": "001134", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[503.08, 267.08], [422.49, 265.68], [478.6, 247.13], [543.57, 248.02], [503.08, 195.13], [422.49, 194.8], [478.6, 190.41], [543.57, 190.62]]}, {\"category\": \"cyclist\", \"corners_3d\": [[738.28, 245.16], [760.93, 245.39], [770.61, 254.35], [745.14, 254.06], [738.28, 178.31], [760.93, 178.32], [770.61, 179.0], [745.14, 178.98]]}, {\"category\": \"car\", \"corners_3d\": [[647.73, 208.99], [617.68, 208.5], [637.54, 205.43], [665.21, 205.84], [647.73, 181.31], [617.68, 181.19], [637.54, 180.47], [665.21, 180.57]]}, {\"category\": \"car\", \"corners_3d\": [[771.67, 197.61], [795.84, 197.81], [795.95, 199.77], [769.91, 199.54], [771.67, 177.77], [795.84, 177.81], [795.95, 178.2], [769.91, 178.16]]}, {\"category\": \"cyclist\", \"corners_3d\": [[747.84, 262.04], [784.41, 262.57], [797.44, 275.97], [755.36, 275.26], [747.84, 169.72], [784.41, 169.7], [797.44, 169.23], [755.36, 169.26]]}]\n```" - } - ] - }, - { - "id": 1516, - "data_source": "Mono3DRefer", - "scene_name": "001134", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[503.08, 267.08], [422.49, 265.68], [478.6, 247.13], [543.57, 248.02], [503.08, 195.13], [422.49, 194.8], [478.6, 190.41], [543.57, 190.62]]}, {\"category\": \"car\", \"corners_3d\": [[647.73, 208.99], [617.68, 208.5], [637.54, 205.43], [665.21, 205.84], [647.73, 181.31], [617.68, 181.19], [637.54, 180.47], [665.21, 180.57]]}, {\"category\": \"van\", \"corners_3d\": [[925.29, 194.45], [933.79, 195.49], [835.7, 194.94], [831.79, 193.95], [925.29, 151.41], [933.79, 150.38], [835.7, 150.92], [831.79, 151.9]]}, {\"category\": \"car\", \"corners_3d\": [[771.67, 197.61], [795.84, 197.81], [795.95, 199.77], [769.91, 199.54], [771.67, 177.77], [795.84, 177.81], [795.95, 178.2], [769.91, 178.16]]}]\n```" - } - ] - }, - { - "id": 1517, - "data_source": "Mono3DRefer", - "scene_name": "001134", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[503.08, 267.08], [422.49, 265.68], [478.6, 247.13], [543.57, 248.02], [503.08, 195.13], [422.49, 194.8], [478.6, 190.41], [543.57, 190.62]]}, {\"category\": \"cyclist\", \"corners_3d\": [[738.28, 245.16], [760.93, 245.39], [770.61, 254.35], [745.14, 254.06], [738.28, 178.31], [760.93, 178.32], [770.61, 179.0], [745.14, 178.98]]}, {\"category\": \"car\", \"corners_3d\": [[647.73, 208.99], [617.68, 208.5], [637.54, 205.43], [665.21, 205.84], [647.73, 181.31], [617.68, 181.19], [637.54, 180.47], [665.21, 180.57]]}, {\"category\": \"van\", \"corners_3d\": [[925.29, 194.45], [933.79, 195.49], [835.7, 194.94], [831.79, 193.95], [925.29, 151.41], [933.79, 150.38], [835.7, 150.92], [831.79, 151.9]]}, {\"category\": \"car\", \"corners_3d\": [[771.67, 197.61], [795.84, 197.81], [795.95, 199.77], [769.91, 199.54], [771.67, 177.77], [795.84, 177.81], [795.95, 178.2], [769.91, 178.16]]}, {\"category\": \"cyclist\", \"corners_3d\": [[747.84, 262.04], [784.41, 262.57], [797.44, 275.97], [755.36, 275.26], [747.84, 169.72], [784.41, 169.7], [797.44, 169.23], [755.36, 169.26]]}]\n```" - } - ] - }, - { - "id": 1518, - "data_source": "Mono3DRefer", - "scene_name": "001135", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[688.93, 249.8], [755.69, 249.59], [813.26, 276.02], [723.64, 276.4], [688.93, 189.85], [755.69, 189.8], [813.26, 195.64], [723.64, 195.72]]}]\n```" - } - ] - }, - { - "id": 1519, - "data_source": "Mono3DRefer", - "scene_name": "001137", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[817.98, 222.42], [873.54, 223.33], [888.65, 232.97], [822.39, 231.68], [817.98, 172.84], [873.54, 172.84], [888.65, 172.84], [822.39, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[226.41, 309.21], [120.77, 302.45], [342.69, 261.11], [421.2, 264.19], [226.41, 193.24], [120.77, 192.23], [342.69, 186.05], [421.2, 186.51]]}]\n```" - } - ] - }, - { - "id": 1520, - "data_source": "Mono3DRefer", - "scene_name": "001147", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[478.53, 294.1], [365.25, 294.68], [419.71, 264.81], [505.11, 264.47], [478.53, 191.76], [365.25, 191.85], [419.71, 187.19], [505.11, 187.14]]}, {\"category\": \"car\", \"corners_3d\": [[699.01, 260.56], [786.02, 260.46], [847.51, 289.88], [731.32, 290.06], [699.01, 176.43], [786.02, 176.42], [847.51, 177.62], [731.32, 177.63]]}, {\"category\": \"car\", \"corners_3d\": [[678.29, 235.62], [736.96, 235.53], [768.58, 249.55], [696.81, 249.7], [678.29, 182.2], [736.96, 182.18], [768.58, 184.27], [696.81, 184.29]]}, {\"category\": \"car\", \"corners_3d\": [[565.02, 210.56], [529.55, 210.59], [536.99, 206.28], [568.39, 206.25], [565.02, 177.9], [529.55, 177.91], [536.99, 177.33], [568.39, 177.33]]}]\n```" - } - ] - }, - { - "id": 1521, - "data_source": "Mono3DRefer", - "scene_name": "001147", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[845.07, 265.39], [865.08, 265.36], [893.35, 275.34], [871.19, 275.37], [845.07, 185.1], [865.08, 185.1], [893.35, 186.42], [871.19, 186.42]]}]\n```" - } - ] - }, - { - "id": 1522, - "data_source": "Mono3DRefer", - "scene_name": "001147", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[660.36, 223.69], [709.79, 223.59], [735.51, 234.35], [675.64, 234.5], [660.36, 171.65], [709.79, 171.65], [735.51, 171.39], [675.64, 171.39]]}, {\"category\": \"van\", \"corners_3d\": [[645.11, 205.93], [677.72, 205.89], [689.72, 210.33], [652.74, 210.39], [645.11, 168.76], [677.72, 168.76], [689.72, 168.21], [652.74, 168.21]]}]\n```" - } - ] - }, - { - "id": 1523, - "data_source": "Mono3DRefer", - "scene_name": "001147", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[845.07, 265.39], [865.08, 265.36], [893.35, 275.34], [871.19, 275.37], [845.07, 185.1], [865.08, 185.1], [893.35, 186.42], [871.19, 186.42]]}, {\"category\": \"car\", \"corners_3d\": [[478.53, 294.1], [365.25, 294.68], [419.71, 264.81], [505.11, 264.47], [478.53, 191.76], [365.25, 191.85], [419.71, 187.19], [505.11, 187.14]]}, {\"category\": \"car\", \"corners_3d\": [[699.01, 260.56], [786.02, 260.46], [847.51, 289.88], [731.32, 290.06], [699.01, 176.43], [786.02, 176.42], [847.51, 177.62], [731.32, 177.63]]}, {\"category\": \"car\", \"corners_3d\": [[678.29, 235.62], [736.96, 235.53], [768.58, 249.55], [696.81, 249.7], [678.29, 182.2], [736.96, 182.18], [768.58, 184.27], [696.81, 184.29]]}, {\"category\": \"car\", \"corners_3d\": [[565.02, 210.56], [529.55, 210.59], [536.99, 206.28], [568.39, 206.25], [565.02, 177.9], [529.55, 177.91], [536.99, 177.33], [568.39, 177.33]]}]\n```" - } - ] - }, - { - "id": 1524, - "data_source": "Mono3DRefer", - "scene_name": "001147", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[478.53, 294.1], [365.25, 294.68], [419.71, 264.81], [505.11, 264.47], [478.53, 191.76], [365.25, 191.85], [419.71, 187.19], [505.11, 187.14]]}, {\"category\": \"car\", \"corners_3d\": [[699.01, 260.56], [786.02, 260.46], [847.51, 289.88], [731.32, 290.06], [699.01, 176.43], [786.02, 176.42], [847.51, 177.62], [731.32, 177.63]]}, {\"category\": \"car\", \"corners_3d\": [[678.29, 235.62], [736.96, 235.53], [768.58, 249.55], [696.81, 249.7], [678.29, 182.2], [736.96, 182.18], [768.58, 184.27], [696.81, 184.29]]}, {\"category\": \"van\", \"corners_3d\": [[660.36, 223.69], [709.79, 223.59], [735.51, 234.35], [675.64, 234.5], [660.36, 171.65], [709.79, 171.65], [735.51, 171.39], [675.64, 171.39]]}, {\"category\": \"car\", \"corners_3d\": [[565.02, 210.56], [529.55, 210.59], [536.99, 206.28], [568.39, 206.25], [565.02, 177.9], [529.55, 177.91], [536.99, 177.33], [568.39, 177.33]]}, {\"category\": \"van\", \"corners_3d\": [[645.11, 205.93], [677.72, 205.89], [689.72, 210.33], [652.74, 210.39], [645.11, 168.76], [677.72, 168.76], [689.72, 168.21], [652.74, 168.21]]}]\n```" - } - ] - }, - { - "id": 1525, - "data_source": "Mono3DRefer", - "scene_name": "001147", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[845.07, 265.39], [865.08, 265.36], [893.35, 275.34], [871.19, 275.37], [845.07, 185.1], [865.08, 185.1], [893.35, 186.42], [871.19, 186.42]]}, {\"category\": \"van\", \"corners_3d\": [[660.36, 223.69], [709.79, 223.59], [735.51, 234.35], [675.64, 234.5], [660.36, 171.65], [709.79, 171.65], [735.51, 171.39], [675.64, 171.39]]}, {\"category\": \"van\", \"corners_3d\": [[645.11, 205.93], [677.72, 205.89], [689.72, 210.33], [652.74, 210.39], [645.11, 168.76], [677.72, 168.76], [689.72, 168.21], [652.74, 168.21]]}]\n```" - } - ] - }, - { - "id": 1526, - "data_source": "Mono3DRefer", - "scene_name": "001147", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[845.07, 265.39], [865.08, 265.36], [893.35, 275.34], [871.19, 275.37], [845.07, 185.1], [865.08, 185.1], [893.35, 186.42], [871.19, 186.42]]}, {\"category\": \"car\", \"corners_3d\": [[478.53, 294.1], [365.25, 294.68], [419.71, 264.81], [505.11, 264.47], [478.53, 191.76], [365.25, 191.85], [419.71, 187.19], [505.11, 187.14]]}, {\"category\": \"car\", \"corners_3d\": [[699.01, 260.56], [786.02, 260.46], [847.51, 289.88], [731.32, 290.06], [699.01, 176.43], [786.02, 176.42], [847.51, 177.62], [731.32, 177.63]]}, {\"category\": \"car\", \"corners_3d\": [[678.29, 235.62], [736.96, 235.53], [768.58, 249.55], [696.81, 249.7], [678.29, 182.2], [736.96, 182.18], [768.58, 184.27], [696.81, 184.29]]}, {\"category\": \"van\", \"corners_3d\": [[660.36, 223.69], [709.79, 223.59], [735.51, 234.35], [675.64, 234.5], [660.36, 171.65], [709.79, 171.65], [735.51, 171.39], [675.64, 171.39]]}, {\"category\": \"car\", \"corners_3d\": [[565.02, 210.56], [529.55, 210.59], [536.99, 206.28], [568.39, 206.25], [565.02, 177.9], [529.55, 177.91], [536.99, 177.33], [568.39, 177.33]]}, {\"category\": \"van\", \"corners_3d\": [[645.11, 205.93], [677.72, 205.89], [689.72, 210.33], [652.74, 210.39], [645.11, 168.76], [677.72, 168.76], [689.72, 168.21], [652.74, 168.21]]}]\n```" - } - ] - }, - { - "id": 1527, - "data_source": "Mono3DRefer", - "scene_name": "001149", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[718.67, 217.33], [741.77, 217.37], [751.32, 221.23], [726.2, 221.18], [718.67, 160.6], [741.77, 160.59], [751.32, 159.53], [726.2, 159.54]]}, {\"category\": \"cyclist\", \"corners_3d\": [[690.16, 201.62], [699.87, 201.64], [701.71, 203.0], [691.54, 202.97], [690.16, 168.21], [699.87, 168.2], [701.71, 167.98], [691.54, 167.99]]}]\n```" - } - ] - }, - { - "id": 1528, - "data_source": "Mono3DRefer", - "scene_name": "001149", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1084.75, 301.6], [1150.34, 301.13], [1216.98, 316.15], [1143.94, 316.74], [1084.75, 143.17], [1150.34, 143.28], [1216.98, 139.82], [1143.94, 139.68]]}]\n```" - } - ] - }, - { - "id": 1529, - "data_source": "Mono3DRefer", - "scene_name": "001149", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1084.75, 301.6], [1150.34, 301.13], [1216.98, 316.15], [1143.94, 316.74], [1084.75, 143.17], [1150.34, 143.28], [1216.98, 139.82], [1143.94, 139.68]]}, {\"category\": \"cyclist\", \"corners_3d\": [[718.67, 217.33], [741.77, 217.37], [751.32, 221.23], [726.2, 221.18], [718.67, 160.6], [741.77, 160.59], [751.32, 159.53], [726.2, 159.54]]}, {\"category\": \"cyclist\", \"corners_3d\": [[690.16, 201.62], [699.87, 201.64], [701.71, 203.0], [691.54, 202.97], [690.16, 168.21], [699.87, 168.2], [701.71, 167.98], [691.54, 167.99]]}]\n```" - } - ] - }, - { - "id": 1530, - "data_source": "Mono3DRefer", - "scene_name": "001150", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[346.24, 271.3], [251.29, 271.32], [325.8, 250.81], [400.97, 250.8], [346.24, 185.29], [251.29, 185.29], [325.8, 182.7], [400.97, 182.7]]}, {\"category\": \"car\", \"corners_3d\": [[965.62, 394.21], [1180.94, 392.93], [2301.08, 814.16], [1685.77, 825.16], [965.62, 189.11], [1180.94, 189.01], [2301.08, 219.94], [1685.77, 220.74]]}, {\"category\": \"car\", \"corners_3d\": [[768.45, 277.5], [875.76, 276.56], [1035.53, 330.19], [874.16, 332.35], [768.45, 180.03], [875.76, 179.96], [1035.53, 183.64], [874.16, 183.79]]}, {\"category\": \"car\", \"corners_3d\": [[418.22, 246.05], [347.7, 245.91], [397.75, 232.65], [455.53, 232.74], [418.22, 181.12], [347.7, 181.1], [397.75, 179.61], [455.53, 179.62]]}, {\"category\": \"car\", \"corners_3d\": [[730.19, 243.06], [800.67, 243.0], [855.07, 262.29], [765.24, 262.39], [730.19, 177.67], [800.67, 177.66], [855.07, 178.98], [765.24, 178.99]]}]\n```" - } - ] - }, - { - "id": 1531, - "data_source": "Mono3DRefer", - "scene_name": "001150", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[669.49, 186.79], [668.24, 186.32], [738.46, 186.36], [742.16, 186.84], [669.49, 149.82], [668.24, 150.59], [738.46, 150.53], [742.16, 149.75]]}]\n```" - } - ] - }, - { - "id": 1532, - "data_source": "Mono3DRefer", - "scene_name": "001150", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[669.49, 186.79], [668.24, 186.32], [738.46, 186.36], [742.16, 186.84], [669.49, 149.82], [668.24, 150.59], [738.46, 150.53], [742.16, 149.75]]}, {\"category\": \"car\", \"corners_3d\": [[346.24, 271.3], [251.29, 271.32], [325.8, 250.81], [400.97, 250.8], [346.24, 185.29], [251.29, 185.29], [325.8, 182.7], [400.97, 182.7]]}, {\"category\": \"car\", \"corners_3d\": [[965.62, 394.21], [1180.94, 392.93], [2301.08, 814.16], [1685.77, 825.16], [965.62, 189.11], [1180.94, 189.01], [2301.08, 219.94], [1685.77, 220.74]]}, {\"category\": \"car\", \"corners_3d\": [[768.45, 277.5], [875.76, 276.56], [1035.53, 330.19], [874.16, 332.35], [768.45, 180.03], [875.76, 179.96], [1035.53, 183.64], [874.16, 183.79]]}, {\"category\": \"car\", \"corners_3d\": [[418.22, 246.05], [347.7, 245.91], [397.75, 232.65], [455.53, 232.74], [418.22, 181.12], [347.7, 181.1], [397.75, 179.61], [455.53, 179.62]]}, {\"category\": \"car\", \"corners_3d\": [[730.19, 243.06], [800.67, 243.0], [855.07, 262.29], [765.24, 262.39], [730.19, 177.67], [800.67, 177.66], [855.07, 178.98], [765.24, 178.99]]}]\n```" - } - ] - }, - { - "id": 1533, - "data_source": "Mono3DRefer", - "scene_name": "001152", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[584.85, 193.22], [600.22, 193.21], [600.8, 194.22], [584.67, 194.24], [584.85, 181.11], [600.22, 181.11], [600.8, 181.52], [584.67, 181.53]]}, {\"category\": \"car\", \"corners_3d\": [[486.38, 210.31], [460.36, 210.3], [474.44, 206.92], [498.11, 206.93], [486.38, 184.24], [460.36, 184.24], [474.44, 183.21], [498.11, 183.21]]}, {\"category\": \"car\", \"corners_3d\": [[525.2, 195.67], [506.26, 195.67], [512.12, 194.37], [529.98, 194.37], [525.2, 179.17], [506.26, 179.17], [512.12, 178.81], [529.98, 178.81]]}]\n```" - } - ] - }, - { - "id": 1534, - "data_source": "Mono3DRefer", - "scene_name": "001154", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[467.67, 197.76], [446.81, 197.76], [456.46, 196.28], [476.08, 196.28], [467.67, 180.53], [446.81, 180.53], [456.46, 180.08], [476.08, 180.08]]}, {\"category\": \"car\", \"corners_3d\": [[232.75, 230.34], [189.65, 230.27], [241.67, 223.38], [279.65, 223.44], [232.75, 197.04], [189.65, 197.01], [241.67, 194.11], [279.65, 194.14]]}, {\"category\": \"car\", \"corners_3d\": [[391.59, 210.07], [364.08, 210.08], [383.03, 207.19], [408.41, 207.19], [391.59, 185.76], [364.08, 185.76], [383.03, 184.76], [408.41, 184.76]]}]\n```" - } - ] - }, - { - "id": 1535, - "data_source": "Mono3DRefer", - "scene_name": "001154", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[627.81, 196.05], [664.01, 196.11], [669.47, 203.25], [622.17, 203.15], [627.81, 156.93], [664.01, 156.88], [669.47, 151.98], [622.17, 152.05]]}]\n```" - } - ] - }, - { - "id": 1536, - "data_source": "Mono3DRefer", - "scene_name": "001154", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[467.67, 197.76], [446.81, 197.76], [456.46, 196.28], [476.08, 196.28], [467.67, 180.53], [446.81, 180.53], [456.46, 180.08], [476.08, 180.08]]}, {\"category\": \"bus\", \"corners_3d\": [[627.81, 196.05], [664.01, 196.11], [669.47, 203.25], [622.17, 203.15], [627.81, 156.93], [664.01, 156.88], [669.47, 151.98], [622.17, 152.05]]}, {\"category\": \"car\", \"corners_3d\": [[232.75, 230.34], [189.65, 230.27], [241.67, 223.38], [279.65, 223.44], [232.75, 197.04], [189.65, 197.01], [241.67, 194.11], [279.65, 194.14]]}, {\"category\": \"car\", \"corners_3d\": [[391.59, 210.07], [364.08, 210.08], [383.03, 207.19], [408.41, 207.19], [391.59, 185.76], [364.08, 185.76], [383.03, 184.76], [408.41, 184.76]]}]\n```" - } - ] - }, - { - "id": 1537, - "data_source": "Mono3DRefer", - "scene_name": "001159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[627.96, 194.36], [649.16, 194.34], [653.77, 195.97], [630.96, 195.99], [627.96, 174.69], [649.16, 174.68], [653.77, 174.82], [630.96, 174.82]]}, {\"category\": \"car\", \"corners_3d\": [[666.58, 214.11], [705.55, 214.0], [726.17, 220.45], [681.13, 220.6], [666.58, 172.1], [705.55, 172.1], [726.17, 171.98], [681.13, 171.98]]}, {\"category\": \"car\", \"corners_3d\": [[712.9, 234.54], [772.21, 234.5], [806.72, 247.06], [735.34, 247.13], [712.9, 181.17], [772.21, 181.16], [806.72, 182.86], [735.34, 182.86]]}, {\"category\": \"car\", \"corners_3d\": [[854.59, 232.94], [913.93, 232.8], [1003.95, 249.38], [928.43, 249.62], [854.59, 178.81], [913.93, 178.8], [1003.95, 180.44], [928.43, 180.47]]}, {\"category\": \"car\", \"corners_3d\": [[807.69, 201.11], [823.23, 200.6], [875.08, 202.63], [859.42, 203.22], [807.69, 175.45], [823.23, 175.4], [875.08, 175.59], [859.42, 175.64]]}]\n```" - } - ] - }, - { - "id": 1538, - "data_source": "Mono3DRefer", - "scene_name": "001159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[670.79, 194.21], [704.49, 194.16], [726.79, 197.8], [687.38, 197.87], [670.79, 147.11], [704.49, 147.17], [726.79, 142.78], [687.38, 142.7]]}]\n```" - } - ] - }, - { - "id": 1539, - "data_source": "Mono3DRefer", - "scene_name": "001159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[627.96, 194.36], [649.16, 194.34], [653.77, 195.97], [630.96, 195.99], [627.96, 174.69], [649.16, 174.68], [653.77, 174.82], [630.96, 174.82]]}, {\"category\": \"car\", \"corners_3d\": [[666.58, 214.11], [705.55, 214.0], [726.17, 220.45], [681.13, 220.6], [666.58, 172.1], [705.55, 172.1], [726.17, 171.98], [681.13, 171.98]]}, {\"category\": \"car\", \"corners_3d\": [[712.9, 234.54], [772.21, 234.5], [806.72, 247.06], [735.34, 247.13], [712.9, 181.17], [772.21, 181.16], [806.72, 182.86], [735.34, 182.86]]}, {\"category\": \"truck\", \"corners_3d\": [[670.79, 194.21], [704.49, 194.16], [726.79, 197.8], [687.38, 197.87], [670.79, 147.11], [704.49, 147.17], [726.79, 142.78], [687.38, 142.7]]}, {\"category\": \"car\", \"corners_3d\": [[854.59, 232.94], [913.93, 232.8], [1003.95, 249.38], [928.43, 249.62], [854.59, 178.81], [913.93, 178.8], [1003.95, 180.44], [928.43, 180.47]]}, {\"category\": \"car\", \"corners_3d\": [[807.69, 201.11], [823.23, 200.6], [875.08, 202.63], [859.42, 203.22], [807.69, 175.45], [823.23, 175.4], [875.08, 175.59], [859.42, 175.64]]}]\n```" - } - ] - }, - { - "id": 1540, - "data_source": "Mono3DRefer", - "scene_name": "001161", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[249.47, 318.27], [224.44, 328.2], [114.02, 328.42], [146.51, 318.47], [249.47, 143.81], [224.44, 141.17], [114.02, 141.1], [146.51, 143.76]]}]\n```" - } - ] - }, - { - "id": 1541, - "data_source": "Mono3DRefer", - "scene_name": "001163", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[642.34, 191.94], [666.46, 191.96], [670.09, 193.95], [643.45, 193.93], [642.34, 163.0], [666.46, 162.99], [670.09, 161.96], [643.45, 161.97]]}]\n```" - } - ] - }, - { - "id": 1542, - "data_source": "Mono3DRefer", - "scene_name": "001163", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-97.22, 697.98], [-618.23, 682.52], [218.05, 343.19], [397.05, 344.89], [-97.22, 212.88], [-618.23, 211.7], [218.05, 185.84], [397.05, 185.97]]}, {\"category\": \"car\", \"corners_3d\": [[852.38, 384.44], [1038.06, 382.84], [1689.63, 687.16], [1242.09, 696.9], [852.38, 209.67], [1038.06, 209.39], [1689.63, 262.34], [1242.09, 264.03]]}, {\"category\": \"car\", \"corners_3d\": [[400.05, 309.62], [258.67, 309.1], [368.85, 267.95], [467.7, 268.2], [400.05, 177.07], [258.67, 177.06], [368.85, 175.79], [467.7, 175.79]]}, {\"category\": \"car\", \"corners_3d\": [[953.74, 342.46], [798.34, 342.49], [727.99, 279.47], [825.67, 279.45], [953.74, 205.99], [798.34, 206.0], [727.99, 193.68], [825.67, 193.68]]}, {\"category\": \"car\", \"corners_3d\": [[512.51, 234.18], [449.65, 234.29], [472.41, 224.64], [525.38, 224.56], [512.51, 173.67], [449.65, 173.67], [472.41, 173.54], [525.38, 173.54]]}, {\"category\": \"car\", \"corners_3d\": [[740.23, 225.64], [687.39, 225.6], [675.89, 217.08], [720.19, 217.1], [740.23, 179.81], [687.39, 179.81], [675.89, 178.69], [720.19, 178.69]]}, {\"category\": \"car\", \"corners_3d\": [[546.9, 216.96], [496.71, 217.02], [509.74, 211.14], [553.24, 211.09], [546.9, 176.1], [496.71, 176.1], [509.74, 175.67], [553.24, 175.67]]}, {\"category\": \"car\", \"corners_3d\": [[651.5, 198.14], [676.06, 198.16], [680.72, 200.45], [653.94, 200.43], [651.5, 174.42], [676.06, 174.42], [680.72, 174.56], [653.94, 174.56]]}]\n```" - } - ] - }, - { - "id": 1543, - "data_source": "Mono3DRefer", - "scene_name": "001163", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-97.22, 697.98], [-618.23, 682.52], [218.05, 343.19], [397.05, 344.89], [-97.22, 212.88], [-618.23, 211.7], [218.05, 185.84], [397.05, 185.97]]}, {\"category\": \"car\", \"corners_3d\": [[852.38, 384.44], [1038.06, 382.84], [1689.63, 687.16], [1242.09, 696.9], [852.38, 209.67], [1038.06, 209.39], [1689.63, 262.34], [1242.09, 264.03]]}, {\"category\": \"car\", \"corners_3d\": [[400.05, 309.62], [258.67, 309.1], [368.85, 267.95], [467.7, 268.2], [400.05, 177.07], [258.67, 177.06], [368.85, 175.79], [467.7, 175.79]]}, {\"category\": \"car\", \"corners_3d\": [[953.74, 342.46], [798.34, 342.49], [727.99, 279.47], [825.67, 279.45], [953.74, 205.99], [798.34, 206.0], [727.99, 193.68], [825.67, 193.68]]}, {\"category\": \"car\", \"corners_3d\": [[512.51, 234.18], [449.65, 234.29], [472.41, 224.64], [525.38, 224.56], [512.51, 173.67], [449.65, 173.67], [472.41, 173.54], [525.38, 173.54]]}, {\"category\": \"car\", \"corners_3d\": [[740.23, 225.64], [687.39, 225.6], [675.89, 217.08], [720.19, 217.1], [740.23, 179.81], [687.39, 179.81], [675.89, 178.69], [720.19, 178.69]]}, {\"category\": \"car\", \"corners_3d\": [[546.9, 216.96], [496.71, 217.02], [509.74, 211.14], [553.24, 211.09], [546.9, 176.1], [496.71, 176.1], [509.74, 175.67], [553.24, 175.67]]}, {\"category\": \"car\", \"corners_3d\": [[651.5, 198.14], [676.06, 198.16], [680.72, 200.45], [653.94, 200.43], [651.5, 174.42], [676.06, 174.42], [680.72, 174.56], [653.94, 174.56]]}, {\"category\": \"van\", \"corners_3d\": [[642.34, 191.94], [666.46, 191.96], [670.09, 193.95], [643.45, 193.93], [642.34, 163.0], [666.46, 162.99], [670.09, 161.96], [643.45, 161.97]]}]\n```" - } - ] - }, - { - "id": 1544, - "data_source": "Mono3DRefer", - "scene_name": "001167", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[756.4, 214.06], [739.1, 214.08], [736.19, 213.25], [753.14, 213.23], [756.4, 173.77], [739.1, 173.77], [736.19, 173.75], [753.14, 173.75]]}]\n```" - } - ] - }, - { - "id": 1545, - "data_source": "Mono3DRefer", - "scene_name": "001167", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[325.27, 215.99], [287.12, 215.92], [322.2, 211.52], [356.51, 211.58], [325.27, 182.41], [287.12, 182.39], [322.2, 181.42], [356.51, 181.43]]}]\n```" - } - ] - }, - { - "id": 1546, - "data_source": "Mono3DRefer", - "scene_name": "001167", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[756.4, 214.06], [739.1, 214.08], [736.19, 213.25], [753.14, 213.23], [756.4, 173.77], [739.1, 173.77], [736.19, 173.75], [753.14, 173.75]]}, {\"category\": \"car\", \"corners_3d\": [[325.27, 215.99], [287.12, 215.92], [322.2, 211.52], [356.51, 211.58], [325.27, 182.41], [287.12, 182.39], [322.2, 181.42], [356.51, 181.43]]}]\n```" - } - ] - }, - { - "id": 1547, - "data_source": "Mono3DRefer", - "scene_name": "001170", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[561.06, 213.55], [607.2, 213.89], [586.96, 222.43], [531.53, 221.92], [561.06, 166.46], [607.2, 166.41], [586.96, 165.06], [531.53, 165.14]]}, {\"category\": \"car\", \"corners_3d\": [[68.32, 283.48], [125.54, 276.99], [186.03, 285.79], [127.75, 293.47], [68.32, 218.53], [125.54, 215.85], [186.03, 219.49], [127.75, 222.65]]}, {\"category\": \"car\", \"corners_3d\": [[1333.16, 247.81], [1455.49, 261.15], [1130.68, 259.27], [1058.74, 246.46], [1333.16, 148.73], [1455.49, 144.44], [1130.68, 145.04], [1058.74, 149.17]]}, {\"category\": \"car\", \"corners_3d\": [[1119.33, 226.79], [1176.44, 233.39], [912.45, 232.12], [884.67, 225.78], [1119.33, 150.53], [1176.44, 147.79], [912.45, 148.32], [884.67, 150.95]]}, {\"category\": \"car\", \"corners_3d\": [[628.35, 196.79], [592.33, 196.69], [600.88, 194.19], [633.15, 194.28], [628.35, 162.28], [592.33, 162.32], [600.88, 163.42], [633.15, 163.39]]}, {\"category\": \"car\", \"corners_3d\": [[658.9, 176.11], [639.61, 176.1], [642.64, 175.9], [660.74, 175.91], [658.9, 159.59], [639.61, 159.63], [642.64, 160.44], [660.74, 160.41]]}]\n```" - } - ] - }, - { - "id": 1548, - "data_source": "Mono3DRefer", - "scene_name": "001172", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[694.22, 222.03], [739.57, 222.52], [741.58, 230.24], [689.24, 229.58], [694.22, 177.92], [739.57, 177.97], [741.58, 178.77], [689.24, 178.7]]}, {\"category\": \"car\", \"corners_3d\": [[-51.55, 342.36], [-189.03, 337.17], [144.95, 276.17], [236.96, 278.2], [-51.55, 188.52], [-189.03, 188.04], [144.95, 182.4], [236.96, 182.59]]}, {\"category\": \"car\", \"corners_3d\": [[475.72, 230.01], [422.57, 229.31], [473.27, 219.86], [517.95, 220.34], [475.72, 183.21], [422.57, 183.08], [473.27, 181.37], [517.95, 181.45]]}, {\"category\": \"car\", \"corners_3d\": [[557.85, 215.99], [521.38, 215.66], [542.28, 211.11], [574.99, 211.37], [557.85, 182.36], [521.38, 182.28], [542.28, 181.28], [574.99, 181.34]]}, {\"category\": \"car\", \"corners_3d\": [[585.68, 208.1], [559.51, 207.89], [575.19, 204.58], [598.96, 204.75], [585.68, 180.01], [559.51, 179.97], [575.19, 179.29], [598.96, 179.33]]}, {\"category\": \"car\", \"corners_3d\": [[609.95, 203.05], [587.12, 202.89], [595.96, 200.97], [617.37, 201.11], [609.95, 179.51], [587.12, 179.48], [595.96, 179.05], [617.37, 179.08]]}]\n```" - } - ] - }, - { - "id": 1549, - "data_source": "Mono3DRefer", - "scene_name": "001181", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[878.8, 262.26], [963.9, 264.11], [1046.46, 296.08], [927.3, 292.57], [878.8, 188.81], [963.9, 189.0], [1046.46, 192.24], [927.3, 191.89]]}, {\"category\": \"car\", \"corners_3d\": [[851.35, 233.64], [914.4, 234.58], [960.07, 250.11], [878.3, 248.55], [851.35, 179.01], [914.4, 178.98], [960.07, 178.54], [878.3, 178.58]]}, {\"category\": \"car\", \"corners_3d\": [[581.16, 200.37], [555.32, 200.11], [580.06, 198.48], [604.03, 198.7], [581.16, 169.44], [555.32, 169.59], [580.06, 170.49], [604.03, 170.37]]}, {\"category\": \"car\", \"corners_3d\": [[819.14, 211.29], [857.22, 211.82], [857.51, 215.65], [814.86, 214.98], [819.14, 176.13], [857.22, 176.05], [857.51, 175.51], [814.86, 175.6]]}]\n```" - } - ] - }, - { - "id": 1550, - "data_source": "Mono3DRefer", - "scene_name": "001181", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[10.7, 283.59], [-85.02, 279.08], [177.27, 249.33], [251.47, 251.5], [10.7, 147.9], [-85.02, 149.33], [177.27, 158.74], [251.47, 158.05]]}]\n```" - } - ] - }, - { - "id": 1551, - "data_source": "Mono3DRefer", - "scene_name": "001181", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[10.7, 283.59], [-85.02, 279.08], [177.27, 249.33], [251.47, 251.5], [10.7, 147.9], [-85.02, 149.33], [177.27, 158.74], [251.47, 158.05]]}, {\"category\": \"car\", \"corners_3d\": [[878.8, 262.26], [963.9, 264.11], [1046.46, 296.08], [927.3, 292.57], [878.8, 188.81], [963.9, 189.0], [1046.46, 192.24], [927.3, 191.89]]}, {\"category\": \"car\", \"corners_3d\": [[851.35, 233.64], [914.4, 234.58], [960.07, 250.11], [878.3, 248.55], [851.35, 179.01], [914.4, 178.98], [960.07, 178.54], [878.3, 178.58]]}, {\"category\": \"car\", \"corners_3d\": [[581.16, 200.37], [555.32, 200.11], [580.06, 198.48], [604.03, 198.7], [581.16, 169.44], [555.32, 169.59], [580.06, 170.49], [604.03, 170.37]]}, {\"category\": \"car\", \"corners_3d\": [[819.14, 211.29], [857.22, 211.82], [857.51, 215.65], [814.86, 214.98], [819.14, 176.13], [857.22, 176.05], [857.51, 175.51], [814.86, 175.6]]}]\n```" - } - ] - }, - { - "id": 1552, - "data_source": "Mono3DRefer", - "scene_name": "001182", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[838.36, 307.29], [965.89, 307.08], [1270.56, 419.74], [1036.59, 420.48], [838.36, 174.64], [965.89, 174.64], [1270.56, 176.13], [1036.59, 176.14]]}, {\"category\": \"car\", \"corners_3d\": [[729.75, 227.12], [788.33, 227.17], [838.82, 243.21], [762.9, 243.13], [729.75, 173.59], [788.33, 173.6], [838.82, 173.81], [762.9, 173.81]]}, {\"category\": \"car\", \"corners_3d\": [[580.74, 209.61], [617.23, 209.75], [611.22, 214.84], [569.74, 214.67], [580.74, 181.18], [617.23, 181.21], [611.22, 182.37], [569.74, 182.33]]}, {\"category\": \"car\", \"corners_3d\": [[604.83, 201.92], [630.45, 202.05], [623.39, 205.14], [595.11, 204.99], [604.83, 178.06], [630.45, 178.08], [623.39, 178.63], [595.11, 178.61]]}]\n```" - } - ] - }, - { - "id": 1553, - "data_source": "Mono3DRefer", - "scene_name": "001182", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[703.28, 205.35], [748.24, 205.53], [767.88, 214.32], [710.77, 214.03], [703.28, 144.14], [748.24, 143.98], [767.88, 136.21], [710.77, 136.47]]}]\n```" - } - ] - }, - { - "id": 1554, - "data_source": "Mono3DRefer", - "scene_name": "001182", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[838.36, 307.29], [965.89, 307.08], [1270.56, 419.74], [1036.59, 420.48], [838.36, 174.64], [965.89, 174.64], [1270.56, 176.13], [1036.59, 176.14]]}, {\"category\": \"car\", \"corners_3d\": [[729.75, 227.12], [788.33, 227.17], [838.82, 243.21], [762.9, 243.13], [729.75, 173.59], [788.33, 173.6], [838.82, 173.81], [762.9, 173.81]]}, {\"category\": \"car\", \"corners_3d\": [[580.74, 209.61], [617.23, 209.75], [611.22, 214.84], [569.74, 214.67], [580.74, 181.18], [617.23, 181.21], [611.22, 182.37], [569.74, 182.33]]}, {\"category\": \"van\", \"corners_3d\": [[703.28, 205.35], [748.24, 205.53], [767.88, 214.32], [710.77, 214.03], [703.28, 144.14], [748.24, 143.98], [767.88, 136.21], [710.77, 136.47]]}, {\"category\": \"car\", \"corners_3d\": [[604.83, 201.92], [630.45, 202.05], [623.39, 205.14], [595.11, 204.99], [604.83, 178.06], [630.45, 178.08], [623.39, 178.63], [595.11, 178.61]]}]\n```" - } - ] - }, - { - "id": 1555, - "data_source": "Mono3DRefer", - "scene_name": "001185", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[359.24, 251.99], [299.32, 247.92], [442.78, 233.91], [496.44, 236.43], [359.24, 159.83], [299.32, 161.01], [442.78, 165.06], [496.44, 164.33]]}]\n```" - } - ] - }, - { - "id": 1556, - "data_source": "Mono3DRefer", - "scene_name": "001185", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1034.36, 260.74], [1123.84, 264.46], [1183.88, 291.53], [1063.74, 285.12], [1034.36, 192.17], [1123.84, 192.72], [1183.88, 196.65], [1063.74, 195.72]]}, {\"category\": \"car\", \"corners_3d\": [[1027.94, 239.99], [1098.57, 242.13], [1143.81, 258.57], [1053.18, 255.17], [1027.94, 185.47], [1098.57, 185.64], [1143.81, 187.01], [1053.18, 186.73]]}, {\"category\": \"car\", \"corners_3d\": [[781.31, 203.19], [757.72, 202.76], [778.57, 201.31], [800.97, 201.68], [781.31, 175.91], [757.72, 176.0], [778.57, 176.29], [800.97, 176.21]]}, {\"category\": \"car\", \"corners_3d\": [[1000.49, 217.68], [1043.46, 218.66], [1047.34, 223.04], [999.47, 221.83], [1000.49, 181.86], [1043.46, 181.9], [1047.34, 182.06], [999.47, 182.01]]}]\n```" - } - ] - }, - { - "id": 1557, - "data_source": "Mono3DRefer", - "scene_name": "001185", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[359.24, 251.99], [299.32, 247.92], [442.78, 233.91], [496.44, 236.43], [359.24, 159.83], [299.32, 161.01], [442.78, 165.06], [496.44, 164.33]]}, {\"category\": \"car\", \"corners_3d\": [[1034.36, 260.74], [1123.84, 264.46], [1183.88, 291.53], [1063.74, 285.12], [1034.36, 192.17], [1123.84, 192.72], [1183.88, 196.65], [1063.74, 195.72]]}, {\"category\": \"car\", \"corners_3d\": [[1027.94, 239.99], [1098.57, 242.13], [1143.81, 258.57], [1053.18, 255.17], [1027.94, 185.47], [1098.57, 185.64], [1143.81, 187.01], [1053.18, 186.73]]}, {\"category\": \"car\", \"corners_3d\": [[781.31, 203.19], [757.72, 202.76], [778.57, 201.31], [800.97, 201.68], [781.31, 175.91], [757.72, 176.0], [778.57, 176.29], [800.97, 176.21]]}, {\"category\": \"car\", \"corners_3d\": [[1000.49, 217.68], [1043.46, 218.66], [1047.34, 223.04], [999.47, 221.83], [1000.49, 181.86], [1043.46, 181.9], [1047.34, 182.06], [999.47, 182.01]]}]\n```" - } - ] - }, - { - "id": 1558, - "data_source": "Mono3DRefer", - "scene_name": "001190", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[784.89, 354.32], [942.77, 353.95], [1285.41, 536.54], [969.1, 538.02], [784.89, 201.74], [942.77, 201.68], [1285.41, 230.74], [969.1, 230.98]]}, {\"category\": \"car\", \"corners_3d\": [[550.22, 248.35], [485.4, 248.08], [511.87, 235.02], [565.48, 235.2], [550.22, 181.82], [485.4, 181.79], [511.87, 180.24], [565.48, 180.26]]}, {\"category\": \"car\", \"corners_3d\": [[735.92, 275.16], [821.06, 276.47], [865.17, 311.3], [751.15, 308.96], [735.92, 194.96], [821.06, 195.24], [865.17, 202.77], [751.15, 202.26]]}, {\"category\": \"car\", \"corners_3d\": [[695.17, 245.97], [751.16, 246.03], [779.3, 261.41], [711.52, 261.32], [695.17, 186.34], [751.16, 186.35], [779.3, 189.19], [711.52, 189.17]]}, {\"category\": \"car\", \"corners_3d\": [[666.33, 230.31], [712.86, 230.23], [732.26, 239.74], [678.04, 239.83], [666.33, 185.39], [712.86, 185.37], [732.26, 187.44], [678.04, 187.46]]}, {\"category\": \"car\", \"corners_3d\": [[647.83, 214.02], [679.44, 213.98], [689.15, 218.75], [653.88, 218.79], [647.83, 184.72], [679.44, 184.71], [689.15, 186.09], [653.88, 186.1]]}, {\"category\": \"car\", \"corners_3d\": [[599.26, 204.94], [574.66, 204.89], [579.71, 202.34], [602.34, 202.37], [599.26, 176.56], [574.66, 176.56], [579.71, 176.26], [602.34, 176.27]]}]\n```" - } - ] - }, - { - "id": 1559, - "data_source": "Mono3DRefer", - "scene_name": "001192", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[659.57, 220.4], [711.24, 220.37], [738.9, 232.5], [674.05, 232.55], [659.57, 158.21], [711.24, 158.22], [738.9, 154.49], [674.05, 154.47]]}]\n```" - } - ] - }, - { - "id": 1560, - "data_source": "Mono3DRefer", - "scene_name": "001192", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[724.12, 260.21], [802.48, 260.22], [871.89, 291.75], [765.26, 291.74], [724.12, 184.39], [802.48, 184.39], [871.89, 188.55], [765.26, 188.55]]}, {\"category\": \"car\", \"corners_3d\": [[689.15, 237.59], [750.72, 237.49], [788.25, 253.14], [711.82, 253.3], [689.15, 179.83], [750.72, 179.82], [788.25, 181.51], [711.82, 181.52]]}, {\"category\": \"car\", \"corners_3d\": [[548.78, 219.54], [509.79, 219.72], [515.68, 214.04], [549.94, 213.91], [548.78, 180.33], [509.79, 180.36], [515.68, 179.45], [549.94, 179.42]]}]\n```" - } - ] - }, - { - "id": 1561, - "data_source": "Mono3DRefer", - "scene_name": "001192", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[724.12, 260.21], [802.48, 260.22], [871.89, 291.75], [765.26, 291.74], [724.12, 184.39], [802.48, 184.39], [871.89, 188.55], [765.26, 188.55]]}, {\"category\": \"car\", \"corners_3d\": [[689.15, 237.59], [750.72, 237.49], [788.25, 253.14], [711.82, 253.3], [689.15, 179.83], [750.72, 179.82], [788.25, 181.51], [711.82, 181.52]]}, {\"category\": \"van\", \"corners_3d\": [[659.57, 220.4], [711.24, 220.37], [738.9, 232.5], [674.05, 232.55], [659.57, 158.21], [711.24, 158.22], [738.9, 154.49], [674.05, 154.47]]}, {\"category\": \"car\", \"corners_3d\": [[548.78, 219.54], [509.79, 219.72], [515.68, 214.04], [549.94, 213.91], [548.78, 180.33], [509.79, 180.36], [515.68, 179.45], [549.94, 179.42]]}]\n```" - } - ] - }, - { - "id": 1562, - "data_source": "Mono3DRefer", - "scene_name": "001194", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[586.8, 191.37], [602.15, 191.36], [602.83, 192.28], [586.71, 192.29], [586.8, 179.28], [602.15, 179.28], [602.83, 179.6], [586.71, 179.6]]}, {\"category\": \"car\", \"corners_3d\": [[517.3, 199.67], [497.93, 199.64], [507.08, 197.8], [525.12, 197.82], [517.3, 180.2], [497.93, 180.19], [507.08, 179.69], [525.12, 179.69]]}]\n```" - } - ] - }, - { - "id": 1563, - "data_source": "Mono3DRefer", - "scene_name": "001197", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[540.33, 210.81], [507.56, 210.86], [515.35, 207.15], [544.92, 207.11], [540.33, 178.54], [507.56, 178.55], [515.35, 177.99], [544.92, 177.99]]}]\n```" - } - ] - }, - { - "id": 1564, - "data_source": "Mono3DRefer", - "scene_name": "001198", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[507.84, 192.92], [517.74, 193.5], [325.53, 194.72], [324.89, 194.02], [507.84, 143.46], [517.74, 141.52], [325.53, 137.43], [324.89, 139.77]]}]\n```" - } - ] - }, - { - "id": 1565, - "data_source": "Mono3DRefer", - "scene_name": "001198", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[755.62, 276.54], [869.68, 276.1], [1024.36, 326.5], [850.2, 327.55], [755.62, 152.4], [869.68, 152.54], [1024.36, 137.1], [850.2, 136.78]]}]\n```" - } - ] - }, - { - "id": 1566, - "data_source": "Mono3DRefer", - "scene_name": "001198", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[452.8, 265.97], [493.95, 265.93], [487.44, 271.44], [443.59, 271.49], [452.8, 188.47], [493.95, 188.47], [487.44, 188.92], [443.59, 188.92]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[414.86, 265.28], [454.81, 265.24], [445.62, 270.78], [403.03, 270.82], [414.86, 189.05], [454.81, 189.04], [445.62, 189.54], [403.03, 189.55]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[463.72, 264.07], [430.02, 263.99], [438.97, 259.98], [471.04, 260.05], [463.72, 190.04], [430.02, 190.03], [438.97, 189.61], [471.04, 189.62]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[483.11, 250.28], [455.07, 250.29], [460.69, 247.62], [487.64, 247.61], [483.11, 185.86], [455.07, 185.86], [460.69, 185.7], [487.64, 185.7]]}]\n```" - } - ] - }, - { - "id": 1567, - "data_source": "Mono3DRefer", - "scene_name": "001198", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[308.37, 270.53], [347.76, 269.62], [342.58, 273.86], [301.22, 274.86], [308.37, 163.84], [347.76, 164.02], [342.58, 163.18], [301.22, 162.98]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[327.73, 267.59], [295.61, 267.51], [308.17, 264.12], [339.02, 264.19], [327.73, 162.34], [295.61, 162.36], [308.17, 163.11], [339.02, 163.1]]}]\n```" - } - ] - }, - { - "id": 1568, - "data_source": "Mono3DRefer", - "scene_name": "001198", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[755.62, 276.54], [869.68, 276.1], [1024.36, 326.5], [850.2, 327.55], [755.62, 152.4], [869.68, 152.54], [1024.36, 137.1], [850.2, 136.78]]}, {\"category\": \"tram\", \"corners_3d\": [[507.84, 192.92], [517.74, 193.5], [325.53, 194.72], [324.89, 194.02], [507.84, 143.46], [517.74, 141.52], [325.53, 137.43], [324.89, 139.77]]}]\n```" - } - ] - }, - { - "id": 1569, - "data_source": "Mono3DRefer", - "scene_name": "001198", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[507.84, 192.92], [517.74, 193.5], [325.53, 194.72], [324.89, 194.02], [507.84, 143.46], [517.74, 141.52], [325.53, 137.43], [324.89, 139.77]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[452.8, 265.97], [493.95, 265.93], [487.44, 271.44], [443.59, 271.49], [452.8, 188.47], [493.95, 188.47], [487.44, 188.92], [443.59, 188.92]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[414.86, 265.28], [454.81, 265.24], [445.62, 270.78], [403.03, 270.82], [414.86, 189.05], [454.81, 189.04], [445.62, 189.54], [403.03, 189.55]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[463.72, 264.07], [430.02, 263.99], [438.97, 259.98], [471.04, 260.05], [463.72, 190.04], [430.02, 190.03], [438.97, 189.61], [471.04, 189.62]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[483.11, 250.28], [455.07, 250.29], [460.69, 247.62], [487.64, 247.61], [483.11, 185.86], [455.07, 185.86], [460.69, 185.7], [487.64, 185.7]]}]\n```" - } - ] - }, - { - "id": 1570, - "data_source": "Mono3DRefer", - "scene_name": "001198", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[507.84, 192.92], [517.74, 193.5], [325.53, 194.72], [324.89, 194.02], [507.84, 143.46], [517.74, 141.52], [325.53, 137.43], [324.89, 139.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[308.37, 270.53], [347.76, 269.62], [342.58, 273.86], [301.22, 274.86], [308.37, 163.84], [347.76, 164.02], [342.58, 163.18], [301.22, 162.98]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[327.73, 267.59], [295.61, 267.51], [308.17, 264.12], [339.02, 264.19], [327.73, 162.34], [295.61, 162.36], [308.17, 163.11], [339.02, 163.1]]}]\n```" - } - ] - }, - { - "id": 1571, - "data_source": "Mono3DRefer", - "scene_name": "001198", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[755.62, 276.54], [869.68, 276.1], [1024.36, 326.5], [850.2, 327.55], [755.62, 152.4], [869.68, 152.54], [1024.36, 137.1], [850.2, 136.78]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[452.8, 265.97], [493.95, 265.93], [487.44, 271.44], [443.59, 271.49], [452.8, 188.47], [493.95, 188.47], [487.44, 188.92], [443.59, 188.92]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[414.86, 265.28], [454.81, 265.24], [445.62, 270.78], [403.03, 270.82], [414.86, 189.05], [454.81, 189.04], [445.62, 189.54], [403.03, 189.55]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[463.72, 264.07], [430.02, 263.99], [438.97, 259.98], [471.04, 260.05], [463.72, 190.04], [430.02, 190.03], [438.97, 189.61], [471.04, 189.62]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[483.11, 250.28], [455.07, 250.29], [460.69, 247.62], [487.64, 247.61], [483.11, 185.86], [455.07, 185.86], [460.69, 185.7], [487.64, 185.7]]}]\n```" - } - ] - }, - { - "id": 1572, - "data_source": "Mono3DRefer", - "scene_name": "001198", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[755.62, 276.54], [869.68, 276.1], [1024.36, 326.5], [850.2, 327.55], [755.62, 152.4], [869.68, 152.54], [1024.36, 137.1], [850.2, 136.78]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[308.37, 270.53], [347.76, 269.62], [342.58, 273.86], [301.22, 274.86], [308.37, 163.84], [347.76, 164.02], [342.58, 163.18], [301.22, 162.98]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[327.73, 267.59], [295.61, 267.51], [308.17, 264.12], [339.02, 264.19], [327.73, 162.34], [295.61, 162.36], [308.17, 163.11], [339.02, 163.1]]}]\n```" - } - ] - }, - { - "id": 1573, - "data_source": "Mono3DRefer", - "scene_name": "001198", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[452.8, 265.97], [493.95, 265.93], [487.44, 271.44], [443.59, 271.49], [452.8, 188.47], [493.95, 188.47], [487.44, 188.92], [443.59, 188.92]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[414.86, 265.28], [454.81, 265.24], [445.62, 270.78], [403.03, 270.82], [414.86, 189.05], [454.81, 189.04], [445.62, 189.54], [403.03, 189.55]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[463.72, 264.07], [430.02, 263.99], [438.97, 259.98], [471.04, 260.05], [463.72, 190.04], [430.02, 190.03], [438.97, 189.61], [471.04, 189.62]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[483.11, 250.28], [455.07, 250.29], [460.69, 247.62], [487.64, 247.61], [483.11, 185.86], [455.07, 185.86], [460.69, 185.7], [487.64, 185.7]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[308.37, 270.53], [347.76, 269.62], [342.58, 273.86], [301.22, 274.86], [308.37, 163.84], [347.76, 164.02], [342.58, 163.18], [301.22, 162.98]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[327.73, 267.59], [295.61, 267.51], [308.17, 264.12], [339.02, 264.19], [327.73, 162.34], [295.61, 162.36], [308.17, 163.11], [339.02, 163.1]]}]\n```" - } - ] - }, - { - "id": 1574, - "data_source": "Mono3DRefer", - "scene_name": "001198", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, van, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[755.62, 276.54], [869.68, 276.1], [1024.36, 326.5], [850.2, 327.55], [755.62, 152.4], [869.68, 152.54], [1024.36, 137.1], [850.2, 136.78]]}, {\"category\": \"tram\", \"corners_3d\": [[507.84, 192.92], [517.74, 193.5], [325.53, 194.72], [324.89, 194.02], [507.84, 143.46], [517.74, 141.52], [325.53, 137.43], [324.89, 139.77]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[452.8, 265.97], [493.95, 265.93], [487.44, 271.44], [443.59, 271.49], [452.8, 188.47], [493.95, 188.47], [487.44, 188.92], [443.59, 188.92]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[414.86, 265.28], [454.81, 265.24], [445.62, 270.78], [403.03, 270.82], [414.86, 189.05], [454.81, 189.04], [445.62, 189.54], [403.03, 189.55]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[463.72, 264.07], [430.02, 263.99], [438.97, 259.98], [471.04, 260.05], [463.72, 190.04], [430.02, 190.03], [438.97, 189.61], [471.04, 189.62]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[483.11, 250.28], [455.07, 250.29], [460.69, 247.62], [487.64, 247.61], [483.11, 185.86], [455.07, 185.86], [460.69, 185.7], [487.64, 185.7]]}]\n```" - } - ] - }, - { - "id": 1575, - "data_source": "Mono3DRefer", - "scene_name": "001198", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[755.62, 276.54], [869.68, 276.1], [1024.36, 326.5], [850.2, 327.55], [755.62, 152.4], [869.68, 152.54], [1024.36, 137.1], [850.2, 136.78]]}, {\"category\": \"tram\", \"corners_3d\": [[507.84, 192.92], [517.74, 193.5], [325.53, 194.72], [324.89, 194.02], [507.84, 143.46], [517.74, 141.52], [325.53, 137.43], [324.89, 139.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[308.37, 270.53], [347.76, 269.62], [342.58, 273.86], [301.22, 274.86], [308.37, 163.84], [347.76, 164.02], [342.58, 163.18], [301.22, 162.98]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[327.73, 267.59], [295.61, 267.51], [308.17, 264.12], [339.02, 264.19], [327.73, 162.34], [295.61, 162.36], [308.17, 163.11], [339.02, 163.1]]}]\n```" - } - ] - }, - { - "id": 1576, - "data_source": "Mono3DRefer", - "scene_name": "001198", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, person_sitting, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[507.84, 192.92], [517.74, 193.5], [325.53, 194.72], [324.89, 194.02], [507.84, 143.46], [517.74, 141.52], [325.53, 137.43], [324.89, 139.77]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[452.8, 265.97], [493.95, 265.93], [487.44, 271.44], [443.59, 271.49], [452.8, 188.47], [493.95, 188.47], [487.44, 188.92], [443.59, 188.92]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[414.86, 265.28], [454.81, 265.24], [445.62, 270.78], [403.03, 270.82], [414.86, 189.05], [454.81, 189.04], [445.62, 189.54], [403.03, 189.55]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[463.72, 264.07], [430.02, 263.99], [438.97, 259.98], [471.04, 260.05], [463.72, 190.04], [430.02, 190.03], [438.97, 189.61], [471.04, 189.62]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[483.11, 250.28], [455.07, 250.29], [460.69, 247.62], [487.64, 247.61], [483.11, 185.86], [455.07, 185.86], [460.69, 185.7], [487.64, 185.7]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[308.37, 270.53], [347.76, 269.62], [342.58, 273.86], [301.22, 274.86], [308.37, 163.84], [347.76, 164.02], [342.58, 163.18], [301.22, 162.98]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[327.73, 267.59], [295.61, 267.51], [308.17, 264.12], [339.02, 264.19], [327.73, 162.34], [295.61, 162.36], [308.17, 163.11], [339.02, 163.1]]}]\n```" - } - ] - }, - { - "id": 1577, - "data_source": "Mono3DRefer", - "scene_name": "001198", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, person_sitting, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[755.62, 276.54], [869.68, 276.1], [1024.36, 326.5], [850.2, 327.55], [755.62, 152.4], [869.68, 152.54], [1024.36, 137.1], [850.2, 136.78]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[452.8, 265.97], [493.95, 265.93], [487.44, 271.44], [443.59, 271.49], [452.8, 188.47], [493.95, 188.47], [487.44, 188.92], [443.59, 188.92]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[414.86, 265.28], [454.81, 265.24], [445.62, 270.78], [403.03, 270.82], [414.86, 189.05], [454.81, 189.04], [445.62, 189.54], [403.03, 189.55]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[463.72, 264.07], [430.02, 263.99], [438.97, 259.98], [471.04, 260.05], [463.72, 190.04], [430.02, 190.03], [438.97, 189.61], [471.04, 189.62]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[483.11, 250.28], [455.07, 250.29], [460.69, 247.62], [487.64, 247.61], [483.11, 185.86], [455.07, 185.86], [460.69, 185.7], [487.64, 185.7]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[308.37, 270.53], [347.76, 269.62], [342.58, 273.86], [301.22, 274.86], [308.37, 163.84], [347.76, 164.02], [342.58, 163.18], [301.22, 162.98]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[327.73, 267.59], [295.61, 267.51], [308.17, 264.12], [339.02, 264.19], [327.73, 162.34], [295.61, 162.36], [308.17, 163.11], [339.02, 163.1]]}]\n```" - } - ] - }, - { - "id": 1578, - "data_source": "Mono3DRefer", - "scene_name": "001198", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, van, person_sitting, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[755.62, 276.54], [869.68, 276.1], [1024.36, 326.5], [850.2, 327.55], [755.62, 152.4], [869.68, 152.54], [1024.36, 137.1], [850.2, 136.78]]}, {\"category\": \"tram\", \"corners_3d\": [[507.84, 192.92], [517.74, 193.5], [325.53, 194.72], [324.89, 194.02], [507.84, 143.46], [517.74, 141.52], [325.53, 137.43], [324.89, 139.77]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[452.8, 265.97], [493.95, 265.93], [487.44, 271.44], [443.59, 271.49], [452.8, 188.47], [493.95, 188.47], [487.44, 188.92], [443.59, 188.92]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[414.86, 265.28], [454.81, 265.24], [445.62, 270.78], [403.03, 270.82], [414.86, 189.05], [454.81, 189.04], [445.62, 189.54], [403.03, 189.55]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[463.72, 264.07], [430.02, 263.99], [438.97, 259.98], [471.04, 260.05], [463.72, 190.04], [430.02, 190.03], [438.97, 189.61], [471.04, 189.62]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[483.11, 250.28], [455.07, 250.29], [460.69, 247.62], [487.64, 247.61], [483.11, 185.86], [455.07, 185.86], [460.69, 185.7], [487.64, 185.7]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[308.37, 270.53], [347.76, 269.62], [342.58, 273.86], [301.22, 274.86], [308.37, 163.84], [347.76, 164.02], [342.58, 163.18], [301.22, 162.98]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[327.73, 267.59], [295.61, 267.51], [308.17, 264.12], [339.02, 264.19], [327.73, 162.34], [295.61, 162.36], [308.17, 163.11], [339.02, 163.1]]}]\n```" - } - ] - }, - { - "id": 1579, - "data_source": "Mono3DRefer", - "scene_name": "001201", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[638.77, 194.61], [684.2, 194.69], [693.64, 201.4], [634.27, 201.25], [638.77, 139.51], [684.2, 139.38], [693.64, 129.1], [634.27, 129.32]]}]\n```" - } - ] - }, - { - "id": 1580, - "data_source": "Mono3DRefer", - "scene_name": "001201", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-2.46, 236.06], [-55.82, 236.15], [39.58, 226.87], [85.0, 226.8], [-2.46, 185.61], [-55.82, 185.63], [39.58, 183.76], [85.0, 183.74]]}, {\"category\": \"car\", \"corners_3d\": [[114.0, 223.84], [67.48, 223.81], [138.7, 217.19], [179.22, 217.22], [114.0, 186.82], [67.48, 186.81], [138.7, 185.0], [179.22, 185.01]]}, {\"category\": \"car\", \"corners_3d\": [[214.46, 211.7], [180.29, 211.7], [226.95, 207.47], [257.4, 207.47], [214.46, 181.64], [180.29, 181.64], [226.95, 180.68], [257.4, 180.68]]}, {\"category\": \"car\", \"corners_3d\": [[272.34, 206.49], [247.22, 206.48], [276.8, 203.78], [299.91, 203.79], [272.34, 180.56], [247.22, 180.55], [276.8, 179.94], [299.91, 179.94]]}, {\"category\": \"car\", \"corners_3d\": [[314.18, 202.91], [288.5, 202.88], [311.62, 200.85], [335.59, 200.88], [314.18, 180.91], [288.5, 180.9], [311.62, 180.35], [335.59, 180.36]]}, {\"category\": \"car\", \"corners_3d\": [[372.03, 197.17], [354.18, 197.12], [371.29, 195.79], [388.2, 195.83], [372.03, 177.49], [354.18, 177.48], [371.29, 177.22], [388.2, 177.23]]}]\n```" - } - ] - }, - { - "id": 1581, - "data_source": "Mono3DRefer", - "scene_name": "001201", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[638.77, 194.61], [684.2, 194.69], [693.64, 201.4], [634.27, 201.25], [638.77, 139.51], [684.2, 139.38], [693.64, 129.1], [634.27, 129.32]]}, {\"category\": \"car\", \"corners_3d\": [[-2.46, 236.06], [-55.82, 236.15], [39.58, 226.87], [85.0, 226.8], [-2.46, 185.61], [-55.82, 185.63], [39.58, 183.76], [85.0, 183.74]]}, {\"category\": \"car\", \"corners_3d\": [[114.0, 223.84], [67.48, 223.81], [138.7, 217.19], [179.22, 217.22], [114.0, 186.82], [67.48, 186.81], [138.7, 185.0], [179.22, 185.01]]}, {\"category\": \"car\", \"corners_3d\": [[214.46, 211.7], [180.29, 211.7], [226.95, 207.47], [257.4, 207.47], [214.46, 181.64], [180.29, 181.64], [226.95, 180.68], [257.4, 180.68]]}, {\"category\": \"car\", \"corners_3d\": [[272.34, 206.49], [247.22, 206.48], [276.8, 203.78], [299.91, 203.79], [272.34, 180.56], [247.22, 180.55], [276.8, 179.94], [299.91, 179.94]]}, {\"category\": \"car\", \"corners_3d\": [[314.18, 202.91], [288.5, 202.88], [311.62, 200.85], [335.59, 200.88], [314.18, 180.91], [288.5, 180.9], [311.62, 180.35], [335.59, 180.36]]}, {\"category\": \"car\", \"corners_3d\": [[372.03, 197.17], [354.18, 197.12], [371.29, 195.79], [388.2, 195.83], [372.03, 177.49], [354.18, 177.48], [371.29, 177.22], [388.2, 177.23]]}]\n```" - } - ] - }, - { - "id": 1582, - "data_source": "Mono3DRefer", - "scene_name": "001206", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[356.19, 264.2], [474.66, 264.06], [379.84, 331.48], [173.32, 331.9], [356.19, 143.69], [474.66, 143.73], [379.84, 122.2], [173.32, 122.07]]}, {\"category\": \"van\", \"corners_3d\": [[478.02, 217.13], [526.35, 217.07], [513.4, 225.32], [456.01, 225.4], [478.02, 167.48], [526.35, 167.49], [513.4, 166.49], [456.01, 166.48]]}, {\"category\": \"van\", \"corners_3d\": [[672.65, 202.71], [712.73, 202.71], [727.53, 207.02], [681.67, 207.01], [672.65, 165.19], [712.73, 165.19], [727.53, 164.08], [681.67, 164.08]]}]\n```" - } - ] - }, - { - "id": 1583, - "data_source": "Mono3DRefer", - "scene_name": "001206", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[574.07, 290.97], [687.95, 290.98], [735.21, 362.72], [552.18, 362.68], [574.07, 184.93], [687.95, 184.93], [735.21, 192.26], [552.18, 192.26]]}, {\"category\": \"car\", \"corners_3d\": [[442.21, 234.61], [507.36, 234.61], [478.59, 251.9], [395.21, 251.89], [442.21, 179.62], [507.36, 179.62], [478.59, 181.51], [395.21, 181.51]]}, {\"category\": \"car\", \"corners_3d\": [[698.0, 213.35], [740.29, 213.35], [762.83, 220.36], [713.21, 220.36], [698.0, 177.18], [740.29, 177.18], [762.83, 177.92], [713.21, 177.92]]}]\n```" - } - ] - }, - { - "id": 1584, - "data_source": "Mono3DRefer", - "scene_name": "001206", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[742.55, 252.32], [865.05, 252.33], [5033.17, 1551.82], [2902.49, 1548.59], [742.55, 88.35], [865.05, 88.34], [5033.17, -1293.48], [2902.49, -1290.05]]}]\n```" - } - ] - }, - { - "id": 1585, - "data_source": "Mono3DRefer", - "scene_name": "001206", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[574.07, 290.97], [687.95, 290.98], [735.21, 362.72], [552.18, 362.68], [574.07, 184.93], [687.95, 184.93], [735.21, 192.26], [552.18, 192.26]]}, {\"category\": \"van\", \"corners_3d\": [[356.19, 264.2], [474.66, 264.06], [379.84, 331.48], [173.32, 331.9], [356.19, 143.69], [474.66, 143.73], [379.84, 122.2], [173.32, 122.07]]}, {\"category\": \"car\", \"corners_3d\": [[442.21, 234.61], [507.36, 234.61], [478.59, 251.9], [395.21, 251.89], [442.21, 179.62], [507.36, 179.62], [478.59, 181.51], [395.21, 181.51]]}, {\"category\": \"van\", \"corners_3d\": [[478.02, 217.13], [526.35, 217.07], [513.4, 225.32], [456.01, 225.4], [478.02, 167.48], [526.35, 167.49], [513.4, 166.49], [456.01, 166.48]]}, {\"category\": \"car\", \"corners_3d\": [[698.0, 213.35], [740.29, 213.35], [762.83, 220.36], [713.21, 220.36], [698.0, 177.18], [740.29, 177.18], [762.83, 177.92], [713.21, 177.92]]}, {\"category\": \"van\", \"corners_3d\": [[672.65, 202.71], [712.73, 202.71], [727.53, 207.02], [681.67, 207.01], [672.65, 165.19], [712.73, 165.19], [727.53, 164.08], [681.67, 164.08]]}]\n```" - } - ] - }, - { - "id": 1586, - "data_source": "Mono3DRefer", - "scene_name": "001206", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[742.55, 252.32], [865.05, 252.33], [5033.17, 1551.82], [2902.49, 1548.59], [742.55, 88.35], [865.05, 88.34], [5033.17, -1293.48], [2902.49, -1290.05]]}, {\"category\": \"van\", \"corners_3d\": [[356.19, 264.2], [474.66, 264.06], [379.84, 331.48], [173.32, 331.9], [356.19, 143.69], [474.66, 143.73], [379.84, 122.2], [173.32, 122.07]]}, {\"category\": \"van\", \"corners_3d\": [[478.02, 217.13], [526.35, 217.07], [513.4, 225.32], [456.01, 225.4], [478.02, 167.48], [526.35, 167.49], [513.4, 166.49], [456.01, 166.48]]}, {\"category\": \"van\", \"corners_3d\": [[672.65, 202.71], [712.73, 202.71], [727.53, 207.02], [681.67, 207.01], [672.65, 165.19], [712.73, 165.19], [727.53, 164.08], [681.67, 164.08]]}]\n```" - } - ] - }, - { - "id": 1587, - "data_source": "Mono3DRefer", - "scene_name": "001206", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[574.07, 290.97], [687.95, 290.98], [735.21, 362.72], [552.18, 362.68], [574.07, 184.93], [687.95, 184.93], [735.21, 192.26], [552.18, 192.26]]}, {\"category\": \"truck\", \"corners_3d\": [[742.55, 252.32], [865.05, 252.33], [5033.17, 1551.82], [2902.49, 1548.59], [742.55, 88.35], [865.05, 88.34], [5033.17, -1293.48], [2902.49, -1290.05]]}, {\"category\": \"car\", \"corners_3d\": [[442.21, 234.61], [507.36, 234.61], [478.59, 251.9], [395.21, 251.89], [442.21, 179.62], [507.36, 179.62], [478.59, 181.51], [395.21, 181.51]]}, {\"category\": \"car\", \"corners_3d\": [[698.0, 213.35], [740.29, 213.35], [762.83, 220.36], [713.21, 220.36], [698.0, 177.18], [740.29, 177.18], [762.83, 177.92], [713.21, 177.92]]}]\n```" - } - ] - }, - { - "id": 1588, - "data_source": "Mono3DRefer", - "scene_name": "001206", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[574.07, 290.97], [687.95, 290.98], [735.21, 362.72], [552.18, 362.68], [574.07, 184.93], [687.95, 184.93], [735.21, 192.26], [552.18, 192.26]]}, {\"category\": \"truck\", \"corners_3d\": [[742.55, 252.32], [865.05, 252.33], [5033.17, 1551.82], [2902.49, 1548.59], [742.55, 88.35], [865.05, 88.34], [5033.17, -1293.48], [2902.49, -1290.05]]}, {\"category\": \"van\", \"corners_3d\": [[356.19, 264.2], [474.66, 264.06], [379.84, 331.48], [173.32, 331.9], [356.19, 143.69], [474.66, 143.73], [379.84, 122.2], [173.32, 122.07]]}, {\"category\": \"car\", \"corners_3d\": [[442.21, 234.61], [507.36, 234.61], [478.59, 251.9], [395.21, 251.89], [442.21, 179.62], [507.36, 179.62], [478.59, 181.51], [395.21, 181.51]]}, {\"category\": \"van\", \"corners_3d\": [[478.02, 217.13], [526.35, 217.07], [513.4, 225.32], [456.01, 225.4], [478.02, 167.48], [526.35, 167.49], [513.4, 166.49], [456.01, 166.48]]}, {\"category\": \"car\", \"corners_3d\": [[698.0, 213.35], [740.29, 213.35], [762.83, 220.36], [713.21, 220.36], [698.0, 177.18], [740.29, 177.18], [762.83, 177.92], [713.21, 177.92]]}, {\"category\": \"van\", \"corners_3d\": [[672.65, 202.71], [712.73, 202.71], [727.53, 207.02], [681.67, 207.01], [672.65, 165.19], [712.73, 165.19], [727.53, 164.08], [681.67, 164.08]]}]\n```" - } - ] - }, - { - "id": 1589, - "data_source": "Mono3DRefer", - "scene_name": "001208", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[282.69, 313.2], [259.65, 324.76], [158.16, 326.35], [189.65, 314.55], [282.69, 162.87], [259.65, 161.34], [158.16, 161.12], [189.65, 162.69]]}]\n```" - } - ] - }, - { - "id": 1590, - "data_source": "Mono3DRefer", - "scene_name": "001215", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[484.56, 280.34], [484.36, 298.34], [120.68, 308.88], [176.76, 287.98], [484.56, 189.71], [484.36, 192.53], [120.68, 194.18], [176.76, 190.91]]}]\n```" - } - ] - }, - { - "id": 1591, - "data_source": "Mono3DRefer", - "scene_name": "001217", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[559.96, 207.88], [529.03, 207.8], [539.09, 204.77], [567.35, 204.83], [559.96, 176.57], [529.03, 176.56], [539.09, 176.24], [567.35, 176.25]]}, {\"category\": \"van\", \"corners_3d\": [[626.86, 202.69], [652.23, 202.72], [654.03, 205.35], [626.42, 205.32], [626.86, 175.18], [652.23, 175.18], [654.03, 175.39], [626.42, 175.38]]}]\n```" - } - ] - }, - { - "id": 1592, - "data_source": "Mono3DRefer", - "scene_name": "001217", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[361.05, 287.14], [469.05, 287.5], [390.42, 345.46], [228.45, 344.65], [361.05, 196.21], [469.05, 196.28], [390.42, 208.13], [228.45, 207.97]]}, {\"category\": \"car\", \"corners_3d\": [[687.06, 230.7], [639.37, 230.82], [632.43, 223.08], [673.77, 222.99], [687.06, 186.49], [639.37, 186.52], [632.43, 184.7], [673.77, 184.67]]}]\n```" - } - ] - }, - { - "id": 1593, - "data_source": "Mono3DRefer", - "scene_name": "001217", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[361.05, 287.14], [469.05, 287.5], [390.42, 345.46], [228.45, 344.65], [361.05, 196.21], [469.05, 196.28], [390.42, 208.13], [228.45, 207.97]]}, {\"category\": \"car\", \"corners_3d\": [[687.06, 230.7], [639.37, 230.82], [632.43, 223.08], [673.77, 222.99], [687.06, 186.49], [639.37, 186.52], [632.43, 184.7], [673.77, 184.67]]}, {\"category\": \"van\", \"corners_3d\": [[559.96, 207.88], [529.03, 207.8], [539.09, 204.77], [567.35, 204.83], [559.96, 176.57], [529.03, 176.56], [539.09, 176.24], [567.35, 176.25]]}, {\"category\": \"van\", \"corners_3d\": [[626.86, 202.69], [652.23, 202.72], [654.03, 205.35], [626.42, 205.32], [626.86, 175.18], [652.23, 175.18], [654.03, 175.39], [626.42, 175.38]]}]\n```" - } - ] - }, - { - "id": 1594, - "data_source": "Mono3DRefer", - "scene_name": "001218", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[252.96, 233.65], [200.93, 233.57], [257.0, 225.51], [302.18, 225.58], [252.96, 190.95], [200.93, 190.92], [257.0, 188.52], [302.18, 188.54]]}, {\"category\": \"car\", \"corners_3d\": [[321.95, 213.91], [288.66, 213.91], [319.36, 209.98], [349.46, 209.98], [321.95, 187.19], [288.66, 187.19], [319.36, 185.81], [349.46, 185.81]]}]\n```" - } - ] - }, - { - "id": 1595, - "data_source": "Mono3DRefer", - "scene_name": "001218", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[844.99, 224.74], [864.9, 224.69], [891.07, 229.6], [869.3, 229.65], [844.99, 159.78], [864.9, 159.79], [891.07, 158.55], [869.3, 158.54]]}]\n```" - } - ] - }, - { - "id": 1596, - "data_source": "Mono3DRefer", - "scene_name": "001218", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[252.96, 233.65], [200.93, 233.57], [257.0, 225.51], [302.18, 225.58], [252.96, 190.95], [200.93, 190.92], [257.0, 188.52], [302.18, 188.54]]}, {\"category\": \"car\", \"corners_3d\": [[321.95, 213.91], [288.66, 213.91], [319.36, 209.98], [349.46, 209.98], [321.95, 187.19], [288.66, 187.19], [319.36, 185.81], [349.46, 185.81]]}, {\"category\": \"cyclist\", \"corners_3d\": [[844.99, 224.74], [864.9, 224.69], [891.07, 229.6], [869.3, 229.65], [844.99, 159.78], [864.9, 159.79], [891.07, 158.55], [869.3, 158.54]]}]\n```" - } - ] - }, - { - "id": 1597, - "data_source": "Mono3DRefer", - "scene_name": "001220", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[808.41, 244.71], [824.62, 244.32], [879.56, 253.25], [861.44, 253.75], [808.41, 149.06], [824.62, 149.25], [879.56, 144.88], [861.44, 144.63]]}]\n```" - } - ] - }, - { - "id": 1598, - "data_source": "Mono3DRefer", - "scene_name": "001220", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[607.63, 179.68], [624.08, 179.68], [625.3, 179.62], [607.82, 179.62], [607.63, 161.05], [624.08, 161.05], [625.3, 159.83], [607.82, 159.83]]}]\n```" - } - ] - }, - { - "id": 1599, - "data_source": "Mono3DRefer", - "scene_name": "001220", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1108.09, 228.83], [1090.1, 226.97], [1115.72, 227.05], [1134.78, 228.92], [1108.09, 150.78], [1090.1, 151.92], [1115.72, 151.87], [1134.78, 150.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1139.61, 232.4], [1121.25, 230.37], [1156.14, 230.59], [1176.0, 232.63], [1139.61, 143.1], [1121.25, 144.56], [1156.14, 144.4], [1176.0, 142.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1125.39, 227.14], [1112.85, 225.99], [1139.56, 226.02], [1152.78, 227.18], [1125.39, 153.93], [1112.85, 154.59], [1139.56, 154.56], [1152.78, 153.91]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[971.74, 330.97], [882.39, 328.32], [862.89, 311.57], [941.64, 313.65], [971.74, 131.57], [882.39, 132.43], [862.89, 137.88], [941.64, 137.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[889.86, 210.14], [912.39, 210.68], [908.5, 211.24], [885.62, 210.69], [889.86, 153.29], [912.39, 152.8], [908.5, 152.28], [885.62, 152.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[665.94, 263.22], [639.24, 260.99], [669.01, 256.22], [694.85, 258.18], [665.94, 155.74], [639.24, 156.4], [669.01, 157.83], [694.85, 157.24]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[596.25, 256.03], [577.91, 253.68], [623.22, 249.84], [641.94, 251.94], [596.25, 152.33], [577.91, 153.2], [623.22, 154.64], [641.94, 153.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[612.03, 260.79], [589.49, 257.81], [631.24, 254.01], [654.17, 256.71], [612.03, 145.2], [589.49, 146.51], [631.24, 148.18], [654.17, 146.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[886.31, 222.76], [880.3, 221.3], [922.66, 221.66], [930.26, 223.14], [886.31, 153.68], [880.3, 154.61], [922.66, 154.38], [930.26, 153.44]]}]\n```" - } - ] - }, - { - "id": 1600, - "data_source": "Mono3DRefer", - "scene_name": "001220", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[607.63, 179.68], [624.08, 179.68], [625.3, 179.62], [607.82, 179.62], [607.63, 161.05], [624.08, 161.05], [625.3, 159.83], [607.82, 159.83]]}, {\"category\": \"cyclist\", \"corners_3d\": [[808.41, 244.71], [824.62, 244.32], [879.56, 253.25], [861.44, 253.75], [808.41, 149.06], [824.62, 149.25], [879.56, 144.88], [861.44, 144.63]]}]\n```" - } - ] - }, - { - "id": 1601, - "data_source": "Mono3DRefer", - "scene_name": "001220", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1108.09, 228.83], [1090.1, 226.97], [1115.72, 227.05], [1134.78, 228.92], [1108.09, 150.78], [1090.1, 151.92], [1115.72, 151.87], [1134.78, 150.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1139.61, 232.4], [1121.25, 230.37], [1156.14, 230.59], [1176.0, 232.63], [1139.61, 143.1], [1121.25, 144.56], [1156.14, 144.4], [1176.0, 142.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1125.39, 227.14], [1112.85, 225.99], [1139.56, 226.02], [1152.78, 227.18], [1125.39, 153.93], [1112.85, 154.59], [1139.56, 154.56], [1152.78, 153.91]]}, {\"category\": \"cyclist\", \"corners_3d\": [[808.41, 244.71], [824.62, 244.32], [879.56, 253.25], [861.44, 253.75], [808.41, 149.06], [824.62, 149.25], [879.56, 144.88], [861.44, 144.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[971.74, 330.97], [882.39, 328.32], [862.89, 311.57], [941.64, 313.65], [971.74, 131.57], [882.39, 132.43], [862.89, 137.88], [941.64, 137.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[889.86, 210.14], [912.39, 210.68], [908.5, 211.24], [885.62, 210.69], [889.86, 153.29], [912.39, 152.8], [908.5, 152.28], [885.62, 152.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[665.94, 263.22], [639.24, 260.99], [669.01, 256.22], [694.85, 258.18], [665.94, 155.74], [639.24, 156.4], [669.01, 157.83], [694.85, 157.24]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[596.25, 256.03], [577.91, 253.68], [623.22, 249.84], [641.94, 251.94], [596.25, 152.33], [577.91, 153.2], [623.22, 154.64], [641.94, 153.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[612.03, 260.79], [589.49, 257.81], [631.24, 254.01], [654.17, 256.71], [612.03, 145.2], [589.49, 146.51], [631.24, 148.18], [654.17, 146.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[886.31, 222.76], [880.3, 221.3], [922.66, 221.66], [930.26, 223.14], [886.31, 153.68], [880.3, 154.61], [922.66, 154.38], [930.26, 153.44]]}]\n```" - } - ] - }, - { - "id": 1602, - "data_source": "Mono3DRefer", - "scene_name": "001220", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1108.09, 228.83], [1090.1, 226.97], [1115.72, 227.05], [1134.78, 228.92], [1108.09, 150.78], [1090.1, 151.92], [1115.72, 151.87], [1134.78, 150.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1139.61, 232.4], [1121.25, 230.37], [1156.14, 230.59], [1176.0, 232.63], [1139.61, 143.1], [1121.25, 144.56], [1156.14, 144.4], [1176.0, 142.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1125.39, 227.14], [1112.85, 225.99], [1139.56, 226.02], [1152.78, 227.18], [1125.39, 153.93], [1112.85, 154.59], [1139.56, 154.56], [1152.78, 153.91]]}, {\"category\": \"van\", \"corners_3d\": [[607.63, 179.68], [624.08, 179.68], [625.3, 179.62], [607.82, 179.62], [607.63, 161.05], [624.08, 161.05], [625.3, 159.83], [607.82, 159.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[971.74, 330.97], [882.39, 328.32], [862.89, 311.57], [941.64, 313.65], [971.74, 131.57], [882.39, 132.43], [862.89, 137.88], [941.64, 137.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[889.86, 210.14], [912.39, 210.68], [908.5, 211.24], [885.62, 210.69], [889.86, 153.29], [912.39, 152.8], [908.5, 152.28], [885.62, 152.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[665.94, 263.22], [639.24, 260.99], [669.01, 256.22], [694.85, 258.18], [665.94, 155.74], [639.24, 156.4], [669.01, 157.83], [694.85, 157.24]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[596.25, 256.03], [577.91, 253.68], [623.22, 249.84], [641.94, 251.94], [596.25, 152.33], [577.91, 153.2], [623.22, 154.64], [641.94, 153.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[612.03, 260.79], [589.49, 257.81], [631.24, 254.01], [654.17, 256.71], [612.03, 145.2], [589.49, 146.51], [631.24, 148.18], [654.17, 146.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[886.31, 222.76], [880.3, 221.3], [922.66, 221.66], [930.26, 223.14], [886.31, 153.68], [880.3, 154.61], [922.66, 154.38], [930.26, 153.44]]}]\n```" - } - ] - }, - { - "id": 1603, - "data_source": "Mono3DRefer", - "scene_name": "001220", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1108.09, 228.83], [1090.1, 226.97], [1115.72, 227.05], [1134.78, 228.92], [1108.09, 150.78], [1090.1, 151.92], [1115.72, 151.87], [1134.78, 150.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1139.61, 232.4], [1121.25, 230.37], [1156.14, 230.59], [1176.0, 232.63], [1139.61, 143.1], [1121.25, 144.56], [1156.14, 144.4], [1176.0, 142.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1125.39, 227.14], [1112.85, 225.99], [1139.56, 226.02], [1152.78, 227.18], [1125.39, 153.93], [1112.85, 154.59], [1139.56, 154.56], [1152.78, 153.91]]}, {\"category\": \"van\", \"corners_3d\": [[607.63, 179.68], [624.08, 179.68], [625.3, 179.62], [607.82, 179.62], [607.63, 161.05], [624.08, 161.05], [625.3, 159.83], [607.82, 159.83]]}, {\"category\": \"cyclist\", \"corners_3d\": [[808.41, 244.71], [824.62, 244.32], [879.56, 253.25], [861.44, 253.75], [808.41, 149.06], [824.62, 149.25], [879.56, 144.88], [861.44, 144.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[971.74, 330.97], [882.39, 328.32], [862.89, 311.57], [941.64, 313.65], [971.74, 131.57], [882.39, 132.43], [862.89, 137.88], [941.64, 137.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[889.86, 210.14], [912.39, 210.68], [908.5, 211.24], [885.62, 210.69], [889.86, 153.29], [912.39, 152.8], [908.5, 152.28], [885.62, 152.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[665.94, 263.22], [639.24, 260.99], [669.01, 256.22], [694.85, 258.18], [665.94, 155.74], [639.24, 156.4], [669.01, 157.83], [694.85, 157.24]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[596.25, 256.03], [577.91, 253.68], [623.22, 249.84], [641.94, 251.94], [596.25, 152.33], [577.91, 153.2], [623.22, 154.64], [641.94, 153.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[612.03, 260.79], [589.49, 257.81], [631.24, 254.01], [654.17, 256.71], [612.03, 145.2], [589.49, 146.51], [631.24, 148.18], [654.17, 146.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[886.31, 222.76], [880.3, 221.3], [922.66, 221.66], [930.26, 223.14], [886.31, 153.68], [880.3, 154.61], [922.66, 154.38], [930.26, 153.44]]}]\n```" - } - ] - }, - { - "id": 1604, - "data_source": "Mono3DRefer", - "scene_name": "001223", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[753.12, 232.19], [735.5, 226.84], [906.47, 225.91], [940.38, 231.07], [753.12, 160.89], [735.5, 161.97], [906.47, 162.15], [940.38, 161.11]]}, {\"category\": \"car\", \"corners_3d\": [[1138.19, 181.5], [1167.0, 181.95], [1074.63, 182.0], [1050.23, 181.55], [1138.19, 148.83], [1167.0, 147.59], [1074.63, 147.45], [1050.23, 148.71]]}, {\"category\": \"car\", \"corners_3d\": [[1107.88, 181.13], [1132.6, 181.52], [1041.39, 181.57], [1020.65, 181.17], [1107.88, 150.41], [1132.6, 149.36], [1041.39, 149.24], [1020.65, 150.3]]}, {\"category\": \"car\", \"corners_3d\": [[1083.76, 182.49], [1106.46, 182.96], [1013.99, 182.94], [995.59, 182.48], [1083.76, 151.87], [1106.46, 150.85], [1013.99, 150.87], [995.59, 151.89]]}, {\"category\": \"car\", \"corners_3d\": [[501.03, 227.6], [451.66, 227.9], [465.42, 220.29], [507.93, 220.06], [501.03, 183.91], [451.66, 183.97], [465.42, 182.43], [507.93, 182.38]]}, {\"category\": \"car\", \"corners_3d\": [[575.19, 209.96], [538.31, 209.94], [546.72, 205.93], [579.62, 205.95], [575.19, 177.56], [538.31, 177.56], [546.72, 177.05], [579.62, 177.05]]}, {\"category\": \"car\", \"corners_3d\": [[855.39, 196.52], [866.83, 197.69], [771.86, 197.62], [764.91, 196.46], [855.39, 160.25], [866.83, 159.62], [771.86, 159.66], [764.91, 160.28]]}, {\"category\": \"car\", \"corners_3d\": [[891.93, 180.05], [916.38, 180.05], [942.49, 180.62], [916.12, 180.63], [891.93, 153.04], [916.38, 153.06], [942.49, 151.48], [916.12, 151.46]]}, {\"category\": \"car\", \"corners_3d\": [[925.74, 180.85], [950.83, 180.85], [981.2, 181.55], [953.94, 181.55], [925.74, 156.54], [950.83, 156.55], [981.2, 155.13], [953.94, 155.12]]}, {\"category\": \"car\", \"corners_3d\": [[512.86, 219.71], [474.87, 219.72], [491.19, 214.01], [524.56, 214.01], [512.86, 185.29], [474.87, 185.29], [491.19, 183.77], [524.56, 183.77]]}, {\"category\": \"car\", \"corners_3d\": [[684.7, 199.23], [656.82, 199.28], [649.12, 196.86], [674.47, 196.81], [684.7, 174.65], [656.82, 174.65], [649.12, 174.49], [674.47, 174.49]]}, {\"category\": \"car\", \"corners_3d\": [[610.14, 198.03], [630.8, 198.1], [627.33, 199.79], [605.3, 199.71], [610.14, 178.9], [630.8, 178.92], [627.33, 179.33], [605.3, 179.31]]}]\n```" - } - ] - }, - { - "id": 1605, - "data_source": "Mono3DRefer", - "scene_name": "001223", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[988.81, 179.18], [1016.34, 179.21], [1045.53, 179.82], [1015.21, 179.78], [988.81, 148.13], [1016.34, 148.01], [1045.53, 145.62], [1015.21, 145.77]]}]\n```" - } - ] - }, - { - "id": 1606, - "data_source": "Mono3DRefer", - "scene_name": "001223", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[753.12, 232.19], [735.5, 226.84], [906.47, 225.91], [940.38, 231.07], [753.12, 160.89], [735.5, 161.97], [906.47, 162.15], [940.38, 161.11]]}, {\"category\": \"car\", \"corners_3d\": [[1138.19, 181.5], [1167.0, 181.95], [1074.63, 182.0], [1050.23, 181.55], [1138.19, 148.83], [1167.0, 147.59], [1074.63, 147.45], [1050.23, 148.71]]}, {\"category\": \"car\", \"corners_3d\": [[1107.88, 181.13], [1132.6, 181.52], [1041.39, 181.57], [1020.65, 181.17], [1107.88, 150.41], [1132.6, 149.36], [1041.39, 149.24], [1020.65, 150.3]]}, {\"category\": \"car\", \"corners_3d\": [[1083.76, 182.49], [1106.46, 182.96], [1013.99, 182.94], [995.59, 182.48], [1083.76, 151.87], [1106.46, 150.85], [1013.99, 150.87], [995.59, 151.89]]}, {\"category\": \"car\", \"corners_3d\": [[501.03, 227.6], [451.66, 227.9], [465.42, 220.29], [507.93, 220.06], [501.03, 183.91], [451.66, 183.97], [465.42, 182.43], [507.93, 182.38]]}, {\"category\": \"car\", \"corners_3d\": [[575.19, 209.96], [538.31, 209.94], [546.72, 205.93], [579.62, 205.95], [575.19, 177.56], [538.31, 177.56], [546.72, 177.05], [579.62, 177.05]]}, {\"category\": \"car\", \"corners_3d\": [[855.39, 196.52], [866.83, 197.69], [771.86, 197.62], [764.91, 196.46], [855.39, 160.25], [866.83, 159.62], [771.86, 159.66], [764.91, 160.28]]}, {\"category\": \"car\", \"corners_3d\": [[891.93, 180.05], [916.38, 180.05], [942.49, 180.62], [916.12, 180.63], [891.93, 153.04], [916.38, 153.06], [942.49, 151.48], [916.12, 151.46]]}, {\"category\": \"van\", \"corners_3d\": [[988.81, 179.18], [1016.34, 179.21], [1045.53, 179.82], [1015.21, 179.78], [988.81, 148.13], [1016.34, 148.01], [1045.53, 145.62], [1015.21, 145.77]]}, {\"category\": \"car\", \"corners_3d\": [[925.74, 180.85], [950.83, 180.85], [981.2, 181.55], [953.94, 181.55], [925.74, 156.54], [950.83, 156.55], [981.2, 155.13], [953.94, 155.12]]}, {\"category\": \"car\", \"corners_3d\": [[512.86, 219.71], [474.87, 219.72], [491.19, 214.01], [524.56, 214.01], [512.86, 185.29], [474.87, 185.29], [491.19, 183.77], [524.56, 183.77]]}, {\"category\": \"car\", \"corners_3d\": [[684.7, 199.23], [656.82, 199.28], [649.12, 196.86], [674.47, 196.81], [684.7, 174.65], [656.82, 174.65], [649.12, 174.49], [674.47, 174.49]]}, {\"category\": \"car\", \"corners_3d\": [[610.14, 198.03], [630.8, 198.1], [627.33, 199.79], [605.3, 199.71], [610.14, 178.9], [630.8, 178.92], [627.33, 179.33], [605.3, 179.31]]}]\n```" - } - ] - }, - { - "id": 1607, - "data_source": "Mono3DRefer", - "scene_name": "001224", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[89.37, 414.22], [-142.89, 438.35], [44.5, 331.03], [177.62, 322.14], [89.37, 247.97], [-142.89, 255.48], [44.5, 222.08], [177.62, 219.31]]}, {\"category\": \"car\", \"corners_3d\": [[404.42, 235.38], [441.04, 234.54], [446.47, 243.28], [404.67, 244.38], [404.42, 205.0], [441.04, 204.57], [446.47, 209.06], [404.67, 209.63]]}, {\"category\": \"car\", \"corners_3d\": [[307.22, 230.26], [278.89, 231.02], [282.84, 226.44], [308.91, 225.8], [307.22, 205.13], [278.89, 205.56], [282.84, 202.98], [308.91, 202.62]]}, {\"category\": \"car\", \"corners_3d\": [[682.01, 364.66], [827.35, 350.99], [1236.9, 508.12], [1006.93, 560.06], [682.01, 205.33], [827.35, 203.02], [1236.9, 229.62], [1006.93, 238.42]]}, {\"category\": \"car\", \"corners_3d\": [[552.56, 291.27], [651.25, 286.05], [797.81, 345.26], [654.31, 357.66], [552.56, 205.87], [651.25, 204.41], [797.81, 220.92], [654.31, 224.38]]}, {\"category\": \"car\", \"corners_3d\": [[499.5, 270.45], [568.78, 267.41], [629.8, 295.45], [541.67, 300.62], [499.5, 205.99], [568.78, 204.96], [629.8, 214.48], [541.67, 216.23]]}, {\"category\": \"car\", \"corners_3d\": [[440.8, 253.97], [498.42, 251.85], [527.49, 267.87], [458.76, 270.94], [440.8, 197.78], [498.42, 197.13], [527.49, 202.05], [458.76, 202.99]]}, {\"category\": \"car\", \"corners_3d\": [[271.87, 260.9], [214.2, 263.15], [232.54, 249.7], [281.41, 248.06], [271.87, 202.61], [214.2, 203.37], [232.54, 198.83], [281.41, 198.27]]}, {\"category\": \"car\", \"corners_3d\": [[210.24, 303.35], [113.42, 309.09], [155.74, 281.27], [232.03, 277.6], [210.24, 222.13], [113.42, 224.29], [155.74, 213.79], [232.03, 212.41]]}]\n```" - } - ] - }, - { - "id": 1608, - "data_source": "Mono3DRefer", - "scene_name": "001225", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[495.18, 212.72], [473.94, 211.01], [554.75, 207.89], [577.5, 209.33], [495.18, 176.54], [473.94, 176.38], [554.75, 176.09], [577.5, 176.22]]}, {\"category\": \"car\", \"corners_3d\": [[339.16, 327.88], [171.66, 334.04], [277.14, 278.48], [385.69, 275.8], [339.16, 200.61], [171.66, 201.71], [277.14, 191.76], [385.69, 191.28]]}, {\"category\": \"car\", \"corners_3d\": [[640.72, 206.44], [622.97, 204.8], [706.76, 203.08], [727.58, 204.54], [640.72, 174.16], [622.97, 174.09], [706.76, 174.02], [727.58, 174.08]]}, {\"category\": \"car\", \"corners_3d\": [[410.6, 223.03], [395.93, 226.6], [268.07, 226.51], [291.19, 222.95], [410.6, 176.81], [395.93, 177.1], [268.07, 177.09], [291.19, 176.81]]}]\n```" - } - ] - }, - { - "id": 1609, - "data_source": "Mono3DRefer", - "scene_name": "001225", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[90.22, 220.14], [119.0, 217.63], [227.48, 217.85], [204.66, 220.38], [90.22, 160.53], [119.0, 161.18], [227.48, 161.12], [204.66, 160.46]]}]\n```" - } - ] - }, - { - "id": 1610, - "data_source": "Mono3DRefer", - "scene_name": "001225", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[986.85, 281.66], [936.47, 281.67], [911.08, 273.24], [957.56, 273.23], [986.85, 158.22], [936.47, 158.22], [911.08, 159.35], [957.56, 159.35]]}]\n```" - } - ] - }, - { - "id": 1611, - "data_source": "Mono3DRefer", - "scene_name": "001225", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[495.18, 212.72], [473.94, 211.01], [554.75, 207.89], [577.5, 209.33], [495.18, 176.54], [473.94, 176.38], [554.75, 176.09], [577.5, 176.22]]}, {\"category\": \"car\", \"corners_3d\": [[339.16, 327.88], [171.66, 334.04], [277.14, 278.48], [385.69, 275.8], [339.16, 200.61], [171.66, 201.71], [277.14, 191.76], [385.69, 191.28]]}, {\"category\": \"car\", \"corners_3d\": [[640.72, 206.44], [622.97, 204.8], [706.76, 203.08], [727.58, 204.54], [640.72, 174.16], [622.97, 174.09], [706.76, 174.02], [727.58, 174.08]]}, {\"category\": \"truck\", \"corners_3d\": [[90.22, 220.14], [119.0, 217.63], [227.48, 217.85], [204.66, 220.38], [90.22, 160.53], [119.0, 161.18], [227.48, 161.12], [204.66, 160.46]]}, {\"category\": \"car\", \"corners_3d\": [[410.6, 223.03], [395.93, 226.6], [268.07, 226.51], [291.19, 222.95], [410.6, 176.81], [395.93, 177.1], [268.07, 177.09], [291.19, 176.81]]}]\n```" - } - ] - }, - { - "id": 1612, - "data_source": "Mono3DRefer", - "scene_name": "001225", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[986.85, 281.66], [936.47, 281.67], [911.08, 273.24], [957.56, 273.23], [986.85, 158.22], [936.47, 158.22], [911.08, 159.35], [957.56, 159.35]]}, {\"category\": \"car\", \"corners_3d\": [[495.18, 212.72], [473.94, 211.01], [554.75, 207.89], [577.5, 209.33], [495.18, 176.54], [473.94, 176.38], [554.75, 176.09], [577.5, 176.22]]}, {\"category\": \"car\", \"corners_3d\": [[339.16, 327.88], [171.66, 334.04], [277.14, 278.48], [385.69, 275.8], [339.16, 200.61], [171.66, 201.71], [277.14, 191.76], [385.69, 191.28]]}, {\"category\": \"car\", \"corners_3d\": [[640.72, 206.44], [622.97, 204.8], [706.76, 203.08], [727.58, 204.54], [640.72, 174.16], [622.97, 174.09], [706.76, 174.02], [727.58, 174.08]]}, {\"category\": \"car\", \"corners_3d\": [[410.6, 223.03], [395.93, 226.6], [268.07, 226.51], [291.19, 222.95], [410.6, 176.81], [395.93, 177.1], [268.07, 177.09], [291.19, 176.81]]}]\n```" - } - ] - }, - { - "id": 1613, - "data_source": "Mono3DRefer", - "scene_name": "001225", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[986.85, 281.66], [936.47, 281.67], [911.08, 273.24], [957.56, 273.23], [986.85, 158.22], [936.47, 158.22], [911.08, 159.35], [957.56, 159.35]]}, {\"category\": \"truck\", \"corners_3d\": [[90.22, 220.14], [119.0, 217.63], [227.48, 217.85], [204.66, 220.38], [90.22, 160.53], [119.0, 161.18], [227.48, 161.12], [204.66, 160.46]]}]\n```" - } - ] - }, - { - "id": 1614, - "data_source": "Mono3DRefer", - "scene_name": "001225", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[986.85, 281.66], [936.47, 281.67], [911.08, 273.24], [957.56, 273.23], [986.85, 158.22], [936.47, 158.22], [911.08, 159.35], [957.56, 159.35]]}, {\"category\": \"car\", \"corners_3d\": [[495.18, 212.72], [473.94, 211.01], [554.75, 207.89], [577.5, 209.33], [495.18, 176.54], [473.94, 176.38], [554.75, 176.09], [577.5, 176.22]]}, {\"category\": \"car\", \"corners_3d\": [[339.16, 327.88], [171.66, 334.04], [277.14, 278.48], [385.69, 275.8], [339.16, 200.61], [171.66, 201.71], [277.14, 191.76], [385.69, 191.28]]}, {\"category\": \"car\", \"corners_3d\": [[640.72, 206.44], [622.97, 204.8], [706.76, 203.08], [727.58, 204.54], [640.72, 174.16], [622.97, 174.09], [706.76, 174.02], [727.58, 174.08]]}, {\"category\": \"truck\", \"corners_3d\": [[90.22, 220.14], [119.0, 217.63], [227.48, 217.85], [204.66, 220.38], [90.22, 160.53], [119.0, 161.18], [227.48, 161.12], [204.66, 160.46]]}, {\"category\": \"car\", \"corners_3d\": [[410.6, 223.03], [395.93, 226.6], [268.07, 226.51], [291.19, 222.95], [410.6, 176.81], [395.93, 177.1], [268.07, 177.09], [291.19, 176.81]]}]\n```" - } - ] - }, - { - "id": 1615, - "data_source": "Mono3DRefer", - "scene_name": "001229", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[660.08, 191.87], [690.07, 191.82], [708.48, 194.66], [674.05, 194.72], [660.08, 149.93], [690.07, 149.99], [708.48, 146.57], [674.05, 146.5]]}]\n```" - } - ] - }, - { - "id": 1616, - "data_source": "Mono3DRefer", - "scene_name": "001229", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[652.78, 207.49], [685.34, 207.41], [699.55, 211.82], [662.85, 211.92], [652.78, 172.43], [685.34, 172.43], [699.55, 172.38], [662.85, 172.38]]}, {\"category\": \"car\", \"corners_3d\": [[685.5, 220.96], [731.12, 220.9], [751.42, 228.1], [698.98, 228.18], [685.5, 179.84], [731.12, 179.83], [751.42, 180.88], [698.98, 180.89]]}, {\"category\": \"car\", \"corners_3d\": [[799.35, 219.9], [845.93, 219.81], [898.56, 229.41], [842.55, 229.54], [799.35, 177.52], [845.93, 177.51], [898.56, 178.46], [842.55, 178.48]]}, {\"category\": \"car\", \"corners_3d\": [[802.16, 200.85], [817.84, 200.36], [868.43, 202.37], [852.56, 202.95], [802.16, 175.43], [817.84, 175.38], [868.43, 175.56], [852.56, 175.62]]}]\n```" - } - ] - }, - { - "id": 1617, - "data_source": "Mono3DRefer", - "scene_name": "001229", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[652.78, 207.49], [685.34, 207.41], [699.55, 211.82], [662.85, 211.92], [652.78, 172.43], [685.34, 172.43], [699.55, 172.38], [662.85, 172.38]]}, {\"category\": \"car\", \"corners_3d\": [[685.5, 220.96], [731.12, 220.9], [751.42, 228.1], [698.98, 228.18], [685.5, 179.84], [731.12, 179.83], [751.42, 180.88], [698.98, 180.89]]}, {\"category\": \"truck\", \"corners_3d\": [[660.08, 191.87], [690.07, 191.82], [708.48, 194.66], [674.05, 194.72], [660.08, 149.93], [690.07, 149.99], [708.48, 146.57], [674.05, 146.5]]}, {\"category\": \"car\", \"corners_3d\": [[799.35, 219.9], [845.93, 219.81], [898.56, 229.41], [842.55, 229.54], [799.35, 177.52], [845.93, 177.51], [898.56, 178.46], [842.55, 178.48]]}, {\"category\": \"car\", \"corners_3d\": [[802.16, 200.85], [817.84, 200.36], [868.43, 202.37], [852.56, 202.95], [802.16, 175.43], [817.84, 175.38], [868.43, 175.56], [852.56, 175.62]]}]\n```" - } - ] - }, - { - "id": 1618, - "data_source": "Mono3DRefer", - "scene_name": "001235", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[659.71, 210.79], [687.6, 210.79], [696.84, 215.32], [665.63, 215.32], [659.71, 185.91], [687.6, 185.91], [696.84, 187.47], [665.63, 187.47]]}]\n```" - } - ] - }, - { - "id": 1619, - "data_source": "Mono3DRefer", - "scene_name": "001238", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[376.56, 243.35], [390.72, 241.67], [432.46, 244.0], [418.87, 245.8], [376.56, 190.17], [390.72, 189.75], [432.46, 190.33], [418.87, 190.77]]}, {\"category\": \"cyclist\", \"corners_3d\": [[385.6, 277.35], [371.18, 281.78], [276.04, 279.33], [293.94, 275.09], [385.6, 183.59], [371.18, 184.05], [276.04, 183.8], [293.94, 183.36]]}]\n```" - } - ] - }, - { - "id": 1620, - "data_source": "Mono3DRefer", - "scene_name": "001238", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[182.34, 314.15], [240.4, 308.03], [257.46, 319.47], [195.09, 326.7], [182.34, 171.23], [240.4, 171.3], [257.46, 171.17], [195.09, 171.08]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[389.6, 289.03], [360.6, 294.28], [324.65, 289.83], [354.14, 284.96], [389.6, 172.83], [360.6, 172.83], [324.65, 172.83], [354.14, 172.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[412.64, 287.12], [375.29, 291.88], [349.63, 285.89], [386.14, 281.59], [412.64, 170.32], [375.29, 170.21], [349.63, 170.35], [386.14, 170.44]]}]\n```" - } - ] - }, - { - "id": 1621, - "data_source": "Mono3DRefer", - "scene_name": "001238", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-226.86, 495.91], [-599.0, 505.38], [49.98, 320.09], [209.09, 318.2], [-226.86, 235.51], [-599.0, 237.35], [49.98, 201.41], [209.09, 201.04]]}, {\"category\": \"car\", \"corners_3d\": [[1194.84, 220.84], [1137.67, 217.55], [1237.5, 214.7], [1297.64, 217.58], [1194.84, 165.77], [1137.67, 166.26], [1237.5, 166.68], [1297.64, 166.25]]}]\n```" - } - ] - }, - { - "id": 1622, - "data_source": "Mono3DRefer", - "scene_name": "001238", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[182.34, 314.15], [240.4, 308.03], [257.46, 319.47], [195.09, 326.7], [182.34, 171.23], [240.4, 171.3], [257.46, 171.17], [195.09, 171.08]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[389.6, 289.03], [360.6, 294.28], [324.65, 289.83], [354.14, 284.96], [389.6, 172.83], [360.6, 172.83], [324.65, 172.83], [354.14, 172.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[412.64, 287.12], [375.29, 291.88], [349.63, 285.89], [386.14, 281.59], [412.64, 170.32], [375.29, 170.21], [349.63, 170.35], [386.14, 170.44]]}, {\"category\": \"cyclist\", \"corners_3d\": [[376.56, 243.35], [390.72, 241.67], [432.46, 244.0], [418.87, 245.8], [376.56, 190.17], [390.72, 189.75], [432.46, 190.33], [418.87, 190.77]]}, {\"category\": \"cyclist\", \"corners_3d\": [[385.6, 277.35], [371.18, 281.78], [276.04, 279.33], [293.94, 275.09], [385.6, 183.59], [371.18, 184.05], [276.04, 183.8], [293.94, 183.36]]}]\n```" - } - ] - }, - { - "id": 1623, - "data_source": "Mono3DRefer", - "scene_name": "001238", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-226.86, 495.91], [-599.0, 505.38], [49.98, 320.09], [209.09, 318.2], [-226.86, 235.51], [-599.0, 237.35], [49.98, 201.41], [209.09, 201.04]]}, {\"category\": \"cyclist\", \"corners_3d\": [[376.56, 243.35], [390.72, 241.67], [432.46, 244.0], [418.87, 245.8], [376.56, 190.17], [390.72, 189.75], [432.46, 190.33], [418.87, 190.77]]}, {\"category\": \"cyclist\", \"corners_3d\": [[385.6, 277.35], [371.18, 281.78], [276.04, 279.33], [293.94, 275.09], [385.6, 183.59], [371.18, 184.05], [276.04, 183.8], [293.94, 183.36]]}, {\"category\": \"car\", \"corners_3d\": [[1194.84, 220.84], [1137.67, 217.55], [1237.5, 214.7], [1297.64, 217.58], [1194.84, 165.77], [1137.67, 166.26], [1237.5, 166.68], [1297.64, 166.25]]}]\n```" - } - ] - }, - { - "id": 1624, - "data_source": "Mono3DRefer", - "scene_name": "001238", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-226.86, 495.91], [-599.0, 505.38], [49.98, 320.09], [209.09, 318.2], [-226.86, 235.51], [-599.0, 237.35], [49.98, 201.41], [209.09, 201.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[182.34, 314.15], [240.4, 308.03], [257.46, 319.47], [195.09, 326.7], [182.34, 171.23], [240.4, 171.3], [257.46, 171.17], [195.09, 171.08]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[389.6, 289.03], [360.6, 294.28], [324.65, 289.83], [354.14, 284.96], [389.6, 172.83], [360.6, 172.83], [324.65, 172.83], [354.14, 172.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[412.64, 287.12], [375.29, 291.88], [349.63, 285.89], [386.14, 281.59], [412.64, 170.32], [375.29, 170.21], [349.63, 170.35], [386.14, 170.44]]}, {\"category\": \"car\", \"corners_3d\": [[1194.84, 220.84], [1137.67, 217.55], [1237.5, 214.7], [1297.64, 217.58], [1194.84, 165.77], [1137.67, 166.26], [1237.5, 166.68], [1297.64, 166.25]]}]\n```" - } - ] - }, - { - "id": 1625, - "data_source": "Mono3DRefer", - "scene_name": "001238", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-226.86, 495.91], [-599.0, 505.38], [49.98, 320.09], [209.09, 318.2], [-226.86, 235.51], [-599.0, 237.35], [49.98, 201.41], [209.09, 201.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[182.34, 314.15], [240.4, 308.03], [257.46, 319.47], [195.09, 326.7], [182.34, 171.23], [240.4, 171.3], [257.46, 171.17], [195.09, 171.08]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[389.6, 289.03], [360.6, 294.28], [324.65, 289.83], [354.14, 284.96], [389.6, 172.83], [360.6, 172.83], [324.65, 172.83], [354.14, 172.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[412.64, 287.12], [375.29, 291.88], [349.63, 285.89], [386.14, 281.59], [412.64, 170.32], [375.29, 170.21], [349.63, 170.35], [386.14, 170.44]]}, {\"category\": \"cyclist\", \"corners_3d\": [[376.56, 243.35], [390.72, 241.67], [432.46, 244.0], [418.87, 245.8], [376.56, 190.17], [390.72, 189.75], [432.46, 190.33], [418.87, 190.77]]}, {\"category\": \"cyclist\", \"corners_3d\": [[385.6, 277.35], [371.18, 281.78], [276.04, 279.33], [293.94, 275.09], [385.6, 183.59], [371.18, 184.05], [276.04, 183.8], [293.94, 183.36]]}, {\"category\": \"car\", \"corners_3d\": [[1194.84, 220.84], [1137.67, 217.55], [1237.5, 214.7], [1297.64, 217.58], [1194.84, 165.77], [1137.67, 166.26], [1237.5, 166.68], [1297.64, 166.25]]}]\n```" - } - ] - }, - { - "id": 1626, - "data_source": "Mono3DRefer", - "scene_name": "001241", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[579.23, 211.0], [561.12, 209.55], [633.17, 207.09], [652.68, 208.36], [579.23, 171.84], [561.12, 171.88], [633.17, 171.94], [652.68, 171.91]]}, {\"category\": \"car\", \"corners_3d\": [[780.56, 210.26], [815.96, 212.1], [747.27, 215.73], [712.26, 213.54], [780.56, 170.55], [815.96, 170.44], [747.27, 170.22], [712.26, 170.35]]}, {\"category\": \"car\", \"corners_3d\": [[848.26, 203.78], [880.98, 204.99], [824.63, 207.99], [791.28, 206.55], [848.26, 169.15], [880.98, 169.01], [824.63, 168.65], [791.28, 168.82]]}, {\"category\": \"car\", \"corners_3d\": [[908.18, 199.01], [934.79, 199.82], [896.64, 201.52], [869.63, 200.61], [908.18, 169.42], [934.79, 169.31], [896.64, 169.08], [869.63, 169.21]]}, {\"category\": \"car\", \"corners_3d\": [[956.06, 196.13], [990.49, 196.98], [948.1, 199.05], [912.45, 198.05], [956.06, 163.15], [990.49, 162.79], [948.1, 161.93], [912.45, 162.34]]}, {\"category\": \"car\", \"corners_3d\": [[634.58, 241.85], [699.48, 240.79], [753.08, 257.87], [672.61, 259.55], [634.58, 181.14], [699.48, 181.01], [753.08, 183.06], [672.61, 183.26]]}]\n```" - } - ] - }, - { - "id": 1627, - "data_source": "Mono3DRefer", - "scene_name": "001242", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (motorcyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"motorcyclist\", \"corners_3d\": [[462.65, 335.56], [389.02, 333.52], [447.35, 302.99], [507.44, 304.33], [462.65, 150.82], [389.02, 151.1], [447.35, 155.23], [507.44, 155.05]]}]\n```" - } - ] - }, - { - "id": 1628, - "data_source": "Mono3DRefer", - "scene_name": "001242", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[739.91, 156.48], [713.19, 156.66], [723.09, 158.01], [747.66, 157.86], [739.91, 131.69], [713.19, 132.14], [723.09, 135.52], [747.66, 135.15]]}]\n```" - } - ] - }, - { - "id": 1629, - "data_source": "Mono3DRefer", - "scene_name": "001242", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (motorcyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"motorcyclist\", \"corners_3d\": [[462.65, 335.56], [389.02, 333.52], [447.35, 302.99], [507.44, 304.33], [462.65, 150.82], [389.02, 151.1], [447.35, 155.23], [507.44, 155.05]]}, {\"category\": \"car\", \"corners_3d\": [[739.91, 156.48], [713.19, 156.66], [723.09, 158.01], [747.66, 157.86], [739.91, 131.69], [713.19, 132.14], [723.09, 135.52], [747.66, 135.15]]}]\n```" - } - ] - }, - { - "id": 1630, - "data_source": "Mono3DRefer", - "scene_name": "001245", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[267.42, 224.47], [230.23, 224.39], [270.01, 219.27], [303.56, 219.34], [267.42, 195.25], [230.23, 195.22], [270.01, 193.0], [303.56, 193.03]]}, {\"category\": \"car\", \"corners_3d\": [[393.28, 207.68], [368.32, 207.62], [391.14, 204.75], [414.08, 204.8], [393.28, 184.36], [368.32, 184.34], [391.14, 183.39], [414.08, 183.41]]}, {\"category\": \"car\", \"corners_3d\": [[463.72, 200.47], [443.16, 200.46], [453.32, 198.84], [472.67, 198.85], [463.72, 182.06], [443.16, 182.05], [453.32, 181.51], [472.67, 181.52]]}]\n```" - } - ] - }, - { - "id": 1631, - "data_source": "Mono3DRefer", - "scene_name": "001245", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[705.64, 199.58], [714.21, 199.58], [718.03, 200.56], [709.14, 200.56], [705.64, 170.16], [714.21, 170.16], [718.03, 170.06], [709.14, 170.06]]}]\n```" - } - ] - }, - { - "id": 1632, - "data_source": "Mono3DRefer", - "scene_name": "001245", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[514.76, 195.76], [484.82, 195.76], [513.47, 190.48], [536.5, 190.48], [514.76, 160.13], [484.82, 160.13], [513.47, 163.07], [536.5, 163.07]]}]\n```" - } - ] - }, - { - "id": 1633, - "data_source": "Mono3DRefer", - "scene_name": "001245", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[267.42, 224.47], [230.23, 224.39], [270.01, 219.27], [303.56, 219.34], [267.42, 195.25], [230.23, 195.22], [270.01, 193.0], [303.56, 193.03]]}, {\"category\": \"car\", \"corners_3d\": [[393.28, 207.68], [368.32, 207.62], [391.14, 204.75], [414.08, 204.8], [393.28, 184.36], [368.32, 184.34], [391.14, 183.39], [414.08, 183.41]]}, {\"category\": \"car\", \"corners_3d\": [[463.72, 200.47], [443.16, 200.46], [453.32, 198.84], [472.67, 198.85], [463.72, 182.06], [443.16, 182.05], [453.32, 181.51], [472.67, 181.52]]}, {\"category\": \"cyclist\", \"corners_3d\": [[705.64, 199.58], [714.21, 199.58], [718.03, 200.56], [709.14, 200.56], [705.64, 170.16], [714.21, 170.16], [718.03, 170.06], [709.14, 170.06]]}]\n```" - } - ] - }, - { - "id": 1634, - "data_source": "Mono3DRefer", - "scene_name": "001245", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[267.42, 224.47], [230.23, 224.39], [270.01, 219.27], [303.56, 219.34], [267.42, 195.25], [230.23, 195.22], [270.01, 193.0], [303.56, 193.03]]}, {\"category\": \"car\", \"corners_3d\": [[393.28, 207.68], [368.32, 207.62], [391.14, 204.75], [414.08, 204.8], [393.28, 184.36], [368.32, 184.34], [391.14, 183.39], [414.08, 183.41]]}, {\"category\": \"car\", \"corners_3d\": [[463.72, 200.47], [443.16, 200.46], [453.32, 198.84], [472.67, 198.85], [463.72, 182.06], [443.16, 182.05], [453.32, 181.51], [472.67, 181.52]]}, {\"category\": \"tram\", \"corners_3d\": [[514.76, 195.76], [484.82, 195.76], [513.47, 190.48], [536.5, 190.48], [514.76, 160.13], [484.82, 160.13], [513.47, 163.07], [536.5, 163.07]]}]\n```" - } - ] - }, - { - "id": 1635, - "data_source": "Mono3DRefer", - "scene_name": "001245", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[514.76, 195.76], [484.82, 195.76], [513.47, 190.48], [536.5, 190.48], [514.76, 160.13], [484.82, 160.13], [513.47, 163.07], [536.5, 163.07]]}, {\"category\": \"cyclist\", \"corners_3d\": [[705.64, 199.58], [714.21, 199.58], [718.03, 200.56], [709.14, 200.56], [705.64, 170.16], [714.21, 170.16], [718.03, 170.06], [709.14, 170.06]]}]\n```" - } - ] - }, - { - "id": 1636, - "data_source": "Mono3DRefer", - "scene_name": "001245", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[267.42, 224.47], [230.23, 224.39], [270.01, 219.27], [303.56, 219.34], [267.42, 195.25], [230.23, 195.22], [270.01, 193.0], [303.56, 193.03]]}, {\"category\": \"car\", \"corners_3d\": [[393.28, 207.68], [368.32, 207.62], [391.14, 204.75], [414.08, 204.8], [393.28, 184.36], [368.32, 184.34], [391.14, 183.39], [414.08, 183.41]]}, {\"category\": \"car\", \"corners_3d\": [[463.72, 200.47], [443.16, 200.46], [453.32, 198.84], [472.67, 198.85], [463.72, 182.06], [443.16, 182.05], [453.32, 181.51], [472.67, 181.52]]}, {\"category\": \"tram\", \"corners_3d\": [[514.76, 195.76], [484.82, 195.76], [513.47, 190.48], [536.5, 190.48], [514.76, 160.13], [484.82, 160.13], [513.47, 163.07], [536.5, 163.07]]}, {\"category\": \"cyclist\", \"corners_3d\": [[705.64, 199.58], [714.21, 199.58], [718.03, 200.56], [709.14, 200.56], [705.64, 170.16], [714.21, 170.16], [718.03, 170.06], [709.14, 170.06]]}]\n```" - } - ] - }, - { - "id": 1637, - "data_source": "Mono3DRefer", - "scene_name": "001257", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[584.17, 192.55], [599.55, 192.54], [600.1, 193.52], [583.95, 193.53], [584.17, 180.44], [599.55, 180.44], [600.1, 180.82], [583.95, 180.82]]}, {\"category\": \"car\", \"corners_3d\": [[474.22, 212.27], [445.87, 212.24], [463.23, 208.39], [488.82, 208.41], [474.22, 183.79], [445.87, 183.78], [463.23, 182.71], [488.82, 182.72]]}, {\"category\": \"car\", \"corners_3d\": [[520.36, 196.34], [500.31, 196.34], [507.29, 194.92], [526.14, 194.93], [520.36, 178.84], [500.31, 178.84], [507.29, 178.48], [526.14, 178.48]]}]\n```" - } - ] - }, - { - "id": 1638, - "data_source": "Mono3DRefer", - "scene_name": "001258", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[568.6, 202.34], [544.39, 202.37], [547.57, 200.18], [569.98, 200.15], [568.6, 178.47], [544.39, 178.47], [547.57, 178.06], [569.98, 178.05]]}]\n```" - } - ] - }, - { - "id": 1639, - "data_source": "Mono3DRefer", - "scene_name": "001269", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[582.77, 189.21], [600.11, 189.2], [600.77, 190.13], [582.45, 190.14], [582.77, 175.56], [600.11, 175.56], [600.77, 175.71], [582.45, 175.71]]}, {\"category\": \"car\", \"corners_3d\": [[216.76, 283.47], [113.61, 283.48], [231.64, 257.12], [310.21, 257.11], [216.76, 190.43], [113.61, 190.43], [231.64, 186.25], [310.21, 186.24]]}]\n```" - } - ] - }, - { - "id": 1640, - "data_source": "Mono3DRefer", - "scene_name": "001273", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[530.93, 212.26], [491.98, 212.19], [507.59, 207.77], [542.17, 207.81], [530.93, 176.45], [491.98, 176.45], [507.59, 176.04], [542.17, 176.05]]}, {\"category\": \"car\", \"corners_3d\": [[378.78, 312.45], [243.5, 312.73], [344.7, 273.22], [441.68, 273.08], [378.78, 188.9], [243.5, 188.93], [344.7, 184.39], [441.68, 184.38]]}, {\"category\": \"car\", \"corners_3d\": [[-568.05, 869.27], [-1358.96, 929.78], [-42.17, 405.97], [181.36, 399.9], [-568.05, 289.47], [-1358.96, 299.61], [-42.17, 211.89], [181.36, 210.87]]}]\n```" - } - ] - }, - { - "id": 1641, - "data_source": "Mono3DRefer", - "scene_name": "001279", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[868.31, 289.94], [853.22, 284.31], [944.08, 283.75], [963.99, 289.31], [868.31, 133.41], [853.22, 135.83], [944.08, 136.07], [963.99, 133.68]]}]\n```" - } - ] - }, - { - "id": 1642, - "data_source": "Mono3DRefer", - "scene_name": "001282", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[529.54, 235.4], [476.7, 235.31], [497.29, 226.55], [542.73, 226.61], [529.54, 186.1], [476.7, 186.08], [497.29, 184.23], [542.73, 184.24]]}]\n```" - } - ] - }, - { - "id": 1643, - "data_source": "Mono3DRefer", - "scene_name": "001283", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[532.11, 229.05], [487.85, 229.01], [505.53, 221.28], [543.69, 221.3], [532.11, 187.04], [487.85, 187.03], [505.53, 185.08], [543.69, 185.08]]}, {\"category\": \"car\", \"corners_3d\": [[-315.32, 744.66], [-807.47, 741.03], [100.56, 378.64], [280.18, 379.12], [-315.32, 214.51], [-807.47, 214.24], [100.56, 187.85], [280.18, 187.88]]}]\n```" - } - ] - }, - { - "id": 1644, - "data_source": "Mono3DRefer", - "scene_name": "001283", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[357.31, 226.27], [344.14, 226.19], [360.13, 223.54], [372.67, 223.62], [357.31, 185.95], [344.14, 185.93], [360.13, 185.28], [372.67, 185.3]]}]\n```" - } - ] - }, - { - "id": 1645, - "data_source": "Mono3DRefer", - "scene_name": "001283", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[532.11, 229.05], [487.85, 229.01], [505.53, 221.28], [543.69, 221.3], [532.11, 187.04], [487.85, 187.03], [505.53, 185.08], [543.69, 185.08]]}, {\"category\": \"car\", \"corners_3d\": [[-315.32, 744.66], [-807.47, 741.03], [100.56, 378.64], [280.18, 379.12], [-315.32, 214.51], [-807.47, 214.24], [100.56, 187.85], [280.18, 187.88]]}, {\"category\": \"cyclist\", \"corners_3d\": [[357.31, 226.27], [344.14, 226.19], [360.13, 223.54], [372.67, 223.62], [357.31, 185.95], [344.14, 185.93], [360.13, 185.28], [372.67, 185.3]]}]\n```" - } - ] - }, - { - "id": 1646, - "data_source": "Mono3DRefer", - "scene_name": "001284", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[499.07, 244.52], [438.83, 244.41], [469.09, 232.67], [519.49, 232.75], [499.07, 189.17], [438.83, 189.14], [469.09, 186.47], [519.49, 186.49]]}, {\"category\": \"car\", \"corners_3d\": [[555.92, 202.51], [531.62, 202.56], [535.15, 200.02], [557.37, 199.97], [555.92, 182.73], [531.62, 182.75], [535.15, 181.91], [557.37, 181.89]]}, {\"category\": \"car\", \"corners_3d\": [[-82.61, 290.76], [-47.56, 278.02], [231.62, 263.85], [230.59, 273.24], [-82.61, 193.23], [-47.56, 191.03], [231.62, 188.58], [230.59, 190.2]]}, {\"category\": \"car\", \"corners_3d\": [[440.35, 209.16], [435.18, 208.07], [504.56, 206.63], [511.57, 207.63], [440.35, 182.5], [435.18, 182.21], [504.56, 181.82], [511.57, 182.09]]}]\n```" - } - ] - }, - { - "id": 1647, - "data_source": "Mono3DRefer", - "scene_name": "001284", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[570.62, 187.7], [586.83, 187.52], [609.69, 188.43], [592.75, 188.63], [570.62, 163.37], [586.83, 163.48], [609.69, 162.9], [592.75, 162.78]]}]\n```" - } - ] - }, - { - "id": 1648, - "data_source": "Mono3DRefer", - "scene_name": "001284", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[499.07, 244.52], [438.83, 244.41], [469.09, 232.67], [519.49, 232.75], [499.07, 189.17], [438.83, 189.14], [469.09, 186.47], [519.49, 186.49]]}, {\"category\": \"car\", \"corners_3d\": [[555.92, 202.51], [531.62, 202.56], [535.15, 200.02], [557.37, 199.97], [555.92, 182.73], [531.62, 182.75], [535.15, 181.91], [557.37, 181.89]]}, {\"category\": \"van\", \"corners_3d\": [[570.62, 187.7], [586.83, 187.52], [609.69, 188.43], [592.75, 188.63], [570.62, 163.37], [586.83, 163.48], [609.69, 162.9], [592.75, 162.78]]}, {\"category\": \"car\", \"corners_3d\": [[-82.61, 290.76], [-47.56, 278.02], [231.62, 263.85], [230.59, 273.24], [-82.61, 193.23], [-47.56, 191.03], [231.62, 188.58], [230.59, 190.2]]}, {\"category\": \"car\", \"corners_3d\": [[440.35, 209.16], [435.18, 208.07], [504.56, 206.63], [511.57, 207.63], [440.35, 182.5], [435.18, 182.21], [504.56, 181.82], [511.57, 182.09]]}]\n```" - } - ] - }, - { - "id": 1649, - "data_source": "Mono3DRefer", - "scene_name": "001285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[274.01, 228.03], [236.86, 228.04], [276.53, 222.15], [309.72, 222.15], [274.01, 197.33], [236.86, 197.33], [276.53, 194.72], [309.72, 194.72]]}]\n```" - } - ] - }, - { - "id": 1650, - "data_source": "Mono3DRefer", - "scene_name": "001285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[591.39, 199.55], [629.98, 199.55], [634.94, 206.23], [586.7, 206.23], [591.39, 152.55], [629.98, 152.55], [634.94, 147.48], [586.7, 147.48]]}]\n```" - } - ] - }, - { - "id": 1651, - "data_source": "Mono3DRefer", - "scene_name": "001285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[591.39, 199.55], [629.98, 199.55], [634.94, 206.23], [586.7, 206.23], [591.39, 152.55], [629.98, 152.55], [634.94, 147.48], [586.7, 147.48]]}, {\"category\": \"car\", \"corners_3d\": [[274.01, 228.03], [236.86, 228.04], [276.53, 222.15], [309.72, 222.15], [274.01, 197.33], [236.86, 197.33], [276.53, 194.72], [309.72, 194.72]]}]\n```" - } - ] - }, - { - "id": 1652, - "data_source": "Mono3DRefer", - "scene_name": "001298", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[658.55, 192.77], [640.24, 192.64], [649.6, 191.43], [666.84, 191.54], [658.55, 176.74], [640.24, 176.72], [649.6, 176.48], [666.84, 176.51]]}]\n```" - } - ] - }, - { - "id": 1653, - "data_source": "Mono3DRefer", - "scene_name": "001299", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[572.02, 209.87], [605.37, 209.9], [602.83, 214.8], [565.08, 214.75], [572.02, 179.09], [605.37, 179.09], [602.83, 179.92], [565.08, 179.91]]}, {\"category\": \"car\", \"corners_3d\": [[1534.98, 357.15], [1848.08, 427.83], [1056.56, 394.26], [972.13, 338.95], [1534.98, 176.4], [1848.08, 177.76], [1056.56, 177.12], [972.13, 176.05]]}, {\"category\": \"car\", \"corners_3d\": [[520.78, 203.36], [489.83, 203.28], [503.47, 200.73], [531.84, 200.8], [520.78, 174.83], [489.83, 174.82], [503.47, 174.66], [531.84, 174.66]]}]\n```" - } - ] - }, - { - "id": 1654, - "data_source": "Mono3DRefer", - "scene_name": "001300", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[431.65, 322.95], [390.83, 331.67], [323.46, 318.91], [364.59, 311.57], [431.65, 140.14], [390.83, 137.67], [323.46, 141.28], [364.59, 143.37]]}]\n```" - } - ] - }, - { - "id": 1655, - "data_source": "Mono3DRefer", - "scene_name": "001302", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[590.44, 200.09], [629.61, 200.09], [634.57, 207.03], [585.42, 207.03], [590.44, 152.38], [629.61, 152.38], [634.57, 147.17], [585.42, 147.17]]}]\n```" - } - ] - }, - { - "id": 1656, - "data_source": "Mono3DRefer", - "scene_name": "001302", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[370.75, 214.48], [344.41, 214.48], [365.05, 211.23], [389.34, 211.23], [370.75, 192.71], [344.41, 192.71], [365.05, 191.16], [389.34, 191.16]]}]\n```" - } - ] - }, - { - "id": 1657, - "data_source": "Mono3DRefer", - "scene_name": "001302", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[590.44, 200.09], [629.61, 200.09], [634.57, 207.03], [585.42, 207.03], [590.44, 152.38], [629.61, 152.38], [634.57, 147.17], [585.42, 147.17]]}, {\"category\": \"car\", \"corners_3d\": [[370.75, 214.48], [344.41, 214.48], [365.05, 211.23], [389.34, 211.23], [370.75, 192.71], [344.41, 192.71], [365.05, 191.16], [389.34, 191.16]]}]\n```" - } - ] - }, - { - "id": 1658, - "data_source": "Mono3DRefer", - "scene_name": "001311", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[872.54, 280.68], [920.88, 280.68], [942.83, 288.3], [891.08, 288.3], [872.54, 153.22], [920.88, 153.22], [942.83, 151.83], [891.08, 151.83]]}]\n```" - } - ] - }, - { - "id": 1659, - "data_source": "Mono3DRefer", - "scene_name": "001311", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[663.66, 223.77], [710.4, 223.74], [729.8, 232.93], [674.63, 232.97], [663.66, 177.55], [710.4, 177.55], [729.8, 178.4], [674.63, 178.4]]}, {\"category\": \"car\", \"corners_3d\": [[415.26, 217.99], [407.86, 219.85], [329.79, 219.95], [340.3, 218.09], [415.26, 189.28], [407.86, 189.96], [329.79, 189.99], [340.3, 189.32]]}, {\"category\": \"car\", \"corners_3d\": [[311.45, 220.51], [298.44, 222.54], [215.29, 222.48], [231.69, 220.45], [311.45, 190.62], [298.44, 191.38], [215.29, 191.36], [231.69, 190.61]]}, {\"category\": \"car\", \"corners_3d\": [[633.25, 198.74], [654.22, 198.72], [659.45, 200.78], [636.82, 200.81], [633.25, 178.54], [654.22, 178.54], [659.45, 178.99], [636.82, 179.0]]}, {\"category\": \"car\", \"corners_3d\": [[593.93, 199.97], [616.83, 199.96], [617.91, 202.05], [593.24, 202.05], [593.93, 178.33], [616.83, 178.33], [617.91, 178.75], [593.24, 178.75]]}, {\"category\": \"car\", \"corners_3d\": [[411.52, 338.35], [261.07, 338.37], [372.42, 285.4], [474.71, 285.39], [411.52, 201.2], [261.07, 201.2], [372.42, 192.13], [474.71, 192.13]]}, {\"category\": \"car\", \"corners_3d\": [[852.4, 381.64], [1056.3, 381.09], [1986.76, 808.46], [1368.51, 813.56], [852.4, 176.56], [1056.3, 176.55], [1986.76, 184.14], [1368.51, 184.23]]}, {\"category\": \"car\", \"corners_3d\": [[764.52, 307.89], [872.84, 306.88], [994.45, 361.42], [842.87, 363.42], [764.52, 173.62], [872.84, 173.61], [994.45, 173.92], [842.87, 173.94]]}, {\"category\": \"car\", \"corners_3d\": [[676.26, 236.15], [735.03, 235.95], [772.24, 251.21], [699.35, 251.53], [676.26, 180.29], [735.03, 180.26], [772.24, 182.06], [699.35, 182.1]]}, {\"category\": \"car\", \"corners_3d\": [[644.67, 209.47], [677.18, 209.48], [685.09, 214.32], [648.27, 214.3], [644.67, 178.64], [677.18, 178.64], [685.09, 179.41], [648.27, 179.4]]}]\n```" - } - ] - }, - { - "id": 1660, - "data_source": "Mono3DRefer", - "scene_name": "001311", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[663.66, 223.77], [710.4, 223.74], [729.8, 232.93], [674.63, 232.97], [663.66, 177.55], [710.4, 177.55], [729.8, 178.4], [674.63, 178.4]]}, {\"category\": \"car\", \"corners_3d\": [[415.26, 217.99], [407.86, 219.85], [329.79, 219.95], [340.3, 218.09], [415.26, 189.28], [407.86, 189.96], [329.79, 189.99], [340.3, 189.32]]}, {\"category\": \"car\", \"corners_3d\": [[311.45, 220.51], [298.44, 222.54], [215.29, 222.48], [231.69, 220.45], [311.45, 190.62], [298.44, 191.38], [215.29, 191.36], [231.69, 190.61]]}, {\"category\": \"car\", \"corners_3d\": [[633.25, 198.74], [654.22, 198.72], [659.45, 200.78], [636.82, 200.81], [633.25, 178.54], [654.22, 178.54], [659.45, 178.99], [636.82, 179.0]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[872.54, 280.68], [920.88, 280.68], [942.83, 288.3], [891.08, 288.3], [872.54, 153.22], [920.88, 153.22], [942.83, 151.83], [891.08, 151.83]]}, {\"category\": \"car\", \"corners_3d\": [[593.93, 199.97], [616.83, 199.96], [617.91, 202.05], [593.24, 202.05], [593.93, 178.33], [616.83, 178.33], [617.91, 178.75], [593.24, 178.75]]}, {\"category\": \"car\", \"corners_3d\": [[411.52, 338.35], [261.07, 338.37], [372.42, 285.4], [474.71, 285.39], [411.52, 201.2], [261.07, 201.2], [372.42, 192.13], [474.71, 192.13]]}, {\"category\": \"car\", \"corners_3d\": [[852.4, 381.64], [1056.3, 381.09], [1986.76, 808.46], [1368.51, 813.56], [852.4, 176.56], [1056.3, 176.55], [1986.76, 184.14], [1368.51, 184.23]]}, {\"category\": \"car\", \"corners_3d\": [[764.52, 307.89], [872.84, 306.88], [994.45, 361.42], [842.87, 363.42], [764.52, 173.62], [872.84, 173.61], [994.45, 173.92], [842.87, 173.94]]}, {\"category\": \"car\", \"corners_3d\": [[676.26, 236.15], [735.03, 235.95], [772.24, 251.21], [699.35, 251.53], [676.26, 180.29], [735.03, 180.26], [772.24, 182.06], [699.35, 182.1]]}, {\"category\": \"car\", \"corners_3d\": [[644.67, 209.47], [677.18, 209.48], [685.09, 214.32], [648.27, 214.3], [644.67, 178.64], [677.18, 178.64], [685.09, 179.41], [648.27, 179.4]]}]\n```" - } - ] - }, - { - "id": 1661, - "data_source": "Mono3DRefer", - "scene_name": "001313", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[628.95, 229.43], [680.08, 229.07], [712.35, 242.51], [649.16, 243.06], [628.95, 182.1], [680.08, 182.04], [712.35, 184.24], [649.16, 184.33]]}, {\"category\": \"car\", \"corners_3d\": [[750.17, 292.47], [874.62, 290.38], [1077.75, 361.24], [881.96, 366.68], [750.17, 167.35], [874.62, 167.45], [1077.75, 164.19], [881.96, 163.94]]}]\n```" - } - ] - }, - { - "id": 1662, - "data_source": "Mono3DRefer", - "scene_name": "001321", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[387.25, 344.39], [205.9, 344.43], [356.1, 280.5], [469.86, 280.48], [387.25, 200.93], [205.9, 200.94], [356.1, 190.47], [469.86, 190.47]]}]\n```" - } - ] - }, - { - "id": 1663, - "data_source": "Mono3DRefer", - "scene_name": "001325", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[436.92, 208.19], [428.11, 208.16], [435.57, 206.99], [444.09, 207.01], [436.92, 181.64], [428.11, 181.64], [435.57, 181.34], [444.09, 181.35]]}]\n```" - } - ] - }, - { - "id": 1664, - "data_source": "Mono3DRefer", - "scene_name": "001325", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[557.48, 203.64], [533.05, 203.65], [538.61, 201.16], [561.07, 201.15], [557.48, 180.51], [533.05, 180.51], [538.61, 179.89], [561.07, 179.89]]}, {\"category\": \"car\", \"corners_3d\": [[504.7, 234.99], [452.68, 234.95], [478.11, 225.29], [522.05, 225.32], [504.7, 179.86], [452.68, 179.85], [478.11, 178.77], [522.05, 178.77]]}]\n```" - } - ] - }, - { - "id": 1665, - "data_source": "Mono3DRefer", - "scene_name": "001325", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[436.92, 208.19], [428.11, 208.16], [435.57, 206.99], [444.09, 207.01], [436.92, 181.64], [428.11, 181.64], [435.57, 181.34], [444.09, 181.35]]}, {\"category\": \"car\", \"corners_3d\": [[557.48, 203.64], [533.05, 203.65], [538.61, 201.16], [561.07, 201.15], [557.48, 180.51], [533.05, 180.51], [538.61, 179.89], [561.07, 179.89]]}, {\"category\": \"car\", \"corners_3d\": [[504.7, 234.99], [452.68, 234.95], [478.11, 225.29], [522.05, 225.32], [504.7, 179.86], [452.68, 179.85], [478.11, 178.77], [522.05, 178.77]]}]\n```" - } - ] - }, - { - "id": 1666, - "data_source": "Mono3DRefer", - "scene_name": "001327", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[581.22, 193.66], [557.15, 193.69], [559.4, 192.13], [581.67, 192.11], [581.22, 168.35], [557.15, 168.35], [559.4, 168.68], [581.67, 168.69]]}, {\"category\": \"car\", \"corners_3d\": [[468.43, 296.71], [356.49, 297.11], [422.36, 262.72], [503.25, 262.52], [468.43, 201.62], [356.49, 201.72], [422.36, 193.73], [503.25, 193.68]]}]\n```" - } - ] - }, - { - "id": 1667, - "data_source": "Mono3DRefer", - "scene_name": "001328", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[619.86, 182.98], [639.04, 182.97], [654.37, 186.55], [628.41, 186.56], [619.86, 156.0], [639.04, 156.01], [654.37, 150.05], [628.41, 150.03]]}]\n```" - } - ] - }, - { - "id": 1668, - "data_source": "Mono3DRefer", - "scene_name": "001328", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1193.27, 230.98], [1217.43, 230.98], [1307.57, 239.61], [1279.83, 239.61], [1193.27, 133.34], [1217.43, 133.34], [1307.57, 127.47], [1279.83, 127.47]]}, {\"category\": \"cyclist\", \"corners_3d\": [[880.17, 195.46], [892.46, 195.46], [915.0, 197.26], [901.73, 197.26], [880.17, 149.74], [892.46, 149.74], [915.0, 147.9], [901.73, 147.9]]}]\n```" - } - ] - }, - { - "id": 1669, - "data_source": "Mono3DRefer", - "scene_name": "001328", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[209.59, 219.74], [189.15, 221.8], [116.88, 221.33], [140.15, 219.31], [209.59, 189.86], [189.15, 190.61], [116.88, 190.44], [140.15, 189.71]]}, {\"category\": \"car\", \"corners_3d\": [[233.21, 215.99], [215.77, 217.59], [159.92, 217.1], [179.15, 215.54], [233.21, 191.34], [215.77, 192.02], [159.92, 191.81], [179.15, 191.14]]}]\n```" - } - ] - }, - { - "id": 1670, - "data_source": "Mono3DRefer", - "scene_name": "001328", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[619.86, 182.98], [639.04, 182.97], [654.37, 186.55], [628.41, 186.56], [619.86, 156.0], [639.04, 156.01], [654.37, 150.05], [628.41, 150.03]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1193.27, 230.98], [1217.43, 230.98], [1307.57, 239.61], [1279.83, 239.61], [1193.27, 133.34], [1217.43, 133.34], [1307.57, 127.47], [1279.83, 127.47]]}, {\"category\": \"cyclist\", \"corners_3d\": [[880.17, 195.46], [892.46, 195.46], [915.0, 197.26], [901.73, 197.26], [880.17, 149.74], [892.46, 149.74], [915.0, 147.9], [901.73, 147.9]]}]\n```" - } - ] - }, - { - "id": 1671, - "data_source": "Mono3DRefer", - "scene_name": "001328", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[619.86, 182.98], [639.04, 182.97], [654.37, 186.55], [628.41, 186.56], [619.86, 156.0], [639.04, 156.01], [654.37, 150.05], [628.41, 150.03]]}, {\"category\": \"car\", \"corners_3d\": [[209.59, 219.74], [189.15, 221.8], [116.88, 221.33], [140.15, 219.31], [209.59, 189.86], [189.15, 190.61], [116.88, 190.44], [140.15, 189.71]]}, {\"category\": \"car\", \"corners_3d\": [[233.21, 215.99], [215.77, 217.59], [159.92, 217.1], [179.15, 215.54], [233.21, 191.34], [215.77, 192.02], [159.92, 191.81], [179.15, 191.14]]}]\n```" - } - ] - }, - { - "id": 1672, - "data_source": "Mono3DRefer", - "scene_name": "001328", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1193.27, 230.98], [1217.43, 230.98], [1307.57, 239.61], [1279.83, 239.61], [1193.27, 133.34], [1217.43, 133.34], [1307.57, 127.47], [1279.83, 127.47]]}, {\"category\": \"cyclist\", \"corners_3d\": [[880.17, 195.46], [892.46, 195.46], [915.0, 197.26], [901.73, 197.26], [880.17, 149.74], [892.46, 149.74], [915.0, 147.9], [901.73, 147.9]]}, {\"category\": \"car\", \"corners_3d\": [[209.59, 219.74], [189.15, 221.8], [116.88, 221.33], [140.15, 219.31], [209.59, 189.86], [189.15, 190.61], [116.88, 190.44], [140.15, 189.71]]}, {\"category\": \"car\", \"corners_3d\": [[233.21, 215.99], [215.77, 217.59], [159.92, 217.1], [179.15, 215.54], [233.21, 191.34], [215.77, 192.02], [159.92, 191.81], [179.15, 191.14]]}]\n```" - } - ] - }, - { - "id": 1673, - "data_source": "Mono3DRefer", - "scene_name": "001328", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[619.86, 182.98], [639.04, 182.97], [654.37, 186.55], [628.41, 186.56], [619.86, 156.0], [639.04, 156.01], [654.37, 150.05], [628.41, 150.03]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1193.27, 230.98], [1217.43, 230.98], [1307.57, 239.61], [1279.83, 239.61], [1193.27, 133.34], [1217.43, 133.34], [1307.57, 127.47], [1279.83, 127.47]]}, {\"category\": \"cyclist\", \"corners_3d\": [[880.17, 195.46], [892.46, 195.46], [915.0, 197.26], [901.73, 197.26], [880.17, 149.74], [892.46, 149.74], [915.0, 147.9], [901.73, 147.9]]}, {\"category\": \"car\", \"corners_3d\": [[209.59, 219.74], [189.15, 221.8], [116.88, 221.33], [140.15, 219.31], [209.59, 189.86], [189.15, 190.61], [116.88, 190.44], [140.15, 189.71]]}, {\"category\": \"car\", \"corners_3d\": [[233.21, 215.99], [215.77, 217.59], [159.92, 217.1], [179.15, 215.54], [233.21, 191.34], [215.77, 192.02], [159.92, 191.81], [179.15, 191.14]]}]\n```" - } - ] - }, - { - "id": 1674, - "data_source": "Mono3DRefer", - "scene_name": "001333", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[893.05, 279.45], [916.5, 279.42], [956.92, 293.15], [930.46, 293.19], [893.05, 185.26], [916.5, 185.25], [956.92, 186.85], [930.46, 186.86]]}]\n```" - } - ] - }, - { - "id": 1675, - "data_source": "Mono3DRefer", - "scene_name": "001333", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[413.96, 353.88], [240.76, 354.78], [358.18, 294.39], [473.7, 293.99], [413.96, 197.46], [240.76, 197.58], [358.18, 189.37], [473.7, 189.32]]}, {\"category\": \"car\", \"corners_3d\": [[735.6, 290.93], [854.03, 290.75], [984.73, 352.12], [804.82, 352.54], [735.6, 176.32], [854.03, 176.32], [984.73, 178.12], [804.82, 178.13]]}, {\"category\": \"car\", \"corners_3d\": [[698.8, 250.47], [771.91, 250.4], [821.75, 273.27], [727.11, 273.39], [698.8, 184.0], [771.91, 183.99], [821.75, 187.27], [727.11, 187.29]]}, {\"category\": \"car\", \"corners_3d\": [[564.23, 210.54], [524.57, 210.58], [533.4, 205.82], [568.05, 205.78], [564.23, 174.02], [524.57, 174.02], [533.4, 173.88], [568.05, 173.88]]}]\n```" - } - ] - }, - { - "id": 1676, - "data_source": "Mono3DRefer", - "scene_name": "001333", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[671.8, 231.26], [729.91, 231.12], [766.54, 246.22], [693.44, 246.44], [671.8, 170.03], [729.91, 170.03], [766.54, 169.3], [693.44, 169.29]]}, {\"category\": \"van\", \"corners_3d\": [[652.23, 206.03], [688.16, 205.99], [703.13, 210.96], [661.83, 211.03], [652.23, 165.06], [688.16, 165.07], [703.13, 163.9], [661.83, 163.89]]}]\n```" - } - ] - }, - { - "id": 1677, - "data_source": "Mono3DRefer", - "scene_name": "001333", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[893.05, 279.45], [916.5, 279.42], [956.92, 293.15], [930.46, 293.19], [893.05, 185.26], [916.5, 185.25], [956.92, 186.85], [930.46, 186.86]]}, {\"category\": \"car\", \"corners_3d\": [[413.96, 353.88], [240.76, 354.78], [358.18, 294.39], [473.7, 293.99], [413.96, 197.46], [240.76, 197.58], [358.18, 189.37], [473.7, 189.32]]}, {\"category\": \"car\", \"corners_3d\": [[735.6, 290.93], [854.03, 290.75], [984.73, 352.12], [804.82, 352.54], [735.6, 176.32], [854.03, 176.32], [984.73, 178.12], [804.82, 178.13]]}, {\"category\": \"car\", \"corners_3d\": [[698.8, 250.47], [771.91, 250.4], [821.75, 273.27], [727.11, 273.39], [698.8, 184.0], [771.91, 183.99], [821.75, 187.27], [727.11, 187.29]]}, {\"category\": \"car\", \"corners_3d\": [[564.23, 210.54], [524.57, 210.58], [533.4, 205.82], [568.05, 205.78], [564.23, 174.02], [524.57, 174.02], [533.4, 173.88], [568.05, 173.88]]}]\n```" - } - ] - }, - { - "id": 1678, - "data_source": "Mono3DRefer", - "scene_name": "001333", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[893.05, 279.45], [916.5, 279.42], [956.92, 293.15], [930.46, 293.19], [893.05, 185.26], [916.5, 185.25], [956.92, 186.85], [930.46, 186.86]]}, {\"category\": \"van\", \"corners_3d\": [[671.8, 231.26], [729.91, 231.12], [766.54, 246.22], [693.44, 246.44], [671.8, 170.03], [729.91, 170.03], [766.54, 169.3], [693.44, 169.29]]}, {\"category\": \"van\", \"corners_3d\": [[652.23, 206.03], [688.16, 205.99], [703.13, 210.96], [661.83, 211.03], [652.23, 165.06], [688.16, 165.07], [703.13, 163.9], [661.83, 163.89]]}]\n```" - } - ] - }, - { - "id": 1679, - "data_source": "Mono3DRefer", - "scene_name": "001333", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[413.96, 353.88], [240.76, 354.78], [358.18, 294.39], [473.7, 293.99], [413.96, 197.46], [240.76, 197.58], [358.18, 189.37], [473.7, 189.32]]}, {\"category\": \"car\", \"corners_3d\": [[735.6, 290.93], [854.03, 290.75], [984.73, 352.12], [804.82, 352.54], [735.6, 176.32], [854.03, 176.32], [984.73, 178.12], [804.82, 178.13]]}, {\"category\": \"car\", \"corners_3d\": [[698.8, 250.47], [771.91, 250.4], [821.75, 273.27], [727.11, 273.39], [698.8, 184.0], [771.91, 183.99], [821.75, 187.27], [727.11, 187.29]]}, {\"category\": \"van\", \"corners_3d\": [[671.8, 231.26], [729.91, 231.12], [766.54, 246.22], [693.44, 246.44], [671.8, 170.03], [729.91, 170.03], [766.54, 169.3], [693.44, 169.29]]}, {\"category\": \"car\", \"corners_3d\": [[564.23, 210.54], [524.57, 210.58], [533.4, 205.82], [568.05, 205.78], [564.23, 174.02], [524.57, 174.02], [533.4, 173.88], [568.05, 173.88]]}, {\"category\": \"van\", \"corners_3d\": [[652.23, 206.03], [688.16, 205.99], [703.13, 210.96], [661.83, 211.03], [652.23, 165.06], [688.16, 165.07], [703.13, 163.9], [661.83, 163.89]]}]\n```" - } - ] - }, - { - "id": 1680, - "data_source": "Mono3DRefer", - "scene_name": "001333", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[893.05, 279.45], [916.5, 279.42], [956.92, 293.15], [930.46, 293.19], [893.05, 185.26], [916.5, 185.25], [956.92, 186.85], [930.46, 186.86]]}, {\"category\": \"car\", \"corners_3d\": [[413.96, 353.88], [240.76, 354.78], [358.18, 294.39], [473.7, 293.99], [413.96, 197.46], [240.76, 197.58], [358.18, 189.37], [473.7, 189.32]]}, {\"category\": \"car\", \"corners_3d\": [[735.6, 290.93], [854.03, 290.75], [984.73, 352.12], [804.82, 352.54], [735.6, 176.32], [854.03, 176.32], [984.73, 178.12], [804.82, 178.13]]}, {\"category\": \"car\", \"corners_3d\": [[698.8, 250.47], [771.91, 250.4], [821.75, 273.27], [727.11, 273.39], [698.8, 184.0], [771.91, 183.99], [821.75, 187.27], [727.11, 187.29]]}, {\"category\": \"van\", \"corners_3d\": [[671.8, 231.26], [729.91, 231.12], [766.54, 246.22], [693.44, 246.44], [671.8, 170.03], [729.91, 170.03], [766.54, 169.3], [693.44, 169.29]]}, {\"category\": \"car\", \"corners_3d\": [[564.23, 210.54], [524.57, 210.58], [533.4, 205.82], [568.05, 205.78], [564.23, 174.02], [524.57, 174.02], [533.4, 173.88], [568.05, 173.88]]}, {\"category\": \"van\", \"corners_3d\": [[652.23, 206.03], [688.16, 205.99], [703.13, 210.96], [661.83, 211.03], [652.23, 165.06], [688.16, 165.07], [703.13, 163.9], [661.83, 163.89]]}]\n```" - } - ] - }, - { - "id": 1681, - "data_source": "Mono3DRefer", - "scene_name": "001334", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1004.79, 247.2], [1035.06, 246.32], [1065.2, 250.0], [1033.77, 250.97], [1004.79, 157.54], [1035.06, 157.72], [1065.2, 156.96], [1033.77, 156.76]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1049.35, 245.04], [1074.83, 245.07], [1105.88, 249.97], [1078.67, 249.94], [1049.35, 155.44], [1074.83, 155.44], [1105.88, 154.25], [1078.67, 154.26]]}]\n```" - } - ] - }, - { - "id": 1682, - "data_source": "Mono3DRefer", - "scene_name": "001334", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[965.21, 256.77], [998.66, 257.22], [1050.72, 271.78], [1011.2, 271.17], [965.21, 147.29], [998.66, 147.15], [1050.72, 142.71], [1011.2, 142.9]]}, {\"category\": \"cyclist\", \"corners_3d\": [[789.36, 208.51], [800.65, 208.56], [807.4, 210.6], [795.46, 210.54], [789.36, 167.32], [800.65, 167.31], [807.4, 166.99], [795.46, 167.0]]}]\n```" - } - ] - }, - { - "id": 1683, - "data_source": "Mono3DRefer", - "scene_name": "001334", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[965.21, 256.77], [998.66, 257.22], [1050.72, 271.78], [1011.2, 271.17], [965.21, 147.29], [998.66, 147.15], [1050.72, 142.71], [1011.2, 142.9]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1004.79, 247.2], [1035.06, 246.32], [1065.2, 250.0], [1033.77, 250.97], [1004.79, 157.54], [1035.06, 157.72], [1065.2, 156.96], [1033.77, 156.76]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1049.35, 245.04], [1074.83, 245.07], [1105.88, 249.97], [1078.67, 249.94], [1049.35, 155.44], [1074.83, 155.44], [1105.88, 154.25], [1078.67, 154.26]]}, {\"category\": \"cyclist\", \"corners_3d\": [[789.36, 208.51], [800.65, 208.56], [807.4, 210.6], [795.46, 210.54], [789.36, 167.32], [800.65, 167.31], [807.4, 166.99], [795.46, 167.0]]}]\n```" - } - ] - }, - { - "id": 1684, - "data_source": "Mono3DRefer", - "scene_name": "001337", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[243.16, 405.45], [3.56, 404.73], [276.52, 301.42], [409.65, 301.64], [243.16, 187.53], [3.56, 187.48], [276.52, 180.96], [409.65, 180.98]]}, {\"category\": \"car\", \"corners_3d\": [[767.07, 302.64], [900.17, 302.65], [1120.61, 401.31], [886.32, 401.26], [767.07, 193.11], [900.17, 193.11], [1120.61, 208.51], [886.32, 208.5]]}, {\"category\": \"car\", \"corners_3d\": [[774.93, 255.33], [700.61, 255.5], [677.89, 237.8], [736.34, 237.69], [774.93, 188.77], [700.61, 188.81], [677.89, 185.39], [736.34, 185.37]]}, {\"category\": \"car\", \"corners_3d\": [[501.2, 245.7], [427.51, 246.07], [454.15, 232.65], [514.28, 232.4], [501.2, 182.61], [427.51, 182.66], [454.15, 180.86], [514.28, 180.83]]}, {\"category\": \"car\", \"corners_3d\": [[662.81, 221.72], [707.91, 221.69], [725.69, 229.96], [672.96, 230.0], [662.81, 182.28], [707.91, 182.27], [725.69, 183.87], [672.96, 183.87]]}, {\"category\": \"car\", \"corners_3d\": [[704.58, 221.06], [660.57, 221.12], [652.24, 215.03], [690.7, 214.98], [704.58, 181.94], [660.57, 181.95], [652.24, 180.8], [690.7, 180.79]]}]\n```" - } - ] - }, - { - "id": 1685, - "data_source": "Mono3DRefer", - "scene_name": "001339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[986.75, 342.31], [1182.11, 353.81], [1601.62, 562.1], [1156.6, 512.52], [986.75, 201.07], [1182.11, 202.98], [1601.62, 237.66], [1156.6, 229.4]]}, {\"category\": \"car\", \"corners_3d\": [[851.98, 222.81], [893.89, 223.52], [909.41, 232.09], [860.28, 231.13], [851.98, 188.69], [893.89, 188.91], [909.41, 191.63], [860.28, 191.32]]}]\n```" - } - ] - }, - { - "id": 1686, - "data_source": "Mono3DRefer", - "scene_name": "001343", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[553.19, 215.8], [511.63, 215.88], [520.88, 210.62], [557.36, 210.56], [553.19, 172.18], [511.63, 172.18], [520.88, 172.26], [557.36, 172.26]]}, {\"category\": \"car\", \"corners_3d\": [[571.18, 196.9], [551.84, 196.89], [556.34, 195.38], [574.46, 195.39], [571.18, 175.38], [551.84, 175.38], [556.34, 175.22], [574.46, 175.22]]}]\n```" - } - ] - }, - { - "id": 1687, - "data_source": "Mono3DRefer", - "scene_name": "001352", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[89.81, 257.35], [10.91, 257.36], [133.5, 240.04], [196.22, 240.03], [89.81, 196.7], [10.91, 196.7], [133.5, 191.81], [196.22, 191.81]]}, {\"category\": \"car\", \"corners_3d\": [[432.15, 208.39], [403.71, 208.39], [419.56, 205.65], [445.8, 205.65], [432.15, 182.74], [403.71, 182.75], [419.56, 181.98], [445.8, 181.98]]}]\n```" - } - ] - }, - { - "id": 1688, - "data_source": "Mono3DRefer", - "scene_name": "001352", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[607.92, 196.75], [638.71, 196.84], [629.56, 202.85], [591.13, 202.7], [607.92, 163.36], [638.71, 163.32], [629.56, 160.94], [591.13, 161.0]]}]\n```" - } - ] - }, - { - "id": 1689, - "data_source": "Mono3DRefer", - "scene_name": "001352", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[89.81, 257.35], [10.91, 257.36], [133.5, 240.04], [196.22, 240.03], [89.81, 196.7], [10.91, 196.7], [133.5, 191.81], [196.22, 191.81]]}, {\"category\": \"car\", \"corners_3d\": [[432.15, 208.39], [403.71, 208.39], [419.56, 205.65], [445.8, 205.65], [432.15, 182.74], [403.71, 182.75], [419.56, 181.98], [445.8, 181.98]]}, {\"category\": \"bus\", \"corners_3d\": [[607.92, 196.75], [638.71, 196.84], [629.56, 202.85], [591.13, 202.7], [607.92, 163.36], [638.71, 163.32], [629.56, 160.94], [591.13, 161.0]]}]\n```" - } - ] - }, - { - "id": 1690, - "data_source": "Mono3DRefer", - "scene_name": "001354", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[445.97, 191.91], [437.04, 191.75], [467.67, 191.32], [476.68, 191.47], [445.97, 145.57], [437.04, 146.06], [467.67, 147.38], [476.68, 146.93]]}]\n```" - } - ] - }, - { - "id": 1691, - "data_source": "Mono3DRefer", - "scene_name": "001354", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1295.68, 564.39], [1256.24, 628.28], [1013.89, 537.13], [1074.61, 495.39], [1295.68, 105.75], [1256.24, 93.31], [1013.89, 111.06], [1074.61, 119.19]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[670.57, 247.59], [645.43, 248.58], [626.6, 244.63], [650.57, 243.75], [670.57, 144.03], [645.43, 143.49], [626.6, 145.64], [650.57, 146.12]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[574.01, 219.46], [545.38, 219.7], [543.18, 217.93], [570.54, 217.72], [574.01, 130.95], [545.38, 130.65], [543.18, 132.9], [570.54, 133.17]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[608.96, 219.49], [583.22, 219.98], [576.42, 218.77], [601.46, 218.31], [608.96, 142.71], [583.22, 142.24], [576.42, 143.41], [601.46, 143.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[665.63, 218.11], [645.99, 218.35], [638.05, 216.85], [656.95, 216.63], [665.63, 147.62], [645.99, 147.41], [638.05, 148.72], [656.95, 148.92]]}]\n```" - } - ] - }, - { - "id": 1692, - "data_source": "Mono3DRefer", - "scene_name": "001354", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1295.68, 564.39], [1256.24, 628.28], [1013.89, 537.13], [1074.61, 495.39], [1295.68, 105.75], [1256.24, 93.31], [1013.89, 111.06], [1074.61, 119.19]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[670.57, 247.59], [645.43, 248.58], [626.6, 244.63], [650.57, 243.75], [670.57, 144.03], [645.43, 143.49], [626.6, 145.64], [650.57, 146.12]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[574.01, 219.46], [545.38, 219.7], [543.18, 217.93], [570.54, 217.72], [574.01, 130.95], [545.38, 130.65], [543.18, 132.9], [570.54, 133.17]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[608.96, 219.49], [583.22, 219.98], [576.42, 218.77], [601.46, 218.31], [608.96, 142.71], [583.22, 142.24], [576.42, 143.41], [601.46, 143.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[665.63, 218.11], [645.99, 218.35], [638.05, 216.85], [656.95, 216.63], [665.63, 147.62], [645.99, 147.41], [638.05, 148.72], [656.95, 148.92]]}, {\"category\": \"cyclist\", \"corners_3d\": [[445.97, 191.91], [437.04, 191.75], [467.67, 191.32], [476.68, 191.47], [445.97, 145.57], [437.04, 146.06], [467.67, 147.38], [476.68, 146.93]]}]\n```" - } - ] - }, - { - "id": 1693, - "data_source": "Mono3DRefer", - "scene_name": "001357", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[245.49, 202.71], [259.89, 202.43], [270.63, 203.22], [255.85, 203.52], [245.49, 167.36], [259.89, 167.52], [270.63, 167.06], [255.85, 166.88]]}, {\"category\": \"cyclist\", \"corners_3d\": [[-7.91, 380.19], [98.18, 372.09], [0.81, 444.84], [-153.74, 460.51], [-7.91, 148.65], [98.18, 149.95], [0.81, 138.34], [-153.74, 135.84]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1182.87, 198.35], [1192.66, 198.78], [1142.53, 198.51], [1134.05, 198.09], [1182.87, 152.89], [1192.66, 152.23], [1142.53, 152.63], [1134.05, 153.28]]}]\n```" - } - ] - }, - { - "id": 1694, - "data_source": "Mono3DRefer", - "scene_name": "001357", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[487.44, 268.79], [520.73, 266.76], [544.55, 273.77], [509.09, 276.15], [487.44, 161.72], [520.73, 162.15], [544.55, 160.66], [509.09, 160.16]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[441.1, 272.98], [477.81, 270.35], [506.27, 278.44], [467.06, 281.56], [441.1, 151.66], [477.81, 152.48], [506.27, 149.96], [467.06, 148.98]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[642.19, 281.98], [591.62, 285.88], [562.36, 277.23], [609.92, 273.93], [642.19, 153.91], [591.62, 152.89], [562.36, 155.15], [609.92, 156.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[920.2, 197.41], [904.96, 197.38], [901.32, 197.13], [916.33, 197.15], [920.2, 155.38], [904.96, 155.43], [901.32, 155.8], [916.33, 155.76]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[915.39, 192.83], [926.66, 192.79], [937.94, 193.09], [926.42, 193.14], [915.39, 157.04], [926.66, 157.12], [937.94, 156.54], [926.42, 156.46]]}]\n```" - } - ] - }, - { - "id": 1695, - "data_source": "Mono3DRefer", - "scene_name": "001357", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[245.49, 202.71], [259.89, 202.43], [270.63, 203.22], [255.85, 203.52], [245.49, 167.36], [259.89, 167.52], [270.63, 167.06], [255.85, 166.88]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[487.44, 268.79], [520.73, 266.76], [544.55, 273.77], [509.09, 276.15], [487.44, 161.72], [520.73, 162.15], [544.55, 160.66], [509.09, 160.16]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[441.1, 272.98], [477.81, 270.35], [506.27, 278.44], [467.06, 281.56], [441.1, 151.66], [477.81, 152.48], [506.27, 149.96], [467.06, 148.98]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[642.19, 281.98], [591.62, 285.88], [562.36, 277.23], [609.92, 273.93], [642.19, 153.91], [591.62, 152.89], [562.36, 155.15], [609.92, 156.02]]}, {\"category\": \"cyclist\", \"corners_3d\": [[-7.91, 380.19], [98.18, 372.09], [0.81, 444.84], [-153.74, 460.51], [-7.91, 148.65], [98.18, 149.95], [0.81, 138.34], [-153.74, 135.84]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[920.2, 197.41], [904.96, 197.38], [901.32, 197.13], [916.33, 197.15], [920.2, 155.38], [904.96, 155.43], [901.32, 155.8], [916.33, 155.76]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[915.39, 192.83], [926.66, 192.79], [937.94, 193.09], [926.42, 193.14], [915.39, 157.04], [926.66, 157.12], [937.94, 156.54], [926.42, 156.46]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1182.87, 198.35], [1192.66, 198.78], [1142.53, 198.51], [1134.05, 198.09], [1182.87, 152.89], [1192.66, 152.23], [1142.53, 152.63], [1134.05, 153.28]]}]\n```" - } - ] - }, - { - "id": 1696, - "data_source": "Mono3DRefer", - "scene_name": "001362", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[510.6, 205.97], [504.27, 207.31], [427.94, 207.01], [437.16, 205.7], [510.6, 183.22], [504.27, 183.64], [427.94, 183.55], [437.16, 183.13]]}]\n```" - } - ] - }, - { - "id": 1697, - "data_source": "Mono3DRefer", - "scene_name": "001367", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1012.0, 196.24], [1055.81, 196.27], [1121.55, 199.84], [1070.97, 199.8], [1012.0, 144.92], [1055.81, 144.89], [1121.55, 140.63], [1070.97, 140.67]]}, {\"category\": \"van\", \"corners_3d\": [[933.0, 195.86], [976.17, 195.94], [1027.3, 199.6], [977.11, 199.49], [933.0, 151.27], [976.17, 151.19], [1027.3, 147.76], [977.11, 147.86]]}]\n```" - } - ] - }, - { - "id": 1698, - "data_source": "Mono3DRefer", - "scene_name": "001367", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[814.34, 176.89], [844.16, 176.96], [831.45, 177.61], [797.19, 177.52], [814.34, 142.19], [844.16, 141.7], [831.45, 136.73], [797.19, 137.39]]}]\n```" - } - ] - }, - { - "id": 1699, - "data_source": "Mono3DRefer", - "scene_name": "001367", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-153.21, 563.73], [-585.4, 577.12], [91.12, 339.51], [263.41, 337.19], [-153.21, 148.93], [-585.4, 148.11], [91.12, 162.65], [263.41, 162.8]]}, {\"category\": \"car\", \"corners_3d\": [[418.12, 288.61], [322.9, 285.88], [425.02, 257.72], [497.93, 259.25], [418.12, 171.4], [322.9, 171.43], [425.02, 171.79], [497.93, 171.77]]}, {\"category\": \"car\", \"corners_3d\": [[845.61, 316.2], [997.22, 317.71], [1344.09, 460.86], [1038.6, 454.97], [845.61, 174.58], [997.22, 174.6], [1344.09, 176.33], [1038.6, 176.25]]}, {\"category\": \"car\", \"corners_3d\": [[776.96, 260.04], [862.53, 261.36], [934.13, 301.45], [809.08, 298.68], [776.96, 180.67], [862.53, 180.79], [934.13, 184.38], [809.08, 184.13]]}, {\"category\": \"car\", \"corners_3d\": [[650.64, 229.44], [598.34, 227.66], [648.1, 219.46], [693.73, 220.74], [650.64, 175.3], [598.34, 175.23], [648.1, 174.87], [693.73, 174.93]]}, {\"category\": \"car\", \"corners_3d\": [[704.32, 200.89], [671.87, 200.45], [692.03, 197.67], [721.46, 198.03], [704.32, 169.5], [671.87, 169.55], [692.03, 169.88], [721.46, 169.84]]}]\n```" - } - ] - }, - { - "id": 1700, - "data_source": "Mono3DRefer", - "scene_name": "001367", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1012.0, 196.24], [1055.81, 196.27], [1121.55, 199.84], [1070.97, 199.8], [1012.0, 144.92], [1055.81, 144.89], [1121.55, 140.63], [1070.97, 140.67]]}, {\"category\": \"van\", \"corners_3d\": [[933.0, 195.86], [976.17, 195.94], [1027.3, 199.6], [977.11, 199.49], [933.0, 151.27], [976.17, 151.19], [1027.3, 147.76], [977.11, 147.86]]}, {\"category\": \"truck\", \"corners_3d\": [[814.34, 176.89], [844.16, 176.96], [831.45, 177.61], [797.19, 177.52], [814.34, 142.19], [844.16, 141.7], [831.45, 136.73], [797.19, 137.39]]}]\n```" - } - ] - }, - { - "id": 1701, - "data_source": "Mono3DRefer", - "scene_name": "001367", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-153.21, 563.73], [-585.4, 577.12], [91.12, 339.51], [263.41, 337.19], [-153.21, 148.93], [-585.4, 148.11], [91.12, 162.65], [263.41, 162.8]]}, {\"category\": \"car\", \"corners_3d\": [[418.12, 288.61], [322.9, 285.88], [425.02, 257.72], [497.93, 259.25], [418.12, 171.4], [322.9, 171.43], [425.02, 171.79], [497.93, 171.77]]}, {\"category\": \"car\", \"corners_3d\": [[845.61, 316.2], [997.22, 317.71], [1344.09, 460.86], [1038.6, 454.97], [845.61, 174.58], [997.22, 174.6], [1344.09, 176.33], [1038.6, 176.25]]}, {\"category\": \"car\", \"corners_3d\": [[776.96, 260.04], [862.53, 261.36], [934.13, 301.45], [809.08, 298.68], [776.96, 180.67], [862.53, 180.79], [934.13, 184.38], [809.08, 184.13]]}, {\"category\": \"car\", \"corners_3d\": [[650.64, 229.44], [598.34, 227.66], [648.1, 219.46], [693.73, 220.74], [650.64, 175.3], [598.34, 175.23], [648.1, 174.87], [693.73, 174.93]]}, {\"category\": \"van\", \"corners_3d\": [[1012.0, 196.24], [1055.81, 196.27], [1121.55, 199.84], [1070.97, 199.8], [1012.0, 144.92], [1055.81, 144.89], [1121.55, 140.63], [1070.97, 140.67]]}, {\"category\": \"van\", \"corners_3d\": [[933.0, 195.86], [976.17, 195.94], [1027.3, 199.6], [977.11, 199.49], [933.0, 151.27], [976.17, 151.19], [1027.3, 147.76], [977.11, 147.86]]}, {\"category\": \"car\", \"corners_3d\": [[704.32, 200.89], [671.87, 200.45], [692.03, 197.67], [721.46, 198.03], [704.32, 169.5], [671.87, 169.55], [692.03, 169.88], [721.46, 169.84]]}]\n```" - } - ] - }, - { - "id": 1702, - "data_source": "Mono3DRefer", - "scene_name": "001367", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-153.21, 563.73], [-585.4, 577.12], [91.12, 339.51], [263.41, 337.19], [-153.21, 148.93], [-585.4, 148.11], [91.12, 162.65], [263.41, 162.8]]}, {\"category\": \"car\", \"corners_3d\": [[418.12, 288.61], [322.9, 285.88], [425.02, 257.72], [497.93, 259.25], [418.12, 171.4], [322.9, 171.43], [425.02, 171.79], [497.93, 171.77]]}, {\"category\": \"car\", \"corners_3d\": [[845.61, 316.2], [997.22, 317.71], [1344.09, 460.86], [1038.6, 454.97], [845.61, 174.58], [997.22, 174.6], [1344.09, 176.33], [1038.6, 176.25]]}, {\"category\": \"car\", \"corners_3d\": [[776.96, 260.04], [862.53, 261.36], [934.13, 301.45], [809.08, 298.68], [776.96, 180.67], [862.53, 180.79], [934.13, 184.38], [809.08, 184.13]]}, {\"category\": \"car\", \"corners_3d\": [[650.64, 229.44], [598.34, 227.66], [648.1, 219.46], [693.73, 220.74], [650.64, 175.3], [598.34, 175.23], [648.1, 174.87], [693.73, 174.93]]}, {\"category\": \"car\", \"corners_3d\": [[704.32, 200.89], [671.87, 200.45], [692.03, 197.67], [721.46, 198.03], [704.32, 169.5], [671.87, 169.55], [692.03, 169.88], [721.46, 169.84]]}, {\"category\": \"truck\", \"corners_3d\": [[814.34, 176.89], [844.16, 176.96], [831.45, 177.61], [797.19, 177.52], [814.34, 142.19], [844.16, 141.7], [831.45, 136.73], [797.19, 137.39]]}]\n```" - } - ] - }, - { - "id": 1703, - "data_source": "Mono3DRefer", - "scene_name": "001367", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-153.21, 563.73], [-585.4, 577.12], [91.12, 339.51], [263.41, 337.19], [-153.21, 148.93], [-585.4, 148.11], [91.12, 162.65], [263.41, 162.8]]}, {\"category\": \"car\", \"corners_3d\": [[418.12, 288.61], [322.9, 285.88], [425.02, 257.72], [497.93, 259.25], [418.12, 171.4], [322.9, 171.43], [425.02, 171.79], [497.93, 171.77]]}, {\"category\": \"car\", \"corners_3d\": [[845.61, 316.2], [997.22, 317.71], [1344.09, 460.86], [1038.6, 454.97], [845.61, 174.58], [997.22, 174.6], [1344.09, 176.33], [1038.6, 176.25]]}, {\"category\": \"car\", \"corners_3d\": [[776.96, 260.04], [862.53, 261.36], [934.13, 301.45], [809.08, 298.68], [776.96, 180.67], [862.53, 180.79], [934.13, 184.38], [809.08, 184.13]]}, {\"category\": \"car\", \"corners_3d\": [[650.64, 229.44], [598.34, 227.66], [648.1, 219.46], [693.73, 220.74], [650.64, 175.3], [598.34, 175.23], [648.1, 174.87], [693.73, 174.93]]}, {\"category\": \"van\", \"corners_3d\": [[1012.0, 196.24], [1055.81, 196.27], [1121.55, 199.84], [1070.97, 199.8], [1012.0, 144.92], [1055.81, 144.89], [1121.55, 140.63], [1070.97, 140.67]]}, {\"category\": \"van\", \"corners_3d\": [[933.0, 195.86], [976.17, 195.94], [1027.3, 199.6], [977.11, 199.49], [933.0, 151.27], [976.17, 151.19], [1027.3, 147.76], [977.11, 147.86]]}, {\"category\": \"car\", \"corners_3d\": [[704.32, 200.89], [671.87, 200.45], [692.03, 197.67], [721.46, 198.03], [704.32, 169.5], [671.87, 169.55], [692.03, 169.88], [721.46, 169.84]]}, {\"category\": \"truck\", \"corners_3d\": [[814.34, 176.89], [844.16, 176.96], [831.45, 177.61], [797.19, 177.52], [814.34, 142.19], [844.16, 141.7], [831.45, 136.73], [797.19, 137.39]]}]\n```" - } - ] - }, - { - "id": 1704, - "data_source": "Mono3DRefer", - "scene_name": "001369", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-104.35, 337.27], [-242.93, 336.66], [43.41, 282.52], [136.79, 282.79], [-104.35, 209.37], [-242.93, 209.23], [43.41, 197.21], [136.79, 197.27]]}, {\"category\": \"car\", \"corners_3d\": [[228.43, 263.32], [152.2, 263.24], [245.75, 245.01], [306.67, 245.07], [228.43, 177.6], [152.2, 177.59], [245.75, 176.64], [306.67, 176.64]]}, {\"category\": \"car\", \"corners_3d\": [[324.02, 241.97], [275.52, 241.44], [333.96, 231.73], [375.95, 232.13], [324.02, 190.57], [275.52, 190.43], [333.96, 187.94], [375.95, 188.04]]}, {\"category\": \"car\", \"corners_3d\": [[385.31, 230.51], [336.29, 230.35], [380.84, 221.86], [422.71, 221.97], [385.31, 179.54], [336.29, 179.52], [380.84, 178.54], [422.71, 178.55]]}, {\"category\": \"car\", \"corners_3d\": [[460.16, 215.21], [425.89, 215.13], [447.87, 210.75], [478.61, 210.81], [460.16, 183.28], [425.89, 183.26], [447.87, 182.18], [478.61, 182.2]]}, {\"category\": \"car\", \"corners_3d\": [[484.52, 209.11], [455.93, 209.02], [473.39, 205.81], [499.47, 205.88], [484.52, 182.27], [455.93, 182.25], [473.39, 181.42], [499.47, 181.44]]}]\n```" - } - ] - }, - { - "id": 1705, - "data_source": "Mono3DRefer", - "scene_name": "001369", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[688.82, 184.05], [707.86, 184.06], [731.82, 187.96], [706.12, 187.94], [688.82, 157.42], [707.86, 157.4], [731.82, 152.02], [706.12, 152.05]]}]\n```" - } - ] - }, - { - "id": 1706, - "data_source": "Mono3DRefer", - "scene_name": "001369", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-104.35, 337.27], [-242.93, 336.66], [43.41, 282.52], [136.79, 282.79], [-104.35, 209.37], [-242.93, 209.23], [43.41, 197.21], [136.79, 197.27]]}, {\"category\": \"tram\", \"corners_3d\": [[688.82, 184.05], [707.86, 184.06], [731.82, 187.96], [706.12, 187.94], [688.82, 157.42], [707.86, 157.4], [731.82, 152.02], [706.12, 152.05]]}, {\"category\": \"car\", \"corners_3d\": [[228.43, 263.32], [152.2, 263.24], [245.75, 245.01], [306.67, 245.07], [228.43, 177.6], [152.2, 177.59], [245.75, 176.64], [306.67, 176.64]]}, {\"category\": \"car\", \"corners_3d\": [[324.02, 241.97], [275.52, 241.44], [333.96, 231.73], [375.95, 232.13], [324.02, 190.57], [275.52, 190.43], [333.96, 187.94], [375.95, 188.04]]}, {\"category\": \"car\", \"corners_3d\": [[385.31, 230.51], [336.29, 230.35], [380.84, 221.86], [422.71, 221.97], [385.31, 179.54], [336.29, 179.52], [380.84, 178.54], [422.71, 178.55]]}, {\"category\": \"car\", \"corners_3d\": [[460.16, 215.21], [425.89, 215.13], [447.87, 210.75], [478.61, 210.81], [460.16, 183.28], [425.89, 183.26], [447.87, 182.18], [478.61, 182.2]]}, {\"category\": \"car\", \"corners_3d\": [[484.52, 209.11], [455.93, 209.02], [473.39, 205.81], [499.47, 205.88], [484.52, 182.27], [455.93, 182.25], [473.39, 181.42], [499.47, 181.44]]}]\n```" - } - ] - }, - { - "id": 1707, - "data_source": "Mono3DRefer", - "scene_name": "001370", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[346.26, 326.09], [330.13, 340.11], [259.15, 342.01], [281.46, 327.68], [346.26, 148.07], [330.13, 145.8], [259.15, 145.49], [281.46, 147.81]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[663.87, 210.75], [681.37, 211.19], [675.26, 212.0], [657.47, 211.54], [663.87, 176.61], [681.37, 176.66], [675.26, 176.74], [657.47, 176.69]]}]\n```" - } - ] - }, - { - "id": 1708, - "data_source": "Mono3DRefer", - "scene_name": "001370", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[731.69, 248.23], [817.81, 248.68], [884.72, 278.25], [764.74, 277.38], [731.69, 151.36], [817.81, 151.23], [884.72, 142.8], [764.74, 143.05]]}, {\"category\": \"van\", \"corners_3d\": [[228.57, 222.67], [171.39, 222.67], [247.1, 214.05], [294.39, 214.05], [228.57, 160.02], [171.39, 160.02], [247.1, 162.24], [294.39, 162.24]]}]\n```" - } - ] - }, - { - "id": 1709, - "data_source": "Mono3DRefer", - "scene_name": "001370", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[589.95, 229.51], [649.8, 229.51], [657.69, 240.78], [585.93, 240.77], [589.95, 181.34], [649.8, 181.34], [657.69, 183.03], [585.93, 183.03]]}, {\"category\": \"car\", \"corners_3d\": [[8.82, 259.98], [-88.91, 259.87], [76.43, 239.47], [151.38, 239.53], [8.82, 168.82], [-88.91, 168.83], [76.43, 169.77], [151.38, 169.77]]}, {\"category\": \"car\", \"corners_3d\": [[314.09, 239.58], [378.93, 234.99], [470.53, 244.93], [402.36, 251.17], [314.09, 169.76], [378.93, 169.98], [470.53, 169.52], [402.36, 169.23]]}, {\"category\": \"car\", \"corners_3d\": [[351.16, 229.94], [401.1, 226.21], [493.13, 232.19], [444.36, 236.84], [351.16, 176.97], [401.1, 176.7], [493.13, 177.14], [444.36, 177.47]]}, {\"category\": \"car\", \"corners_3d\": [[391.99, 225.27], [432.56, 222.42], [523.66, 228.0], [484.09, 231.55], [391.99, 180.83], [432.56, 180.4], [523.66, 181.25], [484.09, 181.79]]}, {\"category\": \"car\", \"corners_3d\": [[464.95, 210.62], [496.24, 209.32], [547.38, 212.58], [515.18, 214.14], [464.95, 174.76], [496.24, 174.69], [547.38, 174.86], [515.18, 174.94]]}, {\"category\": \"car\", \"corners_3d\": [[507.72, 198.19], [526.72, 197.48], [569.78, 198.77], [551.07, 199.56], [507.72, 174.85], [526.72, 174.79], [569.78, 174.89], [551.07, 174.96]]}, {\"category\": \"car\", \"corners_3d\": [[690.2, 216.49], [734.66, 216.49], [756.98, 224.31], [704.55, 224.3], [690.2, 179.39], [734.66, 179.39], [756.98, 180.56], [704.55, 180.56]]}, {\"category\": \"car\", \"corners_3d\": [[674.47, 207.64], [710.91, 207.62], [726.99, 212.8], [685.13, 212.82], [674.47, 172.42], [710.91, 172.42], [726.99, 172.36], [685.13, 172.36]]}, {\"category\": \"car\", \"corners_3d\": [[591.23, 209.99], [627.51, 209.99], [629.63, 214.52], [588.93, 214.52], [591.23, 176.97], [627.51, 176.97], [629.63, 177.48], [588.93, 177.48]]}, {\"category\": \"car\", \"corners_3d\": [[301.06, 204.54], [262.75, 204.44], [307.82, 200.78], [341.81, 200.86], [301.06, 171.73], [262.75, 171.73], [307.82, 171.86], [341.81, 171.86]]}]\n```" - } - ] - }, - { - "id": 1710, - "data_source": "Mono3DRefer", - "scene_name": "001370", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[731.69, 248.23], [817.81, 248.68], [884.72, 278.25], [764.74, 277.38], [731.69, 151.36], [817.81, 151.23], [884.72, 142.8], [764.74, 143.05]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[346.26, 326.09], [330.13, 340.11], [259.15, 342.01], [281.46, 327.68], [346.26, 148.07], [330.13, 145.8], [259.15, 145.49], [281.46, 147.81]]}, {\"category\": \"van\", \"corners_3d\": [[228.57, 222.67], [171.39, 222.67], [247.1, 214.05], [294.39, 214.05], [228.57, 160.02], [171.39, 160.02], [247.1, 162.24], [294.39, 162.24]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[663.87, 210.75], [681.37, 211.19], [675.26, 212.0], [657.47, 211.54], [663.87, 176.61], [681.37, 176.66], [675.26, 176.74], [657.47, 176.69]]}]\n```" - } - ] - }, - { - "id": 1711, - "data_source": "Mono3DRefer", - "scene_name": "001370", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[589.95, 229.51], [649.8, 229.51], [657.69, 240.78], [585.93, 240.77], [589.95, 181.34], [649.8, 181.34], [657.69, 183.03], [585.93, 183.03]]}, {\"category\": \"car\", \"corners_3d\": [[8.82, 259.98], [-88.91, 259.87], [76.43, 239.47], [151.38, 239.53], [8.82, 168.82], [-88.91, 168.83], [76.43, 169.77], [151.38, 169.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[346.26, 326.09], [330.13, 340.11], [259.15, 342.01], [281.46, 327.68], [346.26, 148.07], [330.13, 145.8], [259.15, 145.49], [281.46, 147.81]]}, {\"category\": \"car\", \"corners_3d\": [[314.09, 239.58], [378.93, 234.99], [470.53, 244.93], [402.36, 251.17], [314.09, 169.76], [378.93, 169.98], [470.53, 169.52], [402.36, 169.23]]}, {\"category\": \"car\", \"corners_3d\": [[351.16, 229.94], [401.1, 226.21], [493.13, 232.19], [444.36, 236.84], [351.16, 176.97], [401.1, 176.7], [493.13, 177.14], [444.36, 177.47]]}, {\"category\": \"car\", \"corners_3d\": [[391.99, 225.27], [432.56, 222.42], [523.66, 228.0], [484.09, 231.55], [391.99, 180.83], [432.56, 180.4], [523.66, 181.25], [484.09, 181.79]]}, {\"category\": \"car\", \"corners_3d\": [[464.95, 210.62], [496.24, 209.32], [547.38, 212.58], [515.18, 214.14], [464.95, 174.76], [496.24, 174.69], [547.38, 174.86], [515.18, 174.94]]}, {\"category\": \"car\", \"corners_3d\": [[507.72, 198.19], [526.72, 197.48], [569.78, 198.77], [551.07, 199.56], [507.72, 174.85], [526.72, 174.79], [569.78, 174.89], [551.07, 174.96]]}, {\"category\": \"car\", \"corners_3d\": [[690.2, 216.49], [734.66, 216.49], [756.98, 224.31], [704.55, 224.3], [690.2, 179.39], [734.66, 179.39], [756.98, 180.56], [704.55, 180.56]]}, {\"category\": \"car\", \"corners_3d\": [[674.47, 207.64], [710.91, 207.62], [726.99, 212.8], [685.13, 212.82], [674.47, 172.42], [710.91, 172.42], [726.99, 172.36], [685.13, 172.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[663.87, 210.75], [681.37, 211.19], [675.26, 212.0], [657.47, 211.54], [663.87, 176.61], [681.37, 176.66], [675.26, 176.74], [657.47, 176.69]]}, {\"category\": \"car\", \"corners_3d\": [[591.23, 209.99], [627.51, 209.99], [629.63, 214.52], [588.93, 214.52], [591.23, 176.97], [627.51, 176.97], [629.63, 177.48], [588.93, 177.48]]}, {\"category\": \"car\", \"corners_3d\": [[301.06, 204.54], [262.75, 204.44], [307.82, 200.78], [341.81, 200.86], [301.06, 171.73], [262.75, 171.73], [307.82, 171.86], [341.81, 171.86]]}]\n```" - } - ] - }, - { - "id": 1712, - "data_source": "Mono3DRefer", - "scene_name": "001370", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[589.95, 229.51], [649.8, 229.51], [657.69, 240.78], [585.93, 240.77], [589.95, 181.34], [649.8, 181.34], [657.69, 183.03], [585.93, 183.03]]}, {\"category\": \"car\", \"corners_3d\": [[8.82, 259.98], [-88.91, 259.87], [76.43, 239.47], [151.38, 239.53], [8.82, 168.82], [-88.91, 168.83], [76.43, 169.77], [151.38, 169.77]]}, {\"category\": \"van\", \"corners_3d\": [[731.69, 248.23], [817.81, 248.68], [884.72, 278.25], [764.74, 277.38], [731.69, 151.36], [817.81, 151.23], [884.72, 142.8], [764.74, 143.05]]}, {\"category\": \"van\", \"corners_3d\": [[228.57, 222.67], [171.39, 222.67], [247.1, 214.05], [294.39, 214.05], [228.57, 160.02], [171.39, 160.02], [247.1, 162.24], [294.39, 162.24]]}, {\"category\": \"car\", \"corners_3d\": [[314.09, 239.58], [378.93, 234.99], [470.53, 244.93], [402.36, 251.17], [314.09, 169.76], [378.93, 169.98], [470.53, 169.52], [402.36, 169.23]]}, {\"category\": \"car\", \"corners_3d\": [[351.16, 229.94], [401.1, 226.21], [493.13, 232.19], [444.36, 236.84], [351.16, 176.97], [401.1, 176.7], [493.13, 177.14], [444.36, 177.47]]}, {\"category\": \"car\", \"corners_3d\": [[391.99, 225.27], [432.56, 222.42], [523.66, 228.0], [484.09, 231.55], [391.99, 180.83], [432.56, 180.4], [523.66, 181.25], [484.09, 181.79]]}, {\"category\": \"car\", \"corners_3d\": [[464.95, 210.62], [496.24, 209.32], [547.38, 212.58], [515.18, 214.14], [464.95, 174.76], [496.24, 174.69], [547.38, 174.86], [515.18, 174.94]]}, {\"category\": \"car\", \"corners_3d\": [[507.72, 198.19], [526.72, 197.48], [569.78, 198.77], [551.07, 199.56], [507.72, 174.85], [526.72, 174.79], [569.78, 174.89], [551.07, 174.96]]}, {\"category\": \"car\", \"corners_3d\": [[690.2, 216.49], [734.66, 216.49], [756.98, 224.31], [704.55, 224.3], [690.2, 179.39], [734.66, 179.39], [756.98, 180.56], [704.55, 180.56]]}, {\"category\": \"car\", \"corners_3d\": [[674.47, 207.64], [710.91, 207.62], [726.99, 212.8], [685.13, 212.82], [674.47, 172.42], [710.91, 172.42], [726.99, 172.36], [685.13, 172.36]]}, {\"category\": \"car\", \"corners_3d\": [[591.23, 209.99], [627.51, 209.99], [629.63, 214.52], [588.93, 214.52], [591.23, 176.97], [627.51, 176.97], [629.63, 177.48], [588.93, 177.48]]}, {\"category\": \"car\", \"corners_3d\": [[301.06, 204.54], [262.75, 204.44], [307.82, 200.78], [341.81, 200.86], [301.06, 171.73], [262.75, 171.73], [307.82, 171.86], [341.81, 171.86]]}]\n```" - } - ] - }, - { - "id": 1713, - "data_source": "Mono3DRefer", - "scene_name": "001370", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[589.95, 229.51], [649.8, 229.51], [657.69, 240.78], [585.93, 240.77], [589.95, 181.34], [649.8, 181.34], [657.69, 183.03], [585.93, 183.03]]}, {\"category\": \"car\", \"corners_3d\": [[8.82, 259.98], [-88.91, 259.87], [76.43, 239.47], [151.38, 239.53], [8.82, 168.82], [-88.91, 168.83], [76.43, 169.77], [151.38, 169.77]]}, {\"category\": \"van\", \"corners_3d\": [[731.69, 248.23], [817.81, 248.68], [884.72, 278.25], [764.74, 277.38], [731.69, 151.36], [817.81, 151.23], [884.72, 142.8], [764.74, 143.05]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[346.26, 326.09], [330.13, 340.11], [259.15, 342.01], [281.46, 327.68], [346.26, 148.07], [330.13, 145.8], [259.15, 145.49], [281.46, 147.81]]}, {\"category\": \"van\", \"corners_3d\": [[228.57, 222.67], [171.39, 222.67], [247.1, 214.05], [294.39, 214.05], [228.57, 160.02], [171.39, 160.02], [247.1, 162.24], [294.39, 162.24]]}, {\"category\": \"car\", \"corners_3d\": [[314.09, 239.58], [378.93, 234.99], [470.53, 244.93], [402.36, 251.17], [314.09, 169.76], [378.93, 169.98], [470.53, 169.52], [402.36, 169.23]]}, {\"category\": \"car\", \"corners_3d\": [[351.16, 229.94], [401.1, 226.21], [493.13, 232.19], [444.36, 236.84], [351.16, 176.97], [401.1, 176.7], [493.13, 177.14], [444.36, 177.47]]}, {\"category\": \"car\", \"corners_3d\": [[391.99, 225.27], [432.56, 222.42], [523.66, 228.0], [484.09, 231.55], [391.99, 180.83], [432.56, 180.4], [523.66, 181.25], [484.09, 181.79]]}, {\"category\": \"car\", \"corners_3d\": [[464.95, 210.62], [496.24, 209.32], [547.38, 212.58], [515.18, 214.14], [464.95, 174.76], [496.24, 174.69], [547.38, 174.86], [515.18, 174.94]]}, {\"category\": \"car\", \"corners_3d\": [[507.72, 198.19], [526.72, 197.48], [569.78, 198.77], [551.07, 199.56], [507.72, 174.85], [526.72, 174.79], [569.78, 174.89], [551.07, 174.96]]}, {\"category\": \"car\", \"corners_3d\": [[690.2, 216.49], [734.66, 216.49], [756.98, 224.31], [704.55, 224.3], [690.2, 179.39], [734.66, 179.39], [756.98, 180.56], [704.55, 180.56]]}, {\"category\": \"car\", \"corners_3d\": [[674.47, 207.64], [710.91, 207.62], [726.99, 212.8], [685.13, 212.82], [674.47, 172.42], [710.91, 172.42], [726.99, 172.36], [685.13, 172.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[663.87, 210.75], [681.37, 211.19], [675.26, 212.0], [657.47, 211.54], [663.87, 176.61], [681.37, 176.66], [675.26, 176.74], [657.47, 176.69]]}, {\"category\": \"car\", \"corners_3d\": [[591.23, 209.99], [627.51, 209.99], [629.63, 214.52], [588.93, 214.52], [591.23, 176.97], [627.51, 176.97], [629.63, 177.48], [588.93, 177.48]]}, {\"category\": \"car\", \"corners_3d\": [[301.06, 204.54], [262.75, 204.44], [307.82, 200.78], [341.81, 200.86], [301.06, 171.73], [262.75, 171.73], [307.82, 171.86], [341.81, 171.86]]}]\n```" - } - ] - }, - { - "id": 1714, - "data_source": "Mono3DRefer", - "scene_name": "001376", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[777.51, 320.78], [905.41, 320.53], [1212.21, 470.31], [955.17, 471.3], [777.51, 191.53], [905.41, 191.5], [1212.21, 210.41], [955.17, 210.53]]}, {\"category\": \"car\", \"corners_3d\": [[483.53, 281.01], [396.86, 281.02], [447.51, 255.19], [513.48, 255.19], [483.53, 202.64], [396.86, 202.64], [447.51, 195.53], [513.48, 195.53]]}, {\"category\": \"car\", \"corners_3d\": [[687.16, 265.8], [772.71, 265.26], [858.25, 305.03], [736.27, 306.15], [687.16, 190.27], [772.71, 190.16], [858.25, 197.62], [736.27, 197.83]]}, {\"category\": \"car\", \"corners_3d\": [[556.47, 215.9], [524.4, 215.95], [530.31, 211.9], [559.37, 211.85], [556.47, 187.06], [524.4, 187.08], [530.31, 185.74], [559.37, 185.72]]}, {\"category\": \"car\", \"corners_3d\": [[631.48, 205.98], [658.53, 205.97], [664.18, 209.34], [634.38, 209.35], [631.48, 181.54], [658.53, 181.54], [664.18, 182.42], [634.38, 182.43]]}]\n```" - } - ] - }, - { - "id": 1715, - "data_source": "Mono3DRefer", - "scene_name": "001382", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1012.32, 287.3], [1148.34, 296.68], [1208.41, 356.23], [1007.55, 336.39], [1012.32, 193.58], [1148.34, 195.28], [1208.41, 206.07], [1007.55, 202.47]]}, {\"category\": \"car\", \"corners_3d\": [[1002.7, 220.87], [1046.27, 222.04], [1055.1, 229.24], [1005.09, 227.7], [1002.7, 188.39], [1046.27, 188.77], [1055.1, 191.1], [1005.09, 190.6]]}]\n```" - } - ] - }, - { - "id": 1716, - "data_source": "Mono3DRefer", - "scene_name": "001385", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[566.02, 204.9], [537.51, 204.9], [543.42, 202.25], [569.57, 202.25], [566.02, 181.54], [537.51, 181.54], [543.42, 180.83], [569.57, 180.83]]}]\n```" - } - ] - }, - { - "id": 1717, - "data_source": "Mono3DRefer", - "scene_name": "001390", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[632.35, 194.21], [653.01, 194.2], [657.05, 195.53], [635.1, 195.55], [632.35, 175.62], [653.01, 175.62], [657.05, 175.79], [635.1, 175.79]]}, {\"category\": \"car\", \"corners_3d\": [[683.88, 194.92], [705.73, 194.88], [717.73, 196.78], [694.01, 196.83], [683.88, 175.04], [705.73, 175.04], [717.73, 175.23], [694.01, 175.23]]}, {\"category\": \"car\", \"corners_3d\": [[800.45, 200.1], [816.12, 199.62], [866.45, 201.57], [850.58, 202.12], [800.45, 174.73], [816.12, 174.7], [866.45, 174.83], [850.58, 174.87]]}]\n```" - } - ] - }, - { - "id": 1718, - "data_source": "Mono3DRefer", - "scene_name": "001390", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[768.36, 216.49], [816.4, 216.41], [861.24, 224.97], [803.83, 225.1], [768.36, 161.33], [816.4, 161.35], [861.24, 159.09], [803.83, 159.06]]}]\n```" - } - ] - }, - { - "id": 1719, - "data_source": "Mono3DRefer", - "scene_name": "001390", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[632.35, 194.21], [653.01, 194.2], [657.05, 195.53], [635.1, 195.55], [632.35, 175.62], [653.01, 175.62], [657.05, 175.79], [635.1, 175.79]]}, {\"category\": \"car\", \"corners_3d\": [[683.88, 194.92], [705.73, 194.88], [717.73, 196.78], [694.01, 196.83], [683.88, 175.04], [705.73, 175.04], [717.73, 175.23], [694.01, 175.23]]}, {\"category\": \"van\", \"corners_3d\": [[768.36, 216.49], [816.4, 216.41], [861.24, 224.97], [803.83, 225.1], [768.36, 161.33], [816.4, 161.35], [861.24, 159.09], [803.83, 159.06]]}, {\"category\": \"car\", \"corners_3d\": [[800.45, 200.1], [816.12, 199.62], [866.45, 201.57], [850.58, 202.12], [800.45, 174.73], [816.12, 174.7], [866.45, 174.83], [850.58, 174.87]]}]\n```" - } - ] - }, - { - "id": 1720, - "data_source": "Mono3DRefer", - "scene_name": "001392", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[492.25, 214.5], [450.43, 214.62], [465.58, 209.45], [502.2, 209.35], [492.25, 176.16], [450.43, 176.17], [465.58, 175.76], [502.2, 175.75]]}, {\"category\": \"car\", \"corners_3d\": [[314.61, 302.81], [188.56, 302.82], [302.17, 267.7], [394.15, 267.69], [314.61, 187.98], [188.56, 187.98], [302.17, 183.89], [394.15, 183.89]]}, {\"category\": \"car\", \"corners_3d\": [[403.38, 268.4], [306.05, 268.79], [378.18, 244.19], [450.4, 243.98], [403.38, 170.59], [306.05, 170.58], [378.18, 171.16], [450.4, 171.17]]}, {\"category\": \"car\", \"corners_3d\": [[474.78, 230.8], [415.89, 231.27], [436.46, 221.32], [485.24, 221.0], [474.78, 183.56], [415.89, 183.65], [436.46, 181.81], [485.24, 181.75]]}]\n```" - } - ] - }, - { - "id": 1721, - "data_source": "Mono3DRefer", - "scene_name": "001399", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[568.76, 201.15], [542.53, 201.15], [547.62, 198.98], [571.84, 198.98], [568.76, 179.66], [542.53, 179.67], [547.62, 179.14], [571.84, 179.14]]}]\n```" - } - ] - }, - { - "id": 1722, - "data_source": "Mono3DRefer", - "scene_name": "001403", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[797.31, 225.38], [878.56, 225.15], [1041.61, 253.83], [916.63, 254.39], [797.31, 110.95], [878.56, 111.23], [1041.61, 77.43], [916.63, 76.77]]}]\n```" - } - ] - }, - { - "id": 1723, - "data_source": "Mono3DRefer", - "scene_name": "001403", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[652.44, 211.81], [687.28, 211.75], [699.22, 216.45], [660.18, 216.52], [652.44, 179.42], [687.28, 179.41], [699.22, 180.2], [660.18, 180.21]]}, {\"category\": \"car\", \"corners_3d\": [[716.9, 240.72], [783.45, 240.53], [839.22, 259.89], [753.76, 260.2], [716.9, 178.63], [783.45, 178.62], [839.22, 180.27], [753.76, 180.29]]}, {\"category\": \"car\", \"corners_3d\": [[706.03, 199.15], [731.16, 199.09], [747.97, 201.66], [720.42, 201.74], [706.03, 175.63], [731.16, 175.62], [747.97, 175.89], [720.42, 175.9]]}, {\"category\": \"car\", \"corners_3d\": [[987.23, 211.42], [998.56, 210.41], [1083.73, 213.5], [1073.94, 214.7], [987.23, 179.53], [998.56, 179.35], [1083.73, 179.89], [1073.94, 180.09]]}]\n```" - } - ] - }, - { - "id": 1724, - "data_source": "Mono3DRefer", - "scene_name": "001403", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[652.44, 211.81], [687.28, 211.75], [699.22, 216.45], [660.18, 216.52], [652.44, 179.42], [687.28, 179.41], [699.22, 180.2], [660.18, 180.21]]}, {\"category\": \"car\", \"corners_3d\": [[716.9, 240.72], [783.45, 240.53], [839.22, 259.89], [753.76, 260.2], [716.9, 178.63], [783.45, 178.62], [839.22, 180.27], [753.76, 180.29]]}, {\"category\": \"car\", \"corners_3d\": [[706.03, 199.15], [731.16, 199.09], [747.97, 201.66], [720.42, 201.74], [706.03, 175.63], [731.16, 175.62], [747.97, 175.89], [720.42, 175.9]]}, {\"category\": \"truck\", \"corners_3d\": [[797.31, 225.38], [878.56, 225.15], [1041.61, 253.83], [916.63, 254.39], [797.31, 110.95], [878.56, 111.23], [1041.61, 77.43], [916.63, 76.77]]}, {\"category\": \"car\", \"corners_3d\": [[987.23, 211.42], [998.56, 210.41], [1083.73, 213.5], [1073.94, 214.7], [987.23, 179.53], [998.56, 179.35], [1083.73, 179.89], [1073.94, 180.09]]}]\n```" - } - ] - }, - { - "id": 1725, - "data_source": "Mono3DRefer", - "scene_name": "001406", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[688.72, 297.92], [696.41, 307.06], [596.8, 307.41], [596.3, 298.23], [688.72, 150.37], [696.41, 148.02], [596.8, 147.93], [596.3, 150.29]]}]\n```" - } - ] - }, - { - "id": 1726, - "data_source": "Mono3DRefer", - "scene_name": "001408", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[559.22, 208.72], [528.05, 208.72], [535.67, 205.35], [563.9, 205.34], [559.22, 182.63], [528.05, 182.63], [535.67, 181.71], [563.9, 181.71]]}]\n```" - } - ] - }, - { - "id": 1727, - "data_source": "Mono3DRefer", - "scene_name": "001409", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1317.86, 315.88], [1369.33, 325.22], [1264.32, 328.0], [1218.52, 318.32], [1317.86, 114.53], [1369.33, 110.72], [1264.32, 109.59], [1218.52, 113.54]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1266.2, 310.15], [1315.4, 320.01], [1225.37, 320.77], [1181.99, 310.82], [1266.2, 129.3], [1315.4, 126.17], [1225.37, 125.93], [1181.99, 129.09]]}]\n```" - } - ] - }, - { - "id": 1728, - "data_source": "Mono3DRefer", - "scene_name": "001411", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[25.59, 527.58], [-285.48, 523.0], [195.76, 339.07], [345.38, 340.09], [25.59, 252.51], [-285.48, 251.48], [195.76, 210.18], [345.38, 210.41]]}, {\"category\": \"car\", \"corners_3d\": [[492.78, 241.09], [435.57, 240.43], [479.88, 228.93], [527.64, 229.39], [492.78, 183.31], [435.57, 183.21], [479.88, 181.45], [527.64, 181.52]]}, {\"category\": \"car\", \"corners_3d\": [[647.87, 229.55], [704.42, 229.69], [715.4, 238.28], [650.29, 238.1], [647.87, 172.84], [704.42, 172.84], [715.4, 172.84], [650.29, 172.84]]}]\n```" - } - ] - }, - { - "id": 1729, - "data_source": "Mono3DRefer", - "scene_name": "001412", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[524.3, 196.87], [518.55, 196.74], [529.8, 196.26], [535.51, 196.39], [524.3, 177.54], [518.55, 177.51], [529.8, 177.42], [535.51, 177.44]]}]\n```" - } - ] - }, - { - "id": 1730, - "data_source": "Mono3DRefer", - "scene_name": "001412", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-554.75, 998.97], [-1296.97, 1016.08], [56.73, 412.61], [262.96, 411.2], [-554.75, 313.43], [-1296.97, 316.34], [56.73, 213.65], [262.96, 213.41]]}, {\"category\": \"car\", \"corners_3d\": [[743.4, 285.31], [848.45, 284.49], [1063.12, 371.82], [877.62, 374.42], [743.4, 188.8], [848.45, 188.68], [1063.12, 201.07], [877.62, 201.43]]}, {\"category\": \"car\", \"corners_3d\": [[318.09, 359.51], [156.36, 357.4], [339.9, 288.1], [441.76, 288.92], [318.09, 215.11], [156.36, 214.63], [339.9, 198.94], [441.76, 199.13]]}, {\"category\": \"car\", \"corners_3d\": [[444.0, 274.74], [358.02, 275.11], [406.9, 253.31], [474.47, 253.08], [444.0, 192.99], [358.02, 193.06], [406.9, 188.76], [474.47, 188.71]]}, {\"category\": \"car\", \"corners_3d\": [[494.99, 249.26], [428.53, 249.26], [460.92, 235.55], [515.45, 235.54], [494.99, 191.53], [428.53, 191.53], [460.92, 188.18], [515.45, 188.18]]}, {\"category\": \"car\", \"corners_3d\": [[708.25, 243.62], [776.99, 243.49], [843.99, 269.6], [749.95, 269.84], [708.25, 180.61], [776.99, 180.59], [843.99, 183.45], [749.95, 183.48]]}, {\"category\": \"car\", \"corners_3d\": [[691.6, 228.12], [744.1, 228.09], [771.29, 238.73], [708.7, 238.78], [691.6, 178.24], [744.1, 178.23], [771.29, 179.27], [708.7, 179.27]]}, {\"category\": \"car\", \"corners_3d\": [[907.62, 215.56], [896.22, 212.86], [1005.16, 213.79], [1024.37, 216.61], [907.62, 159.43], [896.22, 160.28], [1005.16, 159.98], [1024.37, 159.1]]}, {\"category\": \"car\", \"corners_3d\": [[641.79, 204.21], [666.87, 204.02], [682.02, 206.64], [654.92, 206.86], [641.79, 178.75], [666.87, 178.71], [682.02, 179.21], [654.92, 179.25]]}, {\"category\": \"car\", \"corners_3d\": [[596.61, 200.02], [615.07, 199.8], [628.73, 201.42], [609.27, 201.66], [596.61, 182.53], [615.07, 182.46], [628.73, 183.03], [609.27, 183.12]]}, {\"category\": \"car\", \"corners_3d\": [[570.79, 195.15], [561.94, 194.82], [591.32, 194.07], [600.29, 194.38], [570.79, 178.89], [561.94, 178.8], [591.32, 178.6], [600.29, 178.69]]}]\n```" - } - ] - }, - { - "id": 1731, - "data_source": "Mono3DRefer", - "scene_name": "001412", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-554.75, 998.97], [-1296.97, 1016.08], [56.73, 412.61], [262.96, 411.2], [-554.75, 313.43], [-1296.97, 316.34], [56.73, 213.65], [262.96, 213.41]]}, {\"category\": \"car\", \"corners_3d\": [[743.4, 285.31], [848.45, 284.49], [1063.12, 371.82], [877.62, 374.42], [743.4, 188.8], [848.45, 188.68], [1063.12, 201.07], [877.62, 201.43]]}, {\"category\": \"car\", \"corners_3d\": [[318.09, 359.51], [156.36, 357.4], [339.9, 288.1], [441.76, 288.92], [318.09, 215.11], [156.36, 214.63], [339.9, 198.94], [441.76, 199.13]]}, {\"category\": \"car\", \"corners_3d\": [[444.0, 274.74], [358.02, 275.11], [406.9, 253.31], [474.47, 253.08], [444.0, 192.99], [358.02, 193.06], [406.9, 188.76], [474.47, 188.71]]}, {\"category\": \"car\", \"corners_3d\": [[494.99, 249.26], [428.53, 249.26], [460.92, 235.55], [515.45, 235.54], [494.99, 191.53], [428.53, 191.53], [460.92, 188.18], [515.45, 188.18]]}, {\"category\": \"car\", \"corners_3d\": [[708.25, 243.62], [776.99, 243.49], [843.99, 269.6], [749.95, 269.84], [708.25, 180.61], [776.99, 180.59], [843.99, 183.45], [749.95, 183.48]]}, {\"category\": \"car\", \"corners_3d\": [[691.6, 228.12], [744.1, 228.09], [771.29, 238.73], [708.7, 238.78], [691.6, 178.24], [744.1, 178.23], [771.29, 179.27], [708.7, 179.27]]}, {\"category\": \"car\", \"corners_3d\": [[907.62, 215.56], [896.22, 212.86], [1005.16, 213.79], [1024.37, 216.61], [907.62, 159.43], [896.22, 160.28], [1005.16, 159.98], [1024.37, 159.1]]}, {\"category\": \"car\", \"corners_3d\": [[641.79, 204.21], [666.87, 204.02], [682.02, 206.64], [654.92, 206.86], [641.79, 178.75], [666.87, 178.71], [682.02, 179.21], [654.92, 179.25]]}, {\"category\": \"car\", \"corners_3d\": [[596.61, 200.02], [615.07, 199.8], [628.73, 201.42], [609.27, 201.66], [596.61, 182.53], [615.07, 182.46], [628.73, 183.03], [609.27, 183.12]]}, {\"category\": \"car\", \"corners_3d\": [[570.79, 195.15], [561.94, 194.82], [591.32, 194.07], [600.29, 194.38], [570.79, 178.89], [561.94, 178.8], [591.32, 178.6], [600.29, 178.69]]}, {\"category\": \"cyclist\", \"corners_3d\": [[524.3, 196.87], [518.55, 196.74], [529.8, 196.26], [535.51, 196.39], [524.3, 177.54], [518.55, 177.51], [529.8, 177.42], [535.51, 177.44]]}]\n```" - } - ] - }, - { - "id": 1732, - "data_source": "Mono3DRefer", - "scene_name": "001418", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[352.5, 333.56], [188.3, 339.17], [282.74, 287.54], [394.95, 284.85], [352.5, 193.02], [188.3, 193.73], [282.74, 187.25], [394.95, 186.91]]}, {\"category\": \"car\", \"corners_3d\": [[590.67, 204.02], [586.29, 202.73], [665.84, 202.16], [673.49, 203.4], [590.67, 172.67], [586.29, 172.67], [665.84, 172.68], [673.49, 172.67]]}, {\"category\": \"car\", \"corners_3d\": [[712.22, 182.18], [728.23, 182.15], [739.48, 182.62], [722.69, 182.66], [712.22, 164.13], [728.23, 164.15], [739.48, 163.71], [722.69, 163.68]]}]\n```" - } - ] - }, - { - "id": 1733, - "data_source": "Mono3DRefer", - "scene_name": "001422", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[599.22, 177.11], [616.3, 177.11], [617.07, 176.89], [598.87, 176.89], [599.22, 157.78], [616.3, 157.78], [617.07, 156.29], [598.87, 156.29]]}]\n```" - } - ] - }, - { - "id": 1734, - "data_source": "Mono3DRefer", - "scene_name": "001422", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[942.01, 206.39], [967.0, 206.81], [965.71, 207.36], [940.23, 206.92], [942.01, 146.53], [967.0, 145.97], [965.71, 145.25], [940.23, 145.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[513.77, 327.93], [468.56, 323.48], [517.13, 308.07], [558.68, 311.6], [513.77, 147.86], [468.56, 148.85], [517.13, 152.26], [558.68, 151.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[407.94, 311.53], [369.4, 308.0], [431.97, 293.27], [467.48, 296.01], [407.94, 140.19], [369.4, 141.28], [431.97, 145.81], [467.48, 144.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[463.59, 322.73], [421.17, 316.78], [483.16, 303.06], [523.56, 307.85], [463.59, 131.9], [421.17, 133.94], [483.16, 138.62], [523.56, 136.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[813.55, 230.02], [806.95, 228.05], [853.71, 228.24], [862.29, 230.23], [813.55, 150.58], [806.95, 151.77], [853.71, 151.66], [862.29, 150.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1081.49, 234.94], [1060.05, 232.41], [1090.27, 232.46], [1113.21, 234.99], [1081.49, 140.18], [1060.05, 142.05], [1090.27, 142.02], [1113.21, 140.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1119.39, 242.4], [1089.77, 239.46], [1124.96, 238.96], [1156.05, 241.85], [1119.39, 131.86], [1089.77, 134.17], [1124.96, 134.56], [1156.05, 132.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1099.86, 237.77], [1085.1, 236.04], [1117.46, 236.07], [1133.24, 237.81], [1099.86, 147.87], [1085.1, 148.85], [1117.46, 148.84], [1133.24, 147.85]]}]\n```" - } - ] - }, - { - "id": 1735, - "data_source": "Mono3DRefer", - "scene_name": "001422", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[933.79, 235.6], [948.33, 236.84], [883.13, 239.8], [869.35, 238.43], [933.79, 145.3], [948.33, 144.51], [883.13, 142.62], [869.35, 143.49]]}]\n```" - } - ] - }, - { - "id": 1736, - "data_source": "Mono3DRefer", - "scene_name": "001422", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[942.01, 206.39], [967.0, 206.81], [965.71, 207.36], [940.23, 206.92], [942.01, 146.53], [967.0, 145.97], [965.71, 145.25], [940.23, 145.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[513.77, 327.93], [468.56, 323.48], [517.13, 308.07], [558.68, 311.6], [513.77, 147.86], [468.56, 148.85], [517.13, 152.26], [558.68, 151.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[407.94, 311.53], [369.4, 308.0], [431.97, 293.27], [467.48, 296.01], [407.94, 140.19], [369.4, 141.28], [431.97, 145.81], [467.48, 144.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[463.59, 322.73], [421.17, 316.78], [483.16, 303.06], [523.56, 307.85], [463.59, 131.9], [421.17, 133.94], [483.16, 138.62], [523.56, 136.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[813.55, 230.02], [806.95, 228.05], [853.71, 228.24], [862.29, 230.23], [813.55, 150.58], [806.95, 151.77], [853.71, 151.66], [862.29, 150.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1081.49, 234.94], [1060.05, 232.41], [1090.27, 232.46], [1113.21, 234.99], [1081.49, 140.18], [1060.05, 142.05], [1090.27, 142.02], [1113.21, 140.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1119.39, 242.4], [1089.77, 239.46], [1124.96, 238.96], [1156.05, 241.85], [1119.39, 131.86], [1089.77, 134.17], [1124.96, 134.56], [1156.05, 132.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1099.86, 237.77], [1085.1, 236.04], [1117.46, 236.07], [1133.24, 237.81], [1099.86, 147.87], [1085.1, 148.85], [1117.46, 148.84], [1133.24, 147.85]]}, {\"category\": \"van\", \"corners_3d\": [[599.22, 177.11], [616.3, 177.11], [617.07, 176.89], [598.87, 176.89], [599.22, 157.78], [616.3, 157.78], [617.07, 156.29], [598.87, 156.29]]}]\n```" - } - ] - }, - { - "id": 1737, - "data_source": "Mono3DRefer", - "scene_name": "001422", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[933.79, 235.6], [948.33, 236.84], [883.13, 239.8], [869.35, 238.43], [933.79, 145.3], [948.33, 144.51], [883.13, 142.62], [869.35, 143.49]]}, {\"category\": \"van\", \"corners_3d\": [[599.22, 177.11], [616.3, 177.11], [617.07, 176.89], [598.87, 176.89], [599.22, 157.78], [616.3, 157.78], [617.07, 156.29], [598.87, 156.29]]}]\n```" - } - ] - }, - { - "id": 1738, - "data_source": "Mono3DRefer", - "scene_name": "001422", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[933.79, 235.6], [948.33, 236.84], [883.13, 239.8], [869.35, 238.43], [933.79, 145.3], [948.33, 144.51], [883.13, 142.62], [869.35, 143.49]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[942.01, 206.39], [967.0, 206.81], [965.71, 207.36], [940.23, 206.92], [942.01, 146.53], [967.0, 145.97], [965.71, 145.25], [940.23, 145.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[513.77, 327.93], [468.56, 323.48], [517.13, 308.07], [558.68, 311.6], [513.77, 147.86], [468.56, 148.85], [517.13, 152.26], [558.68, 151.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[407.94, 311.53], [369.4, 308.0], [431.97, 293.27], [467.48, 296.01], [407.94, 140.19], [369.4, 141.28], [431.97, 145.81], [467.48, 144.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[463.59, 322.73], [421.17, 316.78], [483.16, 303.06], [523.56, 307.85], [463.59, 131.9], [421.17, 133.94], [483.16, 138.62], [523.56, 136.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[813.55, 230.02], [806.95, 228.05], [853.71, 228.24], [862.29, 230.23], [813.55, 150.58], [806.95, 151.77], [853.71, 151.66], [862.29, 150.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1081.49, 234.94], [1060.05, 232.41], [1090.27, 232.46], [1113.21, 234.99], [1081.49, 140.18], [1060.05, 142.05], [1090.27, 142.02], [1113.21, 140.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1119.39, 242.4], [1089.77, 239.46], [1124.96, 238.96], [1156.05, 241.85], [1119.39, 131.86], [1089.77, 134.17], [1124.96, 134.56], [1156.05, 132.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1099.86, 237.77], [1085.1, 236.04], [1117.46, 236.07], [1133.24, 237.81], [1099.86, 147.87], [1085.1, 148.85], [1117.46, 148.84], [1133.24, 147.85]]}]\n```" - } - ] - }, - { - "id": 1739, - "data_source": "Mono3DRefer", - "scene_name": "001422", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[933.79, 235.6], [948.33, 236.84], [883.13, 239.8], [869.35, 238.43], [933.79, 145.3], [948.33, 144.51], [883.13, 142.62], [869.35, 143.49]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[942.01, 206.39], [967.0, 206.81], [965.71, 207.36], [940.23, 206.92], [942.01, 146.53], [967.0, 145.97], [965.71, 145.25], [940.23, 145.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[513.77, 327.93], [468.56, 323.48], [517.13, 308.07], [558.68, 311.6], [513.77, 147.86], [468.56, 148.85], [517.13, 152.26], [558.68, 151.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[407.94, 311.53], [369.4, 308.0], [431.97, 293.27], [467.48, 296.01], [407.94, 140.19], [369.4, 141.28], [431.97, 145.81], [467.48, 144.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[463.59, 322.73], [421.17, 316.78], [483.16, 303.06], [523.56, 307.85], [463.59, 131.9], [421.17, 133.94], [483.16, 138.62], [523.56, 136.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[813.55, 230.02], [806.95, 228.05], [853.71, 228.24], [862.29, 230.23], [813.55, 150.58], [806.95, 151.77], [853.71, 151.66], [862.29, 150.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1081.49, 234.94], [1060.05, 232.41], [1090.27, 232.46], [1113.21, 234.99], [1081.49, 140.18], [1060.05, 142.05], [1090.27, 142.02], [1113.21, 140.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1119.39, 242.4], [1089.77, 239.46], [1124.96, 238.96], [1156.05, 241.85], [1119.39, 131.86], [1089.77, 134.17], [1124.96, 134.56], [1156.05, 132.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1099.86, 237.77], [1085.1, 236.04], [1117.46, 236.07], [1133.24, 237.81], [1099.86, 147.87], [1085.1, 148.85], [1117.46, 148.84], [1133.24, 147.85]]}, {\"category\": \"van\", \"corners_3d\": [[599.22, 177.11], [616.3, 177.11], [617.07, 176.89], [598.87, 176.89], [599.22, 157.78], [616.3, 157.78], [617.07, 156.29], [598.87, 156.29]]}]\n```" - } - ] - }, - { - "id": 1740, - "data_source": "Mono3DRefer", - "scene_name": "001423", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[332.98, 307.8], [319.23, 320.09], [208.4, 323.8], [232.49, 310.87], [332.98, 160.68], [319.23, 158.77], [208.4, 158.19], [232.49, 160.2]]}]\n```" - } - ] - }, - { - "id": 1741, - "data_source": "Mono3DRefer", - "scene_name": "001424", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[84.87, 227.44], [35.57, 227.48], [126.41, 218.72], [167.75, 218.69], [84.87, 177.0], [35.57, 177.01], [126.41, 176.34], [167.75, 176.34]]}]\n```" - } - ] - }, - { - "id": 1742, - "data_source": "Mono3DRefer", - "scene_name": "001424", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[709.97, 203.08], [698.26, 203.09], [696.84, 202.67], [708.39, 202.66], [709.97, 175.82], [698.26, 175.83], [696.84, 175.78], [708.39, 175.78]]}]\n```" - } - ] - }, - { - "id": 1743, - "data_source": "Mono3DRefer", - "scene_name": "001424", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[455.82, 197.7], [434.74, 197.68], [446.26, 196.22], [466.11, 196.24], [455.82, 179.27], [434.74, 179.26], [446.26, 178.88], [466.11, 178.89]]}, {\"category\": \"car\", \"corners_3d\": [[214.82, 232.98], [165.59, 233.02], [220.09, 225.51], [263.13, 225.47], [214.82, 188.03], [165.59, 188.04], [220.09, 186.14], [263.13, 186.13]]}, {\"category\": \"car\", \"corners_3d\": [[160.61, 231.32], [110.02, 231.32], [182.2, 222.86], [225.47, 222.86], [160.61, 190.64], [110.02, 190.64], [182.2, 188.07], [225.47, 188.07]]}, {\"category\": \"car\", \"corners_3d\": [[369.36, 209.3], [336.71, 209.28], [361.75, 206.02], [391.49, 206.03], [369.36, 179.84], [336.71, 179.84], [361.75, 179.21], [391.49, 179.22]]}, {\"category\": \"car\", \"corners_3d\": [[332.72, 208.91], [302.15, 208.88], [329.44, 205.82], [357.44, 205.84], [332.72, 180.16], [302.15, 180.16], [329.44, 179.54], [357.44, 179.54]]}]\n```" - } - ] - }, - { - "id": 1744, - "data_source": "Mono3DRefer", - "scene_name": "001424", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[709.97, 203.08], [698.26, 203.09], [696.84, 202.67], [708.39, 202.66], [709.97, 175.82], [698.26, 175.83], [696.84, 175.78], [708.39, 175.78]]}, {\"category\": \"van\", \"corners_3d\": [[84.87, 227.44], [35.57, 227.48], [126.41, 218.72], [167.75, 218.69], [84.87, 177.0], [35.57, 177.01], [126.41, 176.34], [167.75, 176.34]]}]\n```" - } - ] - }, - { - "id": 1745, - "data_source": "Mono3DRefer", - "scene_name": "001424", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[455.82, 197.7], [434.74, 197.68], [446.26, 196.22], [466.11, 196.24], [455.82, 179.27], [434.74, 179.26], [446.26, 178.88], [466.11, 178.89]]}, {\"category\": \"van\", \"corners_3d\": [[84.87, 227.44], [35.57, 227.48], [126.41, 218.72], [167.75, 218.69], [84.87, 177.0], [35.57, 177.01], [126.41, 176.34], [167.75, 176.34]]}, {\"category\": \"car\", \"corners_3d\": [[214.82, 232.98], [165.59, 233.02], [220.09, 225.51], [263.13, 225.47], [214.82, 188.03], [165.59, 188.04], [220.09, 186.14], [263.13, 186.13]]}, {\"category\": \"car\", \"corners_3d\": [[160.61, 231.32], [110.02, 231.32], [182.2, 222.86], [225.47, 222.86], [160.61, 190.64], [110.02, 190.64], [182.2, 188.07], [225.47, 188.07]]}, {\"category\": \"car\", \"corners_3d\": [[369.36, 209.3], [336.71, 209.28], [361.75, 206.02], [391.49, 206.03], [369.36, 179.84], [336.71, 179.84], [361.75, 179.21], [391.49, 179.22]]}, {\"category\": \"car\", \"corners_3d\": [[332.72, 208.91], [302.15, 208.88], [329.44, 205.82], [357.44, 205.84], [332.72, 180.16], [302.15, 180.16], [329.44, 179.54], [357.44, 179.54]]}]\n```" - } - ] - }, - { - "id": 1746, - "data_source": "Mono3DRefer", - "scene_name": "001424", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[455.82, 197.7], [434.74, 197.68], [446.26, 196.22], [466.11, 196.24], [455.82, 179.27], [434.74, 179.26], [446.26, 178.88], [466.11, 178.89]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[709.97, 203.08], [698.26, 203.09], [696.84, 202.67], [708.39, 202.66], [709.97, 175.82], [698.26, 175.83], [696.84, 175.78], [708.39, 175.78]]}, {\"category\": \"car\", \"corners_3d\": [[214.82, 232.98], [165.59, 233.02], [220.09, 225.51], [263.13, 225.47], [214.82, 188.03], [165.59, 188.04], [220.09, 186.14], [263.13, 186.13]]}, {\"category\": \"car\", \"corners_3d\": [[160.61, 231.32], [110.02, 231.32], [182.2, 222.86], [225.47, 222.86], [160.61, 190.64], [110.02, 190.64], [182.2, 188.07], [225.47, 188.07]]}, {\"category\": \"car\", \"corners_3d\": [[369.36, 209.3], [336.71, 209.28], [361.75, 206.02], [391.49, 206.03], [369.36, 179.84], [336.71, 179.84], [361.75, 179.21], [391.49, 179.22]]}, {\"category\": \"car\", \"corners_3d\": [[332.72, 208.91], [302.15, 208.88], [329.44, 205.82], [357.44, 205.84], [332.72, 180.16], [302.15, 180.16], [329.44, 179.54], [357.44, 179.54]]}]\n```" - } - ] - }, - { - "id": 1747, - "data_source": "Mono3DRefer", - "scene_name": "001424", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[455.82, 197.7], [434.74, 197.68], [446.26, 196.22], [466.11, 196.24], [455.82, 179.27], [434.74, 179.26], [446.26, 178.88], [466.11, 178.89]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[709.97, 203.08], [698.26, 203.09], [696.84, 202.67], [708.39, 202.66], [709.97, 175.82], [698.26, 175.83], [696.84, 175.78], [708.39, 175.78]]}, {\"category\": \"van\", \"corners_3d\": [[84.87, 227.44], [35.57, 227.48], [126.41, 218.72], [167.75, 218.69], [84.87, 177.0], [35.57, 177.01], [126.41, 176.34], [167.75, 176.34]]}, {\"category\": \"car\", \"corners_3d\": [[214.82, 232.98], [165.59, 233.02], [220.09, 225.51], [263.13, 225.47], [214.82, 188.03], [165.59, 188.04], [220.09, 186.14], [263.13, 186.13]]}, {\"category\": \"car\", \"corners_3d\": [[160.61, 231.32], [110.02, 231.32], [182.2, 222.86], [225.47, 222.86], [160.61, 190.64], [110.02, 190.64], [182.2, 188.07], [225.47, 188.07]]}, {\"category\": \"car\", \"corners_3d\": [[369.36, 209.3], [336.71, 209.28], [361.75, 206.02], [391.49, 206.03], [369.36, 179.84], [336.71, 179.84], [361.75, 179.21], [391.49, 179.22]]}, {\"category\": \"car\", \"corners_3d\": [[332.72, 208.91], [302.15, 208.88], [329.44, 205.82], [357.44, 205.84], [332.72, 180.16], [302.15, 180.16], [329.44, 179.54], [357.44, 179.54]]}]\n```" - } - ] - }, - { - "id": 1748, - "data_source": "Mono3DRefer", - "scene_name": "001425", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[98.88, 266.18], [154.98, 256.09], [340.38, 256.34], [306.63, 266.49], [98.88, 172.84], [154.98, 172.84], [340.38, 172.84], [306.63, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[483.98, 209.31], [478.52, 211.1], [398.11, 211.18], [407.35, 209.38], [483.98, 178.0], [478.52, 178.25], [398.11, 178.26], [407.35, 178.01]]}]\n```" - } - ] - }, - { - "id": 1749, - "data_source": "Mono3DRefer", - "scene_name": "001425", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[919.43, 221.49], [937.54, 221.49], [970.34, 226.37], [950.41, 226.37], [919.43, 156.86], [937.54, 156.86], [970.34, 155.26], [950.41, 155.26]]}, {\"category\": \"cyclist\", \"corners_3d\": [[854.01, 212.98], [867.02, 213.0], [884.89, 216.05], [870.87, 216.03], [854.01, 164.46], [867.02, 164.45], [884.89, 163.81], [870.87, 163.82]]}, {\"category\": \"cyclist\", \"corners_3d\": [[827.45, 204.99], [838.25, 204.99], [853.03, 207.07], [841.53, 207.07], [827.45, 165.24], [838.25, 165.24], [853.03, 164.75], [841.53, 164.75]]}]\n```" - } - ] - }, - { - "id": 1750, - "data_source": "Mono3DRefer", - "scene_name": "001425", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[98.88, 266.18], [154.98, 256.09], [340.38, 256.34], [306.63, 266.49], [98.88, 172.84], [154.98, 172.84], [340.38, 172.84], [306.63, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[483.98, 209.31], [478.52, 211.1], [398.11, 211.18], [407.35, 209.38], [483.98, 178.0], [478.52, 178.25], [398.11, 178.26], [407.35, 178.01]]}, {\"category\": \"cyclist\", \"corners_3d\": [[919.43, 221.49], [937.54, 221.49], [970.34, 226.37], [950.41, 226.37], [919.43, 156.86], [937.54, 156.86], [970.34, 155.26], [950.41, 155.26]]}, {\"category\": \"cyclist\", \"corners_3d\": [[854.01, 212.98], [867.02, 213.0], [884.89, 216.05], [870.87, 216.03], [854.01, 164.46], [867.02, 164.45], [884.89, 163.81], [870.87, 163.82]]}, {\"category\": \"cyclist\", \"corners_3d\": [[827.45, 204.99], [838.25, 204.99], [853.03, 207.07], [841.53, 207.07], [827.45, 165.24], [838.25, 165.24], [853.03, 164.75], [841.53, 164.75]]}]\n```" - } - ] - }, - { - "id": 1751, - "data_source": "Mono3DRefer", - "scene_name": "001427", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[334.39, 200.34], [302.29, 200.38], [328.48, 197.85], [357.61, 197.82], [334.39, 175.78], [302.29, 175.78], [328.48, 175.51], [357.61, 175.51]]}]\n```" - } - ] - }, - { - "id": 1752, - "data_source": "Mono3DRefer", - "scene_name": "001430", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[628.85, 216.28], [662.42, 216.73], [646.66, 223.07], [608.49, 222.48], [628.85, 186.92], [662.42, 187.06], [646.66, 189.12], [608.49, 188.92]]}]\n```" - } - ] - }, - { - "id": 1753, - "data_source": "Mono3DRefer", - "scene_name": "001430", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[682.9, 218.58], [721.9, 219.18], [711.11, 226.89], [665.87, 226.08], [682.9, 172.62], [721.9, 172.62], [711.11, 172.58], [665.87, 172.58]]}]\n```" - } - ] - }, - { - "id": 1754, - "data_source": "Mono3DRefer", - "scene_name": "001430", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[626.28, 202.28], [603.1, 202.66], [559.02, 198.01], [579.1, 197.73], [626.28, 168.79], [603.1, 168.73], [559.02, 169.38], [579.1, 169.41]]}]\n```" - } - ] - }, - { - "id": 1755, - "data_source": "Mono3DRefer", - "scene_name": "001430", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[628.85, 216.28], [662.42, 216.73], [646.66, 223.07], [608.49, 222.48], [628.85, 186.92], [662.42, 187.06], [646.66, 189.12], [608.49, 188.92]]}, {\"category\": \"van\", \"corners_3d\": [[682.9, 218.58], [721.9, 219.18], [711.11, 226.89], [665.87, 226.08], [682.9, 172.62], [721.9, 172.62], [711.11, 172.58], [665.87, 172.58]]}]\n```" - } - ] - }, - { - "id": 1756, - "data_source": "Mono3DRefer", - "scene_name": "001430", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[628.85, 216.28], [662.42, 216.73], [646.66, 223.07], [608.49, 222.48], [628.85, 186.92], [662.42, 187.06], [646.66, 189.12], [608.49, 188.92]]}, {\"category\": \"tram\", \"corners_3d\": [[626.28, 202.28], [603.1, 202.66], [559.02, 198.01], [579.1, 197.73], [626.28, 168.79], [603.1, 168.73], [559.02, 169.38], [579.1, 169.41]]}]\n```" - } - ] - }, - { - "id": 1757, - "data_source": "Mono3DRefer", - "scene_name": "001430", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[682.9, 218.58], [721.9, 219.18], [711.11, 226.89], [665.87, 226.08], [682.9, 172.62], [721.9, 172.62], [711.11, 172.58], [665.87, 172.58]]}, {\"category\": \"tram\", \"corners_3d\": [[626.28, 202.28], [603.1, 202.66], [559.02, 198.01], [579.1, 197.73], [626.28, 168.79], [603.1, 168.73], [559.02, 169.38], [579.1, 169.41]]}]\n```" - } - ] - }, - { - "id": 1758, - "data_source": "Mono3DRefer", - "scene_name": "001430", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[628.85, 216.28], [662.42, 216.73], [646.66, 223.07], [608.49, 222.48], [628.85, 186.92], [662.42, 187.06], [646.66, 189.12], [608.49, 188.92]]}, {\"category\": \"van\", \"corners_3d\": [[682.9, 218.58], [721.9, 219.18], [711.11, 226.89], [665.87, 226.08], [682.9, 172.62], [721.9, 172.62], [711.11, 172.58], [665.87, 172.58]]}, {\"category\": \"tram\", \"corners_3d\": [[626.28, 202.28], [603.1, 202.66], [559.02, 198.01], [579.1, 197.73], [626.28, 168.79], [603.1, 168.73], [559.02, 169.38], [579.1, 169.41]]}]\n```" - } - ] - }, - { - "id": 1759, - "data_source": "Mono3DRefer", - "scene_name": "001431", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1127.87, 303.54], [1185.67, 316.26], [1081.11, 317.3], [1032.76, 304.4], [1127.87, 147.62], [1185.67, 144.22], [1081.11, 143.94], [1032.76, 147.39]]}]\n```" - } - ] - }, - { - "id": 1760, - "data_source": "Mono3DRefer", - "scene_name": "001434", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[762.7, 273.25], [838.93, 284.34], [621.68, 317.08], [553.01, 299.05], [762.7, 188.99], [838.93, 190.77], [621.68, 196.04], [553.01, 193.14]]}, {\"category\": \"car\", \"corners_3d\": [[519.9, 252.4], [559.47, 258.33], [389.21, 274.69], [356.58, 266.38], [519.9, 193.13], [559.47, 194.64], [389.21, 198.81], [356.58, 196.69]]}, {\"category\": \"car\", \"corners_3d\": [[284.89, 260.61], [266.2, 254.6], [421.48, 244.15], [447.49, 248.68], [284.89, 184.73], [266.2, 183.92], [421.48, 182.51], [447.49, 183.12]]}, {\"category\": \"car\", \"corners_3d\": [[833.97, 289.72], [949.59, 304.79], [812.75, 346.35], [685.83, 321.19], [833.97, 191.25], [949.59, 193.62], [812.75, 200.16], [685.83, 196.2]]}]\n```" - } - ] - }, - { - "id": 1761, - "data_source": "Mono3DRefer", - "scene_name": "001438", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[526.91, 196.95], [530.09, 196.01], [611.16, 196.01], [611.27, 196.95], [526.91, 153.81], [530.09, 154.55], [611.16, 154.55], [611.27, 153.81]]}]\n```" - } - ] - }, - { - "id": 1762, - "data_source": "Mono3DRefer", - "scene_name": "001438", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[33.15, 377.13], [-161.35, 378.25], [121.71, 301.34], [242.65, 300.9], [33.15, 203.17], [-161.35, 203.33], [121.71, 191.92], [242.65, 191.86]]}, {\"category\": \"car\", \"corners_3d\": [[857.19, 354.16], [1033.08, 351.94], [1586.79, 567.96], [1208.2, 578.9], [857.19, 186.33], [1033.08, 186.17], [1586.79, 202.22], [1208.2, 203.03]]}, {\"category\": \"car\", \"corners_3d\": [[516.86, 212.41], [482.49, 213.25], [462.92, 209.54], [494.57, 208.85], [516.86, 180.89], [482.49, 181.06], [462.92, 180.31], [494.57, 180.17]]}, {\"category\": \"car\", \"corners_3d\": [[550.89, 197.39], [539.44, 196.68], [591.53, 195.66], [603.96, 196.31], [550.89, 175.66], [539.44, 175.58], [591.53, 175.46], [603.96, 175.53]]}, {\"category\": \"car\", \"corners_3d\": [[418.7, 207.31], [397.22, 208.68], [340.19, 207.08], [362.84, 205.83], [418.7, 181.82], [397.22, 182.18], [340.19, 181.76], [362.84, 181.44]]}, {\"category\": \"car\", \"corners_3d\": [[653.97, 193.72], [644.43, 193.05], [701.37, 192.49], [712.49, 193.13], [653.97, 173.4], [644.43, 173.38], [701.37, 173.37], [712.49, 173.38]]}]\n```" - } - ] - }, - { - "id": 1763, - "data_source": "Mono3DRefer", - "scene_name": "001438", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[697.7, 209.22], [684.49, 209.63], [673.18, 208.78], [686.21, 208.39], [697.7, 169.56], [684.49, 169.52], [673.18, 169.6], [686.21, 169.64]]}]\n```" - } - ] - }, - { - "id": 1764, - "data_source": "Mono3DRefer", - "scene_name": "001438", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[33.15, 377.13], [-161.35, 378.25], [121.71, 301.34], [242.65, 300.9], [33.15, 203.17], [-161.35, 203.33], [121.71, 191.92], [242.65, 191.86]]}, {\"category\": \"car\", \"corners_3d\": [[857.19, 354.16], [1033.08, 351.94], [1586.79, 567.96], [1208.2, 578.9], [857.19, 186.33], [1033.08, 186.17], [1586.79, 202.22], [1208.2, 203.03]]}, {\"category\": \"car\", \"corners_3d\": [[516.86, 212.41], [482.49, 213.25], [462.92, 209.54], [494.57, 208.85], [516.86, 180.89], [482.49, 181.06], [462.92, 180.31], [494.57, 180.17]]}, {\"category\": \"car\", \"corners_3d\": [[550.89, 197.39], [539.44, 196.68], [591.53, 195.66], [603.96, 196.31], [550.89, 175.66], [539.44, 175.58], [591.53, 175.46], [603.96, 175.53]]}, {\"category\": \"car\", \"corners_3d\": [[418.7, 207.31], [397.22, 208.68], [340.19, 207.08], [362.84, 205.83], [418.7, 181.82], [397.22, 182.18], [340.19, 181.76], [362.84, 181.44]]}, {\"category\": \"car\", \"corners_3d\": [[653.97, 193.72], [644.43, 193.05], [701.37, 192.49], [712.49, 193.13], [653.97, 173.4], [644.43, 173.38], [701.37, 173.37], [712.49, 173.38]]}, {\"category\": \"truck\", \"corners_3d\": [[526.91, 196.95], [530.09, 196.01], [611.16, 196.01], [611.27, 196.95], [526.91, 153.81], [530.09, 154.55], [611.16, 154.55], [611.27, 153.81]]}]\n```" - } - ] - }, - { - "id": 1765, - "data_source": "Mono3DRefer", - "scene_name": "001438", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[697.7, 209.22], [684.49, 209.63], [673.18, 208.78], [686.21, 208.39], [697.7, 169.56], [684.49, 169.52], [673.18, 169.6], [686.21, 169.64]]}, {\"category\": \"truck\", \"corners_3d\": [[526.91, 196.95], [530.09, 196.01], [611.16, 196.01], [611.27, 196.95], [526.91, 153.81], [530.09, 154.55], [611.16, 154.55], [611.27, 153.81]]}]\n```" - } - ] - }, - { - "id": 1766, - "data_source": "Mono3DRefer", - "scene_name": "001438", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[33.15, 377.13], [-161.35, 378.25], [121.71, 301.34], [242.65, 300.9], [33.15, 203.17], [-161.35, 203.33], [121.71, 191.92], [242.65, 191.86]]}, {\"category\": \"car\", \"corners_3d\": [[857.19, 354.16], [1033.08, 351.94], [1586.79, 567.96], [1208.2, 578.9], [857.19, 186.33], [1033.08, 186.17], [1586.79, 202.22], [1208.2, 203.03]]}, {\"category\": \"car\", \"corners_3d\": [[516.86, 212.41], [482.49, 213.25], [462.92, 209.54], [494.57, 208.85], [516.86, 180.89], [482.49, 181.06], [462.92, 180.31], [494.57, 180.17]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[697.7, 209.22], [684.49, 209.63], [673.18, 208.78], [686.21, 208.39], [697.7, 169.56], [684.49, 169.52], [673.18, 169.6], [686.21, 169.64]]}, {\"category\": \"car\", \"corners_3d\": [[550.89, 197.39], [539.44, 196.68], [591.53, 195.66], [603.96, 196.31], [550.89, 175.66], [539.44, 175.58], [591.53, 175.46], [603.96, 175.53]]}, {\"category\": \"car\", \"corners_3d\": [[418.7, 207.31], [397.22, 208.68], [340.19, 207.08], [362.84, 205.83], [418.7, 181.82], [397.22, 182.18], [340.19, 181.76], [362.84, 181.44]]}, {\"category\": \"car\", \"corners_3d\": [[653.97, 193.72], [644.43, 193.05], [701.37, 192.49], [712.49, 193.13], [653.97, 173.4], [644.43, 173.38], [701.37, 173.37], [712.49, 173.38]]}]\n```" - } - ] - }, - { - "id": 1767, - "data_source": "Mono3DRefer", - "scene_name": "001438", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[33.15, 377.13], [-161.35, 378.25], [121.71, 301.34], [242.65, 300.9], [33.15, 203.17], [-161.35, 203.33], [121.71, 191.92], [242.65, 191.86]]}, {\"category\": \"car\", \"corners_3d\": [[857.19, 354.16], [1033.08, 351.94], [1586.79, 567.96], [1208.2, 578.9], [857.19, 186.33], [1033.08, 186.17], [1586.79, 202.22], [1208.2, 203.03]]}, {\"category\": \"car\", \"corners_3d\": [[516.86, 212.41], [482.49, 213.25], [462.92, 209.54], [494.57, 208.85], [516.86, 180.89], [482.49, 181.06], [462.92, 180.31], [494.57, 180.17]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[697.7, 209.22], [684.49, 209.63], [673.18, 208.78], [686.21, 208.39], [697.7, 169.56], [684.49, 169.52], [673.18, 169.6], [686.21, 169.64]]}, {\"category\": \"car\", \"corners_3d\": [[550.89, 197.39], [539.44, 196.68], [591.53, 195.66], [603.96, 196.31], [550.89, 175.66], [539.44, 175.58], [591.53, 175.46], [603.96, 175.53]]}, {\"category\": \"car\", \"corners_3d\": [[418.7, 207.31], [397.22, 208.68], [340.19, 207.08], [362.84, 205.83], [418.7, 181.82], [397.22, 182.18], [340.19, 181.76], [362.84, 181.44]]}, {\"category\": \"car\", \"corners_3d\": [[653.97, 193.72], [644.43, 193.05], [701.37, 192.49], [712.49, 193.13], [653.97, 173.4], [644.43, 173.38], [701.37, 173.37], [712.49, 173.38]]}, {\"category\": \"truck\", \"corners_3d\": [[526.91, 196.95], [530.09, 196.01], [611.16, 196.01], [611.27, 196.95], [526.91, 153.81], [530.09, 154.55], [611.16, 154.55], [611.27, 153.81]]}]\n```" - } - ] - }, - { - "id": 1768, - "data_source": "Mono3DRefer", - "scene_name": "001453", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[647.17, 195.54], [690.83, 195.74], [682.62, 202.37], [626.55, 202.03], [647.17, 142.65], [690.83, 142.37], [682.62, 133.55], [626.55, 134.0]]}]\n```" - } - ] - }, - { - "id": 1769, - "data_source": "Mono3DRefer", - "scene_name": "001453", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[271.6, 209.34], [241.23, 209.29], [276.94, 205.95], [304.57, 205.99], [271.6, 179.62], [241.23, 179.61], [276.94, 178.99], [304.57, 179.0]]}, {\"category\": \"car\", \"corners_3d\": [[315.31, 204.03], [285.99, 203.98], [316.49, 201.28], [343.3, 201.32], [315.31, 180.44], [285.99, 180.42], [316.49, 179.77], [343.3, 179.78]]}, {\"category\": \"car\", \"corners_3d\": [[356.63, 198.7], [334.64, 198.58], [364.03, 196.6], [384.46, 196.71], [356.63, 177.93], [334.64, 177.91], [364.03, 177.52], [384.46, 177.54]]}]\n```" - } - ] - }, - { - "id": 1770, - "data_source": "Mono3DRefer", - "scene_name": "001453", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[647.17, 195.54], [690.83, 195.74], [682.62, 202.37], [626.55, 202.03], [647.17, 142.65], [690.83, 142.37], [682.62, 133.55], [626.55, 134.0]]}, {\"category\": \"car\", \"corners_3d\": [[271.6, 209.34], [241.23, 209.29], [276.94, 205.95], [304.57, 205.99], [271.6, 179.62], [241.23, 179.61], [276.94, 178.99], [304.57, 179.0]]}, {\"category\": \"car\", \"corners_3d\": [[315.31, 204.03], [285.99, 203.98], [316.49, 201.28], [343.3, 201.32], [315.31, 180.44], [285.99, 180.42], [316.49, 179.77], [343.3, 179.78]]}, {\"category\": \"car\", \"corners_3d\": [[356.63, 198.7], [334.64, 198.58], [364.03, 196.6], [384.46, 196.71], [356.63, 177.93], [334.64, 177.91], [364.03, 177.52], [384.46, 177.54]]}]\n```" - } - ] - }, - { - "id": 1771, - "data_source": "Mono3DRefer", - "scene_name": "001454", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[764.74, 217.3], [743.53, 217.31], [738.25, 215.66], [758.68, 215.64], [764.74, 166.77], [743.53, 166.77], [738.25, 166.99], [758.68, 167.0]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1151.18, 271.83], [1199.43, 280.31], [1162.05, 280.49], [1116.68, 271.98], [1151.18, 155.2], [1199.43, 153.69], [1162.05, 153.66], [1116.68, 155.17]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1125.45, 266.22], [1157.75, 272.07], [1119.42, 272.07], [1089.38, 266.22], [1125.45, 162.53], [1157.75, 161.88], [1119.42, 161.88], [1089.38, 162.53]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[549.23, 226.11], [569.65, 226.05], [569.21, 228.08], [548.01, 228.14], [549.23, 178.02], [569.65, 178.01], [569.21, 178.21], [548.01, 178.21]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[538.46, 238.6], [510.93, 238.71], [513.42, 236.33], [539.96, 236.23], [538.46, 172.06], [510.93, 172.06], [513.42, 172.08], [539.96, 172.09]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[743.86, 218.32], [726.04, 218.28], [722.25, 216.47], [739.35, 216.5], [743.86, 165.08], [726.04, 165.09], [722.25, 165.4], [739.35, 165.39]]}]\n```" - } - ] - }, - { - "id": 1772, - "data_source": "Mono3DRefer", - "scene_name": "001456", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[254.18, 382.72], [49.03, 379.71], [297.74, 293.4], [418.66, 294.42], [254.18, 185.01], [49.03, 184.84], [297.74, 179.84], [418.66, 179.9]]}, {\"category\": \"car\", \"corners_3d\": [[179.26, 290.49], [51.09, 291.72], [209.27, 255.18], [297.19, 254.59], [179.26, 186.3], [51.09, 186.44], [209.27, 182.26], [297.19, 182.19]]}, {\"category\": \"car\", \"corners_3d\": [[424.82, 223.47], [467.84, 223.75], [439.1, 230.8], [390.34, 230.44], [424.82, 185.02], [467.84, 185.08], [439.1, 186.78], [390.34, 186.69]]}, {\"category\": \"car\", \"corners_3d\": [[380.52, 223.38], [418.67, 223.84], [362.66, 233.11], [318.24, 232.47], [380.52, 185.35], [418.67, 185.46], [362.66, 187.75], [318.24, 187.6]]}]\n```" - } - ] - }, - { - "id": 1773, - "data_source": "Mono3DRefer", - "scene_name": "001461", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[111.68, 238.8], [50.26, 238.8], [143.22, 227.83], [194.42, 227.83], [111.68, 190.15], [50.26, 190.15], [143.22, 187.27], [194.42, 187.27]]}, {\"category\": \"car\", \"corners_3d\": [[403.38, 205.8], [377.37, 205.8], [393.71, 203.48], [417.89, 203.48], [403.38, 181.89], [377.37, 181.89], [393.71, 181.26], [417.89, 181.26]]}, {\"category\": \"car\", \"corners_3d\": [[362.04, 208.46], [334.04, 208.42], [359.67, 205.35], [385.28, 205.38], [362.04, 185.67], [334.04, 185.65], [359.67, 184.55], [385.28, 184.56]]}, {\"category\": \"car\", \"corners_3d\": [[457.94, 200.85], [436.76, 200.82], [448.89, 199.13], [468.8, 199.16], [457.94, 181.62], [436.76, 181.61], [448.89, 181.08], [468.8, 181.09]]}]\n```" - } - ] - }, - { - "id": 1774, - "data_source": "Mono3DRefer", - "scene_name": "001461", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[236.0, 211.22], [200.43, 211.24], [249.16, 206.58], [280.4, 206.56], [236.0, 174.73], [200.43, 174.73], [249.16, 174.5], [280.4, 174.5]]}]\n```" - } - ] - }, - { - "id": 1775, - "data_source": "Mono3DRefer", - "scene_name": "001461", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[111.68, 238.8], [50.26, 238.8], [143.22, 227.83], [194.42, 227.83], [111.68, 190.15], [50.26, 190.15], [143.22, 187.27], [194.42, 187.27]]}, {\"category\": \"van\", \"corners_3d\": [[236.0, 211.22], [200.43, 211.24], [249.16, 206.58], [280.4, 206.56], [236.0, 174.73], [200.43, 174.73], [249.16, 174.5], [280.4, 174.5]]}, {\"category\": \"car\", \"corners_3d\": [[403.38, 205.8], [377.37, 205.8], [393.71, 203.48], [417.89, 203.48], [403.38, 181.89], [377.37, 181.89], [393.71, 181.26], [417.89, 181.26]]}, {\"category\": \"car\", \"corners_3d\": [[362.04, 208.46], [334.04, 208.42], [359.67, 205.35], [385.28, 205.38], [362.04, 185.67], [334.04, 185.65], [359.67, 184.55], [385.28, 184.56]]}, {\"category\": \"car\", \"corners_3d\": [[457.94, 200.85], [436.76, 200.82], [448.89, 199.13], [468.8, 199.16], [457.94, 181.62], [436.76, 181.61], [448.89, 181.08], [468.8, 181.09]]}]\n```" - } - ] - }, - { - "id": 1776, - "data_source": "Mono3DRefer", - "scene_name": "001462", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[571.13, 212.46], [603.95, 212.49], [601.27, 217.64], [564.2, 217.59], [571.13, 182.17], [603.95, 182.18], [601.27, 183.39], [564.2, 183.37]]}, {\"category\": \"car\", \"corners_3d\": [[1362.72, 323.42], [1551.31, 367.51], [958.39, 347.32], [908.54, 311.05], [1362.72, 175.75], [1551.31, 176.6], [958.39, 176.21], [908.54, 175.51]]}, {\"category\": \"car\", \"corners_3d\": [[522.5, 205.92], [492.58, 205.85], [504.97, 203.17], [532.49, 203.23], [522.5, 178.39], [492.58, 178.38], [504.97, 177.93], [532.49, 177.94]]}]\n```" - } - ] - }, - { - "id": 1777, - "data_source": "Mono3DRefer", - "scene_name": "001471", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[735.2, 196.18], [758.27, 195.97], [821.74, 200.29], [794.99, 200.58], [735.2, 168.82], [758.27, 168.85], [821.74, 168.11], [794.99, 168.06]]}]\n```" - } - ] - }, - { - "id": 1778, - "data_source": "Mono3DRefer", - "scene_name": "001471", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[529.0, 191.77], [554.93, 191.75], [546.04, 196.75], [513.21, 196.79], [529.0, 155.73], [554.93, 155.74], [546.04, 151.21], [513.21, 151.18]]}]\n```" - } - ] - }, - { - "id": 1779, - "data_source": "Mono3DRefer", - "scene_name": "001471", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[577.19, 210.37], [610.64, 210.37], [610.69, 213.71], [574.25, 213.71], [577.19, 178.26], [610.64, 178.26], [610.69, 178.74], [574.25, 178.74]]}]\n```" - } - ] - }, - { - "id": 1780, - "data_source": "Mono3DRefer", - "scene_name": "001471", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[529.0, 191.77], [554.93, 191.75], [546.04, 196.75], [513.21, 196.79], [529.0, 155.73], [554.93, 155.74], [546.04, 151.21], [513.21, 151.18]]}, {\"category\": \"bus\", \"corners_3d\": [[735.2, 196.18], [758.27, 195.97], [821.74, 200.29], [794.99, 200.58], [735.2, 168.82], [758.27, 168.85], [821.74, 168.11], [794.99, 168.06]]}]\n```" - } - ] - }, - { - "id": 1781, - "data_source": "Mono3DRefer", - "scene_name": "001471", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[577.19, 210.37], [610.64, 210.37], [610.69, 213.71], [574.25, 213.71], [577.19, 178.26], [610.64, 178.26], [610.69, 178.74], [574.25, 178.74]]}, {\"category\": \"bus\", \"corners_3d\": [[735.2, 196.18], [758.27, 195.97], [821.74, 200.29], [794.99, 200.58], [735.2, 168.82], [758.27, 168.85], [821.74, 168.11], [794.99, 168.06]]}]\n```" - } - ] - }, - { - "id": 1782, - "data_source": "Mono3DRefer", - "scene_name": "001471", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[577.19, 210.37], [610.64, 210.37], [610.69, 213.71], [574.25, 213.71], [577.19, 178.26], [610.64, 178.26], [610.69, 178.74], [574.25, 178.74]]}, {\"category\": \"tram\", \"corners_3d\": [[529.0, 191.77], [554.93, 191.75], [546.04, 196.75], [513.21, 196.79], [529.0, 155.73], [554.93, 155.74], [546.04, 151.21], [513.21, 151.18]]}]\n```" - } - ] - }, - { - "id": 1783, - "data_source": "Mono3DRefer", - "scene_name": "001471", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[577.19, 210.37], [610.64, 210.37], [610.69, 213.71], [574.25, 213.71], [577.19, 178.26], [610.64, 178.26], [610.69, 178.74], [574.25, 178.74]]}, {\"category\": \"tram\", \"corners_3d\": [[529.0, 191.77], [554.93, 191.75], [546.04, 196.75], [513.21, 196.79], [529.0, 155.73], [554.93, 155.74], [546.04, 151.21], [513.21, 151.18]]}, {\"category\": \"bus\", \"corners_3d\": [[735.2, 196.18], [758.27, 195.97], [821.74, 200.29], [794.99, 200.58], [735.2, 168.82], [758.27, 168.85], [821.74, 168.11], [794.99, 168.06]]}]\n```" - } - ] - }, - { - "id": 1784, - "data_source": "Mono3DRefer", - "scene_name": "001475", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[517.52, 295.61], [407.08, 290.88], [513.85, 257.58], [595.29, 259.99], [517.52, 181.01], [407.08, 180.7], [513.85, 178.49], [595.29, 178.65]]}, {\"category\": \"car\", \"corners_3d\": [[824.25, 315.21], [969.22, 319.37], [1164.35, 433.72], [902.18, 420.82], [824.25, 188.45], [969.22, 188.91], [1164.35, 201.43], [902.18, 200.02]]}, {\"category\": \"car\", \"corners_3d\": [[794.86, 265.84], [893.98, 267.34], [968.12, 302.93], [830.86, 300.1], [794.86, 175.69], [893.98, 175.73], [968.12, 176.82], [830.86, 176.73]]}, {\"category\": \"car\", \"corners_3d\": [[1138.52, 213.67], [1195.77, 212.86], [1356.16, 221.54], [1290.18, 222.75], [1138.52, 146.59], [1195.77, 147.12], [1356.16, 141.53], [1290.18, 140.75]]}, {\"category\": \"car\", \"corners_3d\": [[601.07, 219.72], [554.29, 219.2], [580.14, 212.49], [620.32, 212.87], [601.07, 174.65], [554.29, 174.63], [580.14, 174.37], [620.32, 174.38]]}, {\"category\": \"car\", \"corners_3d\": [[906.3, 202.27], [873.35, 200.93], [924.25, 199.45], [957.68, 200.65], [906.3, 163.71], [873.35, 164.13], [924.25, 164.59], [957.68, 164.22]]}, {\"category\": \"car\", \"corners_3d\": [[742.2, 205.45], [778.83, 205.68], [787.6, 209.98], [746.15, 209.69], [742.2, 171.78], [778.83, 171.77], [787.6, 171.63], [746.15, 171.64]]}, {\"category\": \"car\", \"corners_3d\": [[637.03, 204.65], [606.22, 204.35], [622.4, 201.25], [650.29, 201.49], [637.03, 172.23], [606.22, 172.24], [622.4, 172.3], [650.29, 172.29]]}]\n```" - } - ] - }, - { - "id": 1785, - "data_source": "Mono3DRefer", - "scene_name": "001475", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[714.71, 186.91], [745.76, 187.06], [738.22, 189.67], [701.61, 189.47], [714.71, 149.7], [745.76, 149.46], [738.22, 145.16], [701.61, 145.49]]}]\n```" - } - ] - }, - { - "id": 1786, - "data_source": "Mono3DRefer", - "scene_name": "001475", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1034.44, 214.91], [1089.02, 214.38], [1220.6, 223.68], [1155.65, 224.47], [1034.44, 142.18], [1089.02, 142.56], [1220.6, 135.79], [1155.65, 135.21]]}, {\"category\": \"van\", \"corners_3d\": [[925.28, 215.82], [982.0, 215.51], [1085.91, 225.65], [1016.54, 226.14], [925.28, 152.37], [982.0, 152.52], [1085.91, 147.69], [1016.54, 147.46]]}]\n```" - } - ] - }, - { - "id": 1787, - "data_source": "Mono3DRefer", - "scene_name": "001475", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[517.52, 295.61], [407.08, 290.88], [513.85, 257.58], [595.29, 259.99], [517.52, 181.01], [407.08, 180.7], [513.85, 178.49], [595.29, 178.65]]}, {\"category\": \"car\", \"corners_3d\": [[824.25, 315.21], [969.22, 319.37], [1164.35, 433.72], [902.18, 420.82], [824.25, 188.45], [969.22, 188.91], [1164.35, 201.43], [902.18, 200.02]]}, {\"category\": \"car\", \"corners_3d\": [[794.86, 265.84], [893.98, 267.34], [968.12, 302.93], [830.86, 300.1], [794.86, 175.69], [893.98, 175.73], [968.12, 176.82], [830.86, 176.73]]}, {\"category\": \"car\", \"corners_3d\": [[1138.52, 213.67], [1195.77, 212.86], [1356.16, 221.54], [1290.18, 222.75], [1138.52, 146.59], [1195.77, 147.12], [1356.16, 141.53], [1290.18, 140.75]]}, {\"category\": \"car\", \"corners_3d\": [[601.07, 219.72], [554.29, 219.2], [580.14, 212.49], [620.32, 212.87], [601.07, 174.65], [554.29, 174.63], [580.14, 174.37], [620.32, 174.38]]}, {\"category\": \"car\", \"corners_3d\": [[906.3, 202.27], [873.35, 200.93], [924.25, 199.45], [957.68, 200.65], [906.3, 163.71], [873.35, 164.13], [924.25, 164.59], [957.68, 164.22]]}, {\"category\": \"car\", \"corners_3d\": [[742.2, 205.45], [778.83, 205.68], [787.6, 209.98], [746.15, 209.69], [742.2, 171.78], [778.83, 171.77], [787.6, 171.63], [746.15, 171.64]]}, {\"category\": \"car\", \"corners_3d\": [[637.03, 204.65], [606.22, 204.35], [622.4, 201.25], [650.29, 201.49], [637.03, 172.23], [606.22, 172.24], [622.4, 172.3], [650.29, 172.29]]}, {\"category\": \"truck\", \"corners_3d\": [[714.71, 186.91], [745.76, 187.06], [738.22, 189.67], [701.61, 189.47], [714.71, 149.7], [745.76, 149.46], [738.22, 145.16], [701.61, 145.49]]}]\n```" - } - ] - }, - { - "id": 1788, - "data_source": "Mono3DRefer", - "scene_name": "001475", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[517.52, 295.61], [407.08, 290.88], [513.85, 257.58], [595.29, 259.99], [517.52, 181.01], [407.08, 180.7], [513.85, 178.49], [595.29, 178.65]]}, {\"category\": \"car\", \"corners_3d\": [[824.25, 315.21], [969.22, 319.37], [1164.35, 433.72], [902.18, 420.82], [824.25, 188.45], [969.22, 188.91], [1164.35, 201.43], [902.18, 200.02]]}, {\"category\": \"car\", \"corners_3d\": [[794.86, 265.84], [893.98, 267.34], [968.12, 302.93], [830.86, 300.1], [794.86, 175.69], [893.98, 175.73], [968.12, 176.82], [830.86, 176.73]]}, {\"category\": \"car\", \"corners_3d\": [[1138.52, 213.67], [1195.77, 212.86], [1356.16, 221.54], [1290.18, 222.75], [1138.52, 146.59], [1195.77, 147.12], [1356.16, 141.53], [1290.18, 140.75]]}, {\"category\": \"van\", \"corners_3d\": [[1034.44, 214.91], [1089.02, 214.38], [1220.6, 223.68], [1155.65, 224.47], [1034.44, 142.18], [1089.02, 142.56], [1220.6, 135.79], [1155.65, 135.21]]}, {\"category\": \"van\", \"corners_3d\": [[925.28, 215.82], [982.0, 215.51], [1085.91, 225.65], [1016.54, 226.14], [925.28, 152.37], [982.0, 152.52], [1085.91, 147.69], [1016.54, 147.46]]}, {\"category\": \"car\", \"corners_3d\": [[601.07, 219.72], [554.29, 219.2], [580.14, 212.49], [620.32, 212.87], [601.07, 174.65], [554.29, 174.63], [580.14, 174.37], [620.32, 174.38]]}, {\"category\": \"car\", \"corners_3d\": [[906.3, 202.27], [873.35, 200.93], [924.25, 199.45], [957.68, 200.65], [906.3, 163.71], [873.35, 164.13], [924.25, 164.59], [957.68, 164.22]]}, {\"category\": \"car\", \"corners_3d\": [[742.2, 205.45], [778.83, 205.68], [787.6, 209.98], [746.15, 209.69], [742.2, 171.78], [778.83, 171.77], [787.6, 171.63], [746.15, 171.64]]}, {\"category\": \"car\", \"corners_3d\": [[637.03, 204.65], [606.22, 204.35], [622.4, 201.25], [650.29, 201.49], [637.03, 172.23], [606.22, 172.24], [622.4, 172.3], [650.29, 172.29]]}]\n```" - } - ] - }, - { - "id": 1789, - "data_source": "Mono3DRefer", - "scene_name": "001475", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1034.44, 214.91], [1089.02, 214.38], [1220.6, 223.68], [1155.65, 224.47], [1034.44, 142.18], [1089.02, 142.56], [1220.6, 135.79], [1155.65, 135.21]]}, {\"category\": \"van\", \"corners_3d\": [[925.28, 215.82], [982.0, 215.51], [1085.91, 225.65], [1016.54, 226.14], [925.28, 152.37], [982.0, 152.52], [1085.91, 147.69], [1016.54, 147.46]]}, {\"category\": \"truck\", \"corners_3d\": [[714.71, 186.91], [745.76, 187.06], [738.22, 189.67], [701.61, 189.47], [714.71, 149.7], [745.76, 149.46], [738.22, 145.16], [701.61, 145.49]]}]\n```" - } - ] - }, - { - "id": 1790, - "data_source": "Mono3DRefer", - "scene_name": "001475", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[517.52, 295.61], [407.08, 290.88], [513.85, 257.58], [595.29, 259.99], [517.52, 181.01], [407.08, 180.7], [513.85, 178.49], [595.29, 178.65]]}, {\"category\": \"car\", \"corners_3d\": [[824.25, 315.21], [969.22, 319.37], [1164.35, 433.72], [902.18, 420.82], [824.25, 188.45], [969.22, 188.91], [1164.35, 201.43], [902.18, 200.02]]}, {\"category\": \"car\", \"corners_3d\": [[794.86, 265.84], [893.98, 267.34], [968.12, 302.93], [830.86, 300.1], [794.86, 175.69], [893.98, 175.73], [968.12, 176.82], [830.86, 176.73]]}, {\"category\": \"car\", \"corners_3d\": [[1138.52, 213.67], [1195.77, 212.86], [1356.16, 221.54], [1290.18, 222.75], [1138.52, 146.59], [1195.77, 147.12], [1356.16, 141.53], [1290.18, 140.75]]}, {\"category\": \"van\", \"corners_3d\": [[1034.44, 214.91], [1089.02, 214.38], [1220.6, 223.68], [1155.65, 224.47], [1034.44, 142.18], [1089.02, 142.56], [1220.6, 135.79], [1155.65, 135.21]]}, {\"category\": \"van\", \"corners_3d\": [[925.28, 215.82], [982.0, 215.51], [1085.91, 225.65], [1016.54, 226.14], [925.28, 152.37], [982.0, 152.52], [1085.91, 147.69], [1016.54, 147.46]]}, {\"category\": \"car\", \"corners_3d\": [[601.07, 219.72], [554.29, 219.2], [580.14, 212.49], [620.32, 212.87], [601.07, 174.65], [554.29, 174.63], [580.14, 174.37], [620.32, 174.38]]}, {\"category\": \"car\", \"corners_3d\": [[906.3, 202.27], [873.35, 200.93], [924.25, 199.45], [957.68, 200.65], [906.3, 163.71], [873.35, 164.13], [924.25, 164.59], [957.68, 164.22]]}, {\"category\": \"car\", \"corners_3d\": [[742.2, 205.45], [778.83, 205.68], [787.6, 209.98], [746.15, 209.69], [742.2, 171.78], [778.83, 171.77], [787.6, 171.63], [746.15, 171.64]]}, {\"category\": \"car\", \"corners_3d\": [[637.03, 204.65], [606.22, 204.35], [622.4, 201.25], [650.29, 201.49], [637.03, 172.23], [606.22, 172.24], [622.4, 172.3], [650.29, 172.29]]}, {\"category\": \"truck\", \"corners_3d\": [[714.71, 186.91], [745.76, 187.06], [738.22, 189.67], [701.61, 189.47], [714.71, 149.7], [745.76, 149.46], [738.22, 145.16], [701.61, 145.49]]}]\n```" - } - ] - }, - { - "id": 1791, - "data_source": "Mono3DRefer", - "scene_name": "001482", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[566.01, 194.24], [563.05, 194.75], [517.38, 194.6], [521.38, 194.09], [566.01, 180.45], [563.05, 180.64], [517.38, 180.58], [521.38, 180.4]]}, {\"category\": \"car\", \"corners_3d\": [[294.75, 230.15], [313.82, 226.21], [436.74, 225.75], [426.83, 229.62], [294.75, 185.1], [313.82, 184.26], [436.74, 184.16], [426.83, 184.99]]}]\n```" - } - ] - }, - { - "id": 1792, - "data_source": "Mono3DRefer", - "scene_name": "001489", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-583.21, 905.52], [-1357.42, 921.33], [64.55, 376.08], [269.77, 374.89], [-583.21, 226.98], [-1357.42, 228.14], [64.55, 187.87], [269.77, 187.78]]}, {\"category\": \"car\", \"corners_3d\": [[648.28, 207.18], [671.27, 207.19], [677.17, 210.49], [651.96, 210.49], [648.28, 186.67], [671.27, 186.67], [677.17, 188.0], [651.96, 188.0]]}]\n```" - } - ] - }, - { - "id": 1793, - "data_source": "Mono3DRefer", - "scene_name": "001491", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[676.04, 195.13], [711.45, 195.08], [735.04, 199.11], [693.24, 199.17], [676.04, 145.7], [711.45, 145.75], [735.04, 140.84], [693.24, 140.77]]}]\n```" - } - ] - }, - { - "id": 1794, - "data_source": "Mono3DRefer", - "scene_name": "001491", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[727.53, 242.94], [794.48, 242.88], [839.76, 259.43], [757.02, 259.52], [727.53, 182.67], [794.48, 182.66], [839.76, 184.98], [757.02, 184.99]]}, {\"category\": \"car\", \"corners_3d\": [[885.81, 240.61], [952.2, 240.42], [1068.79, 262.08], [981.47, 262.4], [885.81, 179.95], [952.2, 179.93], [1068.79, 182.2], [981.47, 182.23]]}, {\"category\": \"car\", \"corners_3d\": [[812.58, 201.48], [828.13, 200.96], [880.73, 203.04], [865.06, 203.64], [812.58, 175.64], [828.13, 175.59], [880.73, 175.79], [865.06, 175.85]]}, {\"category\": \"car\", \"corners_3d\": [[631.16, 195.65], [653.63, 195.63], [658.91, 197.47], [634.61, 197.5], [631.16, 174.8], [653.63, 174.79], [658.91, 174.95], [634.61, 174.95]]}, {\"category\": \"car\", \"corners_3d\": [[673.47, 217.54], [715.66, 217.41], [740.03, 225.08], [690.64, 225.26], [673.47, 172.04], [715.66, 172.04], [740.03, 171.9], [690.64, 171.9]]}]\n```" - } - ] - }, - { - "id": 1795, - "data_source": "Mono3DRefer", - "scene_name": "001491", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[727.53, 242.94], [794.48, 242.88], [839.76, 259.43], [757.02, 259.52], [727.53, 182.67], [794.48, 182.66], [839.76, 184.98], [757.02, 184.99]]}, {\"category\": \"truck\", \"corners_3d\": [[676.04, 195.13], [711.45, 195.08], [735.04, 199.11], [693.24, 199.17], [676.04, 145.7], [711.45, 145.75], [735.04, 140.84], [693.24, 140.77]]}, {\"category\": \"car\", \"corners_3d\": [[885.81, 240.61], [952.2, 240.42], [1068.79, 262.08], [981.47, 262.4], [885.81, 179.95], [952.2, 179.93], [1068.79, 182.2], [981.47, 182.23]]}, {\"category\": \"car\", \"corners_3d\": [[812.58, 201.48], [828.13, 200.96], [880.73, 203.04], [865.06, 203.64], [812.58, 175.64], [828.13, 175.59], [880.73, 175.79], [865.06, 175.85]]}, {\"category\": \"car\", \"corners_3d\": [[631.16, 195.65], [653.63, 195.63], [658.91, 197.47], [634.61, 197.5], [631.16, 174.8], [653.63, 174.79], [658.91, 174.95], [634.61, 174.95]]}, {\"category\": \"car\", \"corners_3d\": [[673.47, 217.54], [715.66, 217.41], [740.03, 225.08], [690.64, 225.26], [673.47, 172.04], [715.66, 172.04], [740.03, 171.9], [690.64, 171.9]]}]\n```" - } - ] - }, - { - "id": 1796, - "data_source": "Mono3DRefer", - "scene_name": "001492", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[357.86, 216.82], [329.95, 216.82], [352.91, 213.2], [378.53, 213.2], [357.86, 193.74], [329.95, 193.74], [352.91, 192.02], [378.53, 192.02]]}]\n```" - } - ] - }, - { - "id": 1797, - "data_source": "Mono3DRefer", - "scene_name": "001492", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[590.63, 199.87], [629.68, 199.87], [634.64, 206.73], [585.68, 206.73], [590.63, 152.31], [629.68, 152.31], [634.64, 147.09], [585.68, 147.1]]}]\n```" - } - ] - }, - { - "id": 1798, - "data_source": "Mono3DRefer", - "scene_name": "001492", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[590.63, 199.87], [629.68, 199.87], [634.64, 206.73], [585.68, 206.73], [590.63, 152.31], [629.68, 152.31], [634.64, 147.09], [585.68, 147.1]]}, {\"category\": \"car\", \"corners_3d\": [[357.86, 216.82], [329.95, 216.82], [352.91, 213.2], [378.53, 213.2], [357.86, 193.74], [329.95, 193.74], [352.91, 192.02], [378.53, 192.02]]}]\n```" - } - ] - }, - { - "id": 1799, - "data_source": "Mono3DRefer", - "scene_name": "001493", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[565.4, 199.82], [613.14, 199.73], [625.96, 208.53], [562.56, 208.69], [565.4, 141.58], [613.14, 141.68], [625.96, 131.47], [562.56, 131.29]]}]\n```" - } - ] - }, - { - "id": 1800, - "data_source": "Mono3DRefer", - "scene_name": "001493", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[10.44, 230.65], [-36.24, 230.86], [30.1, 224.47], [71.43, 224.3], [10.44, 193.17], [-36.24, 193.24], [30.1, 191.0], [71.43, 190.94]]}, {\"category\": \"car\", \"corners_3d\": [[88.41, 222.76], [53.95, 222.79], [114.69, 217.25], [145.3, 217.23], [88.41, 191.73], [53.95, 191.74], [114.69, 189.65], [145.3, 189.64]]}, {\"category\": \"car\", \"corners_3d\": [[164.33, 214.96], [135.51, 215.06], [170.21, 211.66], [196.63, 211.57], [164.33, 185.31], [135.51, 185.34], [170.21, 184.33], [196.63, 184.31]]}, {\"category\": \"car\", \"corners_3d\": [[209.71, 210.89], [183.81, 211.0], [210.17, 208.26], [234.15, 208.17], [209.71, 186.93], [183.81, 186.96], [210.17, 185.95], [234.15, 185.92]]}, {\"category\": \"car\", \"corners_3d\": [[284.29, 201.8], [266.93, 201.79], [290.46, 199.88], [306.69, 199.89], [284.29, 182.59], [266.93, 182.58], [290.46, 181.94], [306.69, 181.94]]}]\n```" - } - ] - }, - { - "id": 1801, - "data_source": "Mono3DRefer", - "scene_name": "001493", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[565.4, 199.82], [613.14, 199.73], [625.96, 208.53], [562.56, 208.69], [565.4, 141.58], [613.14, 141.68], [625.96, 131.47], [562.56, 131.29]]}, {\"category\": \"car\", \"corners_3d\": [[10.44, 230.65], [-36.24, 230.86], [30.1, 224.47], [71.43, 224.3], [10.44, 193.17], [-36.24, 193.24], [30.1, 191.0], [71.43, 190.94]]}, {\"category\": \"car\", \"corners_3d\": [[88.41, 222.76], [53.95, 222.79], [114.69, 217.25], [145.3, 217.23], [88.41, 191.73], [53.95, 191.74], [114.69, 189.65], [145.3, 189.64]]}, {\"category\": \"car\", \"corners_3d\": [[164.33, 214.96], [135.51, 215.06], [170.21, 211.66], [196.63, 211.57], [164.33, 185.31], [135.51, 185.34], [170.21, 184.33], [196.63, 184.31]]}, {\"category\": \"car\", \"corners_3d\": [[209.71, 210.89], [183.81, 211.0], [210.17, 208.26], [234.15, 208.17], [209.71, 186.93], [183.81, 186.96], [210.17, 185.95], [234.15, 185.92]]}, {\"category\": \"car\", \"corners_3d\": [[284.29, 201.8], [266.93, 201.79], [290.46, 199.88], [306.69, 199.89], [284.29, 182.59], [266.93, 182.58], [290.46, 181.94], [306.69, 181.94]]}]\n```" - } - ] - }, - { - "id": 1802, - "data_source": "Mono3DRefer", - "scene_name": "001494", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[230.66, 238.01], [295.34, 237.35], [227.55, 256.54], [142.47, 257.65], [230.66, 184.43], [295.34, 184.31], [227.55, 187.72], [142.47, 187.92]]}, {\"category\": \"car\", \"corners_3d\": [[114.47, 241.68], [177.91, 241.17], [67.9, 261.64], [-15.81, 262.51], [114.47, 167.38], [177.91, 167.42], [67.9, 165.79], [-15.81, 165.72]]}, {\"category\": \"car\", \"corners_3d\": [[402.19, 211.5], [447.75, 211.19], [436.06, 217.78], [382.48, 218.2], [402.19, 172.84], [447.75, 172.84], [436.06, 172.84], [382.48, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[700.83, 220.33], [755.22, 219.85], [815.27, 231.64], [747.82, 232.38], [700.83, 170.88], [755.22, 170.9], [815.27, 170.41], [747.82, 170.38]]}, {\"category\": \"car\", \"corners_3d\": [[548.17, 195.14], [581.87, 195.05], [586.7, 197.96], [548.58, 198.08], [548.17, 163.29], [581.87, 163.33], [586.7, 162.08], [548.58, 162.03]]}, {\"category\": \"car\", \"corners_3d\": [[772.87, 217.38], [825.63, 217.05], [893.84, 227.52], [829.1, 228.03], [772.87, 165.58], [825.63, 165.63], [893.84, 163.92], [829.1, 163.84]]}, {\"category\": \"car\", \"corners_3d\": [[854.55, 214.24], [908.4, 213.92], [996.88, 223.72], [930.9, 224.21], [854.55, 170.96], [908.4, 170.98], [996.88, 170.53], [930.9, 170.5]]}]\n```" - } - ] - }, - { - "id": 1803, - "data_source": "Mono3DRefer", - "scene_name": "001494", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[583.4, 195.4], [542.97, 195.54], [540.71, 191.79], [574.52, 191.7], [583.4, 131.51], [542.97, 131.25], [540.71, 138.12], [574.52, 138.3]]}]\n```" - } - ] - }, - { - "id": 1804, - "data_source": "Mono3DRefer", - "scene_name": "001494", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[484.62, 229.69], [548.0, 229.2], [550.86, 244.89], [469.66, 245.69], [484.62, 160.21], [548.0, 160.32], [550.86, 156.83], [469.66, 156.65]]}, {\"category\": \"van\", \"corners_3d\": [[1217.44, 234.56], [1326.83, 244.27], [995.58, 248.77], [927.77, 237.89], [1217.44, 111.11], [1326.83, 101.39], [995.58, 96.9], [927.77, 107.79]]}, {\"category\": \"van\", \"corners_3d\": [[637.02, 191.08], [604.02, 191.18], [594.18, 188.98], [623.29, 188.91], [637.02, 146.31], [604.02, 146.17], [594.18, 149.37], [623.29, 149.48]]}]\n```" - } - ] - }, - { - "id": 1805, - "data_source": "Mono3DRefer", - "scene_name": "001494", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[230.66, 238.01], [295.34, 237.35], [227.55, 256.54], [142.47, 257.65], [230.66, 184.43], [295.34, 184.31], [227.55, 187.72], [142.47, 187.92]]}, {\"category\": \"car\", \"corners_3d\": [[114.47, 241.68], [177.91, 241.17], [67.9, 261.64], [-15.81, 262.51], [114.47, 167.38], [177.91, 167.42], [67.9, 165.79], [-15.81, 165.72]]}, {\"category\": \"car\", \"corners_3d\": [[402.19, 211.5], [447.75, 211.19], [436.06, 217.78], [382.48, 218.2], [402.19, 172.84], [447.75, 172.84], [436.06, 172.84], [382.48, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[700.83, 220.33], [755.22, 219.85], [815.27, 231.64], [747.82, 232.38], [700.83, 170.88], [755.22, 170.9], [815.27, 170.41], [747.82, 170.38]]}, {\"category\": \"car\", \"corners_3d\": [[548.17, 195.14], [581.87, 195.05], [586.7, 197.96], [548.58, 198.08], [548.17, 163.29], [581.87, 163.33], [586.7, 162.08], [548.58, 162.03]]}, {\"category\": \"car\", \"corners_3d\": [[772.87, 217.38], [825.63, 217.05], [893.84, 227.52], [829.1, 228.03], [772.87, 165.58], [825.63, 165.63], [893.84, 163.92], [829.1, 163.84]]}, {\"category\": \"car\", \"corners_3d\": [[854.55, 214.24], [908.4, 213.92], [996.88, 223.72], [930.9, 224.21], [854.55, 170.96], [908.4, 170.98], [996.88, 170.53], [930.9, 170.5]]}, {\"category\": \"truck\", \"corners_3d\": [[583.4, 195.4], [542.97, 195.54], [540.71, 191.79], [574.52, 191.7], [583.4, 131.51], [542.97, 131.25], [540.71, 138.12], [574.52, 138.3]]}]\n```" - } - ] - }, - { - "id": 1806, - "data_source": "Mono3DRefer", - "scene_name": "001494", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[484.62, 229.69], [548.0, 229.2], [550.86, 244.89], [469.66, 245.69], [484.62, 160.21], [548.0, 160.32], [550.86, 156.83], [469.66, 156.65]]}, {\"category\": \"van\", \"corners_3d\": [[1217.44, 234.56], [1326.83, 244.27], [995.58, 248.77], [927.77, 237.89], [1217.44, 111.11], [1326.83, 101.39], [995.58, 96.9], [927.77, 107.79]]}, {\"category\": \"car\", \"corners_3d\": [[230.66, 238.01], [295.34, 237.35], [227.55, 256.54], [142.47, 257.65], [230.66, 184.43], [295.34, 184.31], [227.55, 187.72], [142.47, 187.92]]}, {\"category\": \"car\", \"corners_3d\": [[114.47, 241.68], [177.91, 241.17], [67.9, 261.64], [-15.81, 262.51], [114.47, 167.38], [177.91, 167.42], [67.9, 165.79], [-15.81, 165.72]]}, {\"category\": \"car\", \"corners_3d\": [[402.19, 211.5], [447.75, 211.19], [436.06, 217.78], [382.48, 218.2], [402.19, 172.84], [447.75, 172.84], [436.06, 172.84], [382.48, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[700.83, 220.33], [755.22, 219.85], [815.27, 231.64], [747.82, 232.38], [700.83, 170.88], [755.22, 170.9], [815.27, 170.41], [747.82, 170.38]]}, {\"category\": \"car\", \"corners_3d\": [[548.17, 195.14], [581.87, 195.05], [586.7, 197.96], [548.58, 198.08], [548.17, 163.29], [581.87, 163.33], [586.7, 162.08], [548.58, 162.03]]}, {\"category\": \"van\", \"corners_3d\": [[637.02, 191.08], [604.02, 191.18], [594.18, 188.98], [623.29, 188.91], [637.02, 146.31], [604.02, 146.17], [594.18, 149.37], [623.29, 149.48]]}, {\"category\": \"car\", \"corners_3d\": [[772.87, 217.38], [825.63, 217.05], [893.84, 227.52], [829.1, 228.03], [772.87, 165.58], [825.63, 165.63], [893.84, 163.92], [829.1, 163.84]]}, {\"category\": \"car\", \"corners_3d\": [[854.55, 214.24], [908.4, 213.92], [996.88, 223.72], [930.9, 224.21], [854.55, 170.96], [908.4, 170.98], [996.88, 170.53], [930.9, 170.5]]}]\n```" - } - ] - }, - { - "id": 1807, - "data_source": "Mono3DRefer", - "scene_name": "001494", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[484.62, 229.69], [548.0, 229.2], [550.86, 244.89], [469.66, 245.69], [484.62, 160.21], [548.0, 160.32], [550.86, 156.83], [469.66, 156.65]]}, {\"category\": \"van\", \"corners_3d\": [[1217.44, 234.56], [1326.83, 244.27], [995.58, 248.77], [927.77, 237.89], [1217.44, 111.11], [1326.83, 101.39], [995.58, 96.9], [927.77, 107.79]]}, {\"category\": \"van\", \"corners_3d\": [[637.02, 191.08], [604.02, 191.18], [594.18, 188.98], [623.29, 188.91], [637.02, 146.31], [604.02, 146.17], [594.18, 149.37], [623.29, 149.48]]}, {\"category\": \"truck\", \"corners_3d\": [[583.4, 195.4], [542.97, 195.54], [540.71, 191.79], [574.52, 191.7], [583.4, 131.51], [542.97, 131.25], [540.71, 138.12], [574.52, 138.3]]}]\n```" - } - ] - }, - { - "id": 1808, - "data_source": "Mono3DRefer", - "scene_name": "001494", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[484.62, 229.69], [548.0, 229.2], [550.86, 244.89], [469.66, 245.69], [484.62, 160.21], [548.0, 160.32], [550.86, 156.83], [469.66, 156.65]]}, {\"category\": \"van\", \"corners_3d\": [[1217.44, 234.56], [1326.83, 244.27], [995.58, 248.77], [927.77, 237.89], [1217.44, 111.11], [1326.83, 101.39], [995.58, 96.9], [927.77, 107.79]]}, {\"category\": \"car\", \"corners_3d\": [[230.66, 238.01], [295.34, 237.35], [227.55, 256.54], [142.47, 257.65], [230.66, 184.43], [295.34, 184.31], [227.55, 187.72], [142.47, 187.92]]}, {\"category\": \"car\", \"corners_3d\": [[114.47, 241.68], [177.91, 241.17], [67.9, 261.64], [-15.81, 262.51], [114.47, 167.38], [177.91, 167.42], [67.9, 165.79], [-15.81, 165.72]]}, {\"category\": \"car\", \"corners_3d\": [[402.19, 211.5], [447.75, 211.19], [436.06, 217.78], [382.48, 218.2], [402.19, 172.84], [447.75, 172.84], [436.06, 172.84], [382.48, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[700.83, 220.33], [755.22, 219.85], [815.27, 231.64], [747.82, 232.38], [700.83, 170.88], [755.22, 170.9], [815.27, 170.41], [747.82, 170.38]]}, {\"category\": \"car\", \"corners_3d\": [[548.17, 195.14], [581.87, 195.05], [586.7, 197.96], [548.58, 198.08], [548.17, 163.29], [581.87, 163.33], [586.7, 162.08], [548.58, 162.03]]}, {\"category\": \"van\", \"corners_3d\": [[637.02, 191.08], [604.02, 191.18], [594.18, 188.98], [623.29, 188.91], [637.02, 146.31], [604.02, 146.17], [594.18, 149.37], [623.29, 149.48]]}, {\"category\": \"car\", \"corners_3d\": [[772.87, 217.38], [825.63, 217.05], [893.84, 227.52], [829.1, 228.03], [772.87, 165.58], [825.63, 165.63], [893.84, 163.92], [829.1, 163.84]]}, {\"category\": \"car\", \"corners_3d\": [[854.55, 214.24], [908.4, 213.92], [996.88, 223.72], [930.9, 224.21], [854.55, 170.96], [908.4, 170.98], [996.88, 170.53], [930.9, 170.5]]}, {\"category\": \"truck\", \"corners_3d\": [[583.4, 195.4], [542.97, 195.54], [540.71, 191.79], [574.52, 191.7], [583.4, 131.51], [542.97, 131.25], [540.71, 138.12], [574.52, 138.3]]}]\n```" - } - ] - }, - { - "id": 1809, - "data_source": "Mono3DRefer", - "scene_name": "001498", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[648.29, 205.29], [678.81, 205.23], [691.24, 209.07], [657.12, 209.15], [648.29, 172.46], [678.81, 172.46], [691.24, 172.41], [657.12, 172.41]]}, {\"category\": \"car\", \"corners_3d\": [[677.02, 216.77], [718.69, 216.72], [735.31, 222.65], [688.02, 222.71], [677.02, 179.23], [718.69, 179.23], [735.31, 180.09], [688.02, 180.1]]}, {\"category\": \"car\", \"corners_3d\": [[782.31, 216.08], [825.01, 215.98], [870.0, 223.94], [819.53, 224.08], [782.31, 177.14], [825.01, 177.13], [870.0, 177.92], [819.53, 177.93]]}, {\"category\": \"car\", \"corners_3d\": [[801.8, 200.83], [817.47, 200.33], [867.99, 202.34], [852.12, 202.91], [801.8, 175.42], [817.47, 175.38], [867.99, 175.56], [852.12, 175.61]]}]\n```" - } - ] - }, - { - "id": 1810, - "data_source": "Mono3DRefer", - "scene_name": "001498", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[656.66, 191.05], [685.37, 191.01], [702.21, 193.59], [669.46, 193.64], [656.66, 150.92], [685.37, 150.97], [702.21, 147.86], [669.46, 147.79]]}]\n```" - } - ] - }, - { - "id": 1811, - "data_source": "Mono3DRefer", - "scene_name": "001498", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[648.29, 205.29], [678.81, 205.23], [691.24, 209.07], [657.12, 209.15], [648.29, 172.46], [678.81, 172.46], [691.24, 172.41], [657.12, 172.41]]}, {\"category\": \"car\", \"corners_3d\": [[677.02, 216.77], [718.69, 216.72], [735.31, 222.65], [688.02, 222.71], [677.02, 179.23], [718.69, 179.23], [735.31, 180.09], [688.02, 180.1]]}, {\"category\": \"truck\", \"corners_3d\": [[656.66, 191.05], [685.37, 191.01], [702.21, 193.59], [669.46, 193.64], [656.66, 150.92], [685.37, 150.97], [702.21, 147.86], [669.46, 147.79]]}, {\"category\": \"car\", \"corners_3d\": [[782.31, 216.08], [825.01, 215.98], [870.0, 223.94], [819.53, 224.08], [782.31, 177.14], [825.01, 177.13], [870.0, 177.92], [819.53, 177.93]]}, {\"category\": \"car\", \"corners_3d\": [[801.8, 200.83], [817.47, 200.33], [867.99, 202.34], [852.12, 202.91], [801.8, 175.42], [817.47, 175.38], [867.99, 175.56], [852.12, 175.61]]}]\n```" - } - ] - }, - { - "id": 1812, - "data_source": "Mono3DRefer", - "scene_name": "001503", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[760.71, 199.12], [776.82, 198.72], [819.04, 200.54], [802.49, 201.0], [760.71, 175.41], [776.82, 175.37], [819.04, 175.55], [802.49, 175.59]]}]\n```" - } - ] - }, - { - "id": 1813, - "data_source": "Mono3DRefer", - "scene_name": "001503", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[667.26, 192.82], [689.03, 192.79], [698.26, 194.39], [674.76, 194.43], [667.26, 167.89], [689.03, 167.89], [698.26, 167.5], [674.76, 167.49]]}]\n```" - } - ] - }, - { - "id": 1814, - "data_source": "Mono3DRefer", - "scene_name": "001503", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[667.26, 192.82], [689.03, 192.79], [698.26, 194.39], [674.76, 194.43], [667.26, 167.89], [689.03, 167.89], [698.26, 167.5], [674.76, 167.49]]}, {\"category\": \"car\", \"corners_3d\": [[760.71, 199.12], [776.82, 198.72], [819.04, 200.54], [802.49, 201.0], [760.71, 175.41], [776.82, 175.37], [819.04, 175.55], [802.49, 175.59]]}]\n```" - } - ] - }, - { - "id": 1815, - "data_source": "Mono3DRefer", - "scene_name": "001504", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[541.27, 218.14], [556.04, 217.6], [565.46, 218.66], [550.46, 219.23], [541.27, 176.39], [556.04, 176.35], [565.46, 176.43], [550.46, 176.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[525.83, 217.7], [539.39, 217.23], [549.65, 218.53], [535.8, 219.02], [525.83, 173.39], [539.39, 173.39], [549.65, 173.4], [535.8, 173.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[944.58, 360.22], [930.06, 377.37], [863.69, 367.5], [882.89, 351.9], [944.58, 156.52], [930.06, 155.02], [863.69, 155.88], [882.89, 157.24]]}]\n```" - } - ] - }, - { - "id": 1816, - "data_source": "Mono3DRefer", - "scene_name": "001504", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[426.52, 271.78], [390.37, 271.88], [420.76, 257.14], [451.5, 257.07], [426.52, 168.42], [390.37, 168.42], [420.76, 169.08], [451.5, 169.08]]}, {\"category\": \"cyclist\", \"corners_3d\": [[497.43, 243.95], [470.83, 244.11], [480.43, 236.9], [504.32, 236.77], [497.43, 164.94], [470.83, 164.92], [480.43, 165.72], [504.32, 165.74]]}]\n```" - } - ] - }, - { - "id": 1817, - "data_source": "Mono3DRefer", - "scene_name": "001504", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[541.27, 218.14], [556.04, 217.6], [565.46, 218.66], [550.46, 219.23], [541.27, 176.39], [556.04, 176.35], [565.46, 176.43], [550.46, 176.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[525.83, 217.7], [539.39, 217.23], [549.65, 218.53], [535.8, 219.02], [525.83, 173.39], [539.39, 173.39], [549.65, 173.4], [535.8, 173.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[944.58, 360.22], [930.06, 377.37], [863.69, 367.5], [882.89, 351.9], [944.58, 156.52], [930.06, 155.02], [863.69, 155.88], [882.89, 157.24]]}, {\"category\": \"cyclist\", \"corners_3d\": [[426.52, 271.78], [390.37, 271.88], [420.76, 257.14], [451.5, 257.07], [426.52, 168.42], [390.37, 168.42], [420.76, 169.08], [451.5, 169.08]]}, {\"category\": \"cyclist\", \"corners_3d\": [[497.43, 243.95], [470.83, 244.11], [480.43, 236.9], [504.32, 236.77], [497.43, 164.94], [470.83, 164.92], [480.43, 165.72], [504.32, 165.74]]}]\n```" - } - ] - }, - { - "id": 1818, - "data_source": "Mono3DRefer", - "scene_name": "001506", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[583.66, 210.31], [619.37, 210.3], [621.43, 214.99], [581.24, 215.01], [583.66, 176.57], [619.37, 176.57], [621.43, 177.04], [581.24, 177.04]]}, {\"category\": \"car\", \"corners_3d\": [[599.48, 203.44], [628.68, 203.44], [630.82, 206.97], [598.25, 206.96], [599.48, 172.85], [628.68, 172.85], [630.82, 172.85], [598.25, 172.85]]}, {\"category\": \"car\", \"corners_3d\": [[855.36, 338.51], [1017.05, 340.73], [1376.48, 506.68], [1049.78, 498.04], [855.36, 182.92], [1017.05, 183.06], [1376.48, 193.14], [1049.78, 192.62]]}, {\"category\": \"car\", \"corners_3d\": [[683.0, 229.43], [738.57, 229.43], [769.95, 243.26], [700.8, 243.25], [683.0, 169.76], [738.57, 169.75], [769.95, 169.0], [700.8, 169.0]]}, {\"category\": \"car\", \"corners_3d\": [[16.49, 329.58], [-124.96, 329.6], [108.96, 279.64], [205.31, 279.63], [16.49, 200.23], [-124.96, 200.24], [108.96, 191.51], [205.31, 191.51]]}, {\"category\": \"car\", \"corners_3d\": [[667.81, 214.47], [704.54, 214.43], [718.53, 219.78], [677.08, 219.83], [667.81, 175.87], [704.54, 175.87], [718.53, 176.26], [677.08, 176.26]]}, {\"category\": \"car\", \"corners_3d\": [[646.28, 199.54], [671.07, 199.54], [676.09, 201.74], [649.26, 201.74], [646.28, 174.12], [671.07, 174.12], [676.09, 174.22], [649.26, 174.22]]}]\n```" - } - ] - }, - { - "id": 1819, - "data_source": "Mono3DRefer", - "scene_name": "001509", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[719.06, 287.62], [714.62, 282.85], [780.1, 283.06], [787.61, 287.84], [719.06, 153.56], [714.62, 154.76], [780.1, 154.71], [787.61, 153.5]]}]\n```" - } - ] - }, - { - "id": 1820, - "data_source": "Mono3DRefer", - "scene_name": "001513", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-402.74, 364.19], [-188.6, 322.16], [225.17, 317.98], [130.96, 357.37], [-402.74, 181.18], [-188.6, 179.35], [225.17, 179.17], [130.96, 180.89]]}, {\"category\": \"car\", \"corners_3d\": [[402.44, 263.48], [369.75, 277.31], [111.92, 276.93], [178.51, 263.2], [402.44, 180.88], [369.75, 182.1], [111.92, 182.07], [178.51, 180.86]]}, {\"category\": \"car\", \"corners_3d\": [[175.01, 252.27], [214.89, 245.12], [362.97, 245.29], [337.68, 252.48], [175.01, 178.87], [214.89, 178.33], [362.97, 178.34], [337.68, 178.88]]}, {\"category\": \"car\", \"corners_3d\": [[494.32, 229.92], [488.14, 234.38], [335.91, 235.02], [353.29, 230.47], [494.32, 182.62], [488.14, 183.38], [335.91, 183.49], [353.29, 182.71]]}, {\"category\": \"car\", \"corners_3d\": [[670.91, 230.63], [722.64, 230.64], [748.25, 243.79], [684.74, 243.78], [670.91, 185.61], [722.64, 185.61], [748.25, 188.52], [684.74, 188.52]]}, {\"category\": \"car\", \"corners_3d\": [[832.02, 221.93], [843.11, 225.1], [747.84, 224.63], [742.57, 221.52], [832.02, 182.77], [843.11, 183.41], [747.84, 183.32], [742.57, 182.69]]}]\n```" - } - ] - }, - { - "id": 1821, - "data_source": "Mono3DRefer", - "scene_name": "001513", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[511.84, 224.17], [503.47, 228.27], [326.27, 228.13], [347.68, 224.06], [511.84, 143.07], [503.47, 140.69], [326.27, 140.77], [347.68, 143.14]]}]\n```" - } - ] - }, - { - "id": 1822, - "data_source": "Mono3DRefer", - "scene_name": "001513", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-402.74, 364.19], [-188.6, 322.16], [225.17, 317.98], [130.96, 357.37], [-402.74, 181.18], [-188.6, 179.35], [225.17, 179.17], [130.96, 180.89]]}, {\"category\": \"car\", \"corners_3d\": [[402.44, 263.48], [369.75, 277.31], [111.92, 276.93], [178.51, 263.2], [402.44, 180.88], [369.75, 182.1], [111.92, 182.07], [178.51, 180.86]]}, {\"category\": \"car\", \"corners_3d\": [[175.01, 252.27], [214.89, 245.12], [362.97, 245.29], [337.68, 252.48], [175.01, 178.87], [214.89, 178.33], [362.97, 178.34], [337.68, 178.88]]}, {\"category\": \"car\", \"corners_3d\": [[494.32, 229.92], [488.14, 234.38], [335.91, 235.02], [353.29, 230.47], [494.32, 182.62], [488.14, 183.38], [335.91, 183.49], [353.29, 182.71]]}, {\"category\": \"van\", \"corners_3d\": [[511.84, 224.17], [503.47, 228.27], [326.27, 228.13], [347.68, 224.06], [511.84, 143.07], [503.47, 140.69], [326.27, 140.77], [347.68, 143.14]]}, {\"category\": \"car\", \"corners_3d\": [[670.91, 230.63], [722.64, 230.64], [748.25, 243.79], [684.74, 243.78], [670.91, 185.61], [722.64, 185.61], [748.25, 188.52], [684.74, 188.52]]}, {\"category\": \"car\", \"corners_3d\": [[832.02, 221.93], [843.11, 225.1], [747.84, 224.63], [742.57, 221.52], [832.02, 182.77], [843.11, 183.41], [747.84, 183.32], [742.57, 182.69]]}]\n```" - } - ] - }, - { - "id": 1823, - "data_source": "Mono3DRefer", - "scene_name": "001519", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[167.85, 327.55], [200.64, 316.53], [312.99, 316.56], [289.3, 327.59], [167.85, 147.73], [200.64, 150.18], [312.99, 150.18], [289.3, 147.72]]}]\n```" - } - ] - }, - { - "id": 1824, - "data_source": "Mono3DRefer", - "scene_name": "001520", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[33.32, 229.91], [-7.45, 229.84], [72.21, 222.65], [107.9, 222.7], [33.32, 191.52], [-7.45, 191.5], [72.21, 189.15], [107.9, 189.16]]}, {\"category\": \"car\", \"corners_3d\": [[138.09, 224.57], [105.29, 224.55], [156.44, 219.37], [185.98, 219.39], [138.09, 189.88], [105.29, 189.87], [156.44, 188.17], [185.98, 188.17]]}, {\"category\": \"car\", \"corners_3d\": [[195.04, 217.24], [170.29, 217.21], [216.39, 212.7], [238.65, 212.72], [195.04, 192.0], [170.29, 191.99], [216.39, 190.05], [238.65, 190.06]]}, {\"category\": \"car\", \"corners_3d\": [[237.29, 213.68], [214.3, 213.57], [248.7, 210.48], [270.01, 210.57], [237.29, 188.92], [214.3, 188.88], [248.7, 187.66], [270.01, 187.69]]}]\n```" - } - ] - }, - { - "id": 1825, - "data_source": "Mono3DRefer", - "scene_name": "001520", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[583.53, 206.37], [638.59, 206.4], [647.08, 219.8], [570.02, 219.75], [583.53, 139.33], [638.59, 139.3], [647.08, 125.88], [570.02, 125.94]]}]\n```" - } - ] - }, - { - "id": 1826, - "data_source": "Mono3DRefer", - "scene_name": "001520", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[583.53, 206.37], [638.59, 206.4], [647.08, 219.8], [570.02, 219.75], [583.53, 139.33], [638.59, 139.3], [647.08, 125.88], [570.02, 125.94]]}, {\"category\": \"car\", \"corners_3d\": [[33.32, 229.91], [-7.45, 229.84], [72.21, 222.65], [107.9, 222.7], [33.32, 191.52], [-7.45, 191.5], [72.21, 189.15], [107.9, 189.16]]}, {\"category\": \"car\", \"corners_3d\": [[138.09, 224.57], [105.29, 224.55], [156.44, 219.37], [185.98, 219.39], [138.09, 189.88], [105.29, 189.87], [156.44, 188.17], [185.98, 188.17]]}, {\"category\": \"car\", \"corners_3d\": [[195.04, 217.24], [170.29, 217.21], [216.39, 212.7], [238.65, 212.72], [195.04, 192.0], [170.29, 191.99], [216.39, 190.05], [238.65, 190.06]]}, {\"category\": \"car\", \"corners_3d\": [[237.29, 213.68], [214.3, 213.57], [248.7, 210.48], [270.01, 210.57], [237.29, 188.92], [214.3, 188.88], [248.7, 187.66], [270.01, 187.69]]}]\n```" - } - ] - }, - { - "id": 1827, - "data_source": "Mono3DRefer", - "scene_name": "001521", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[178.65, 377.14], [-38.92, 388.82], [151.96, 305.39], [281.85, 300.9], [178.65, 207.06], [-38.92, 209.02], [151.96, 195.05], [281.85, 194.3]]}, {\"category\": \"car\", \"corners_3d\": [[446.15, 223.41], [411.28, 223.62], [424.52, 218.56], [455.87, 218.38], [446.15, 192.17], [411.28, 192.25], [424.52, 190.32], [455.87, 190.25]]}, {\"category\": \"car\", \"corners_3d\": [[362.79, 279.87], [268.97, 281.62], [320.0, 259.03], [393.93, 257.92], [362.79, 198.05], [268.97, 198.46], [320.0, 193.14], [393.93, 192.88]]}]\n```" - } - ] - }, - { - "id": 1828, - "data_source": "Mono3DRefer", - "scene_name": "001526", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[793.93, 219.29], [807.47, 222.34], [682.54, 222.59], [676.68, 219.51], [793.93, 178.96], [807.47, 179.36], [682.54, 179.4], [676.68, 178.99]]}, {\"category\": \"car\", \"corners_3d\": [[647.13, 201.72], [646.86, 200.69], [706.87, 200.79], [709.37, 201.82], [647.13, 178.75], [646.86, 178.54], [706.87, 178.56], [709.37, 178.77]]}]\n```" - } - ] - }, - { - "id": 1829, - "data_source": "Mono3DRefer", - "scene_name": "001527", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[956.36, 200.29], [945.53, 199.45], [993.7, 199.43], [1006.04, 200.27], [956.36, 179.89], [945.53, 179.67], [993.7, 179.67], [1006.04, 179.88]]}, {\"category\": \"car\", \"corners_3d\": [[124.81, 240.81], [75.96, 240.81], [135.44, 233.23], [178.83, 233.23], [124.81, 200.62], [75.96, 200.62], [135.44, 197.53], [178.83, 197.52]]}]\n```" - } - ] - }, - { - "id": 1830, - "data_source": "Mono3DRefer", - "scene_name": "001527", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[589.81, 192.05], [612.85, 192.04], [615.88, 195.41], [588.77, 195.43], [589.81, 167.07], [612.85, 167.07], [615.88, 166.05], [588.77, 166.05]]}]\n```" - } - ] - }, - { - "id": 1831, - "data_source": "Mono3DRefer", - "scene_name": "001527", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[956.36, 200.29], [945.53, 199.45], [993.7, 199.43], [1006.04, 200.27], [956.36, 179.89], [945.53, 179.67], [993.7, 179.67], [1006.04, 179.88]]}, {\"category\": \"bus\", \"corners_3d\": [[589.81, 192.05], [612.85, 192.04], [615.88, 195.41], [588.77, 195.43], [589.81, 167.07], [612.85, 167.07], [615.88, 166.05], [588.77, 166.05]]}, {\"category\": \"car\", \"corners_3d\": [[124.81, 240.81], [75.96, 240.81], [135.44, 233.23], [178.83, 233.23], [124.81, 200.62], [75.96, 200.62], [135.44, 197.53], [178.83, 197.52]]}]\n```" - } - ] - }, - { - "id": 1832, - "data_source": "Mono3DRefer", - "scene_name": "001529", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1360.38, 530.51], [1164.55, 523.42], [1054.69, 448.72], [1205.4, 453.04], [1360.38, 98.48], [1164.55, 100.14], [1054.69, 117.65], [1205.4, 116.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[925.99, 204.74], [950.01, 205.21], [946.32, 205.68], [921.92, 205.2], [925.99, 145.65], [950.01, 144.98], [946.32, 144.3], [921.92, 144.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[582.3, 297.24], [545.2, 294.09], [581.4, 284.23], [616.12, 286.86], [582.3, 152.73], [545.2, 153.48], [581.4, 155.82], [616.12, 155.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[488.78, 288.21], [460.19, 284.94], [514.04, 275.82], [541.6, 278.53], [488.78, 149.26], [460.19, 150.21], [514.04, 152.85], [541.6, 152.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[530.52, 291.93], [493.86, 288.07], [539.93, 279.1], [574.95, 282.34], [530.52, 137.16], [493.86, 138.66], [539.93, 142.15], [574.95, 140.89]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[844.43, 227.71], [839.17, 225.93], [884.99, 226.33], [892.12, 228.14], [844.43, 151.98], [839.17, 153.05], [884.99, 152.81], [892.12, 151.71]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1088.68, 232.76], [1067.69, 230.49], [1095.58, 230.49], [1117.84, 232.76], [1088.68, 144.38], [1067.69, 145.95], [1095.58, 145.94], [1117.84, 144.37]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1129.04, 232.69], [1104.74, 230.37], [1140.91, 230.29], [1166.85, 232.6], [1129.04, 130.83], [1104.74, 133.03], [1140.91, 133.11], [1166.85, 130.91]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1107.62, 234.1], [1093.69, 232.6], [1123.76, 232.62], [1138.56, 234.13], [1107.62, 150.76], [1093.69, 151.6], [1123.76, 151.58], [1138.56, 150.75]]}]\n```" - } - ] - }, - { - "id": 1833, - "data_source": "Mono3DRefer", - "scene_name": "001529", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[875.01, 239.37], [891.1, 240.51], [841.14, 245.46], [824.77, 244.13], [875.01, 148.68], [891.1, 148.06], [841.14, 145.38], [824.77, 146.1]]}]\n```" - } - ] - }, - { - "id": 1834, - "data_source": "Mono3DRefer", - "scene_name": "001529", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[601.79, 179.66], [618.64, 179.66], [619.54, 179.6], [601.6, 179.6], [601.79, 160.58], [618.64, 160.58], [619.54, 159.3], [601.6, 159.3]]}]\n```" - } - ] - }, - { - "id": 1835, - "data_source": "Mono3DRefer", - "scene_name": "001529", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[875.01, 239.37], [891.1, 240.51], [841.14, 245.46], [824.77, 244.13], [875.01, 148.68], [891.1, 148.06], [841.14, 145.38], [824.77, 146.1]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1360.38, 530.51], [1164.55, 523.42], [1054.69, 448.72], [1205.4, 453.04], [1360.38, 98.48], [1164.55, 100.14], [1054.69, 117.65], [1205.4, 116.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[925.99, 204.74], [950.01, 205.21], [946.32, 205.68], [921.92, 205.2], [925.99, 145.65], [950.01, 144.98], [946.32, 144.3], [921.92, 144.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[582.3, 297.24], [545.2, 294.09], [581.4, 284.23], [616.12, 286.86], [582.3, 152.73], [545.2, 153.48], [581.4, 155.82], [616.12, 155.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[488.78, 288.21], [460.19, 284.94], [514.04, 275.82], [541.6, 278.53], [488.78, 149.26], [460.19, 150.21], [514.04, 152.85], [541.6, 152.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[530.52, 291.93], [493.86, 288.07], [539.93, 279.1], [574.95, 282.34], [530.52, 137.16], [493.86, 138.66], [539.93, 142.15], [574.95, 140.89]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[844.43, 227.71], [839.17, 225.93], [884.99, 226.33], [892.12, 228.14], [844.43, 151.98], [839.17, 153.05], [884.99, 152.81], [892.12, 151.71]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1088.68, 232.76], [1067.69, 230.49], [1095.58, 230.49], [1117.84, 232.76], [1088.68, 144.38], [1067.69, 145.95], [1095.58, 145.94], [1117.84, 144.37]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1129.04, 232.69], [1104.74, 230.37], [1140.91, 230.29], [1166.85, 232.6], [1129.04, 130.83], [1104.74, 133.03], [1140.91, 133.11], [1166.85, 130.91]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1107.62, 234.1], [1093.69, 232.6], [1123.76, 232.62], [1138.56, 234.13], [1107.62, 150.76], [1093.69, 151.6], [1123.76, 151.58], [1138.56, 150.75]]}]\n```" - } - ] - }, - { - "id": 1836, - "data_source": "Mono3DRefer", - "scene_name": "001529", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1360.38, 530.51], [1164.55, 523.42], [1054.69, 448.72], [1205.4, 453.04], [1360.38, 98.48], [1164.55, 100.14], [1054.69, 117.65], [1205.4, 116.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[925.99, 204.74], [950.01, 205.21], [946.32, 205.68], [921.92, 205.2], [925.99, 145.65], [950.01, 144.98], [946.32, 144.3], [921.92, 144.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[582.3, 297.24], [545.2, 294.09], [581.4, 284.23], [616.12, 286.86], [582.3, 152.73], [545.2, 153.48], [581.4, 155.82], [616.12, 155.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[488.78, 288.21], [460.19, 284.94], [514.04, 275.82], [541.6, 278.53], [488.78, 149.26], [460.19, 150.21], [514.04, 152.85], [541.6, 152.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[530.52, 291.93], [493.86, 288.07], [539.93, 279.1], [574.95, 282.34], [530.52, 137.16], [493.86, 138.66], [539.93, 142.15], [574.95, 140.89]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[844.43, 227.71], [839.17, 225.93], [884.99, 226.33], [892.12, 228.14], [844.43, 151.98], [839.17, 153.05], [884.99, 152.81], [892.12, 151.71]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1088.68, 232.76], [1067.69, 230.49], [1095.58, 230.49], [1117.84, 232.76], [1088.68, 144.38], [1067.69, 145.95], [1095.58, 145.94], [1117.84, 144.37]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1129.04, 232.69], [1104.74, 230.37], [1140.91, 230.29], [1166.85, 232.6], [1129.04, 130.83], [1104.74, 133.03], [1140.91, 133.11], [1166.85, 130.91]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1107.62, 234.1], [1093.69, 232.6], [1123.76, 232.62], [1138.56, 234.13], [1107.62, 150.76], [1093.69, 151.6], [1123.76, 151.58], [1138.56, 150.75]]}, {\"category\": \"van\", \"corners_3d\": [[601.79, 179.66], [618.64, 179.66], [619.54, 179.6], [601.6, 179.6], [601.79, 160.58], [618.64, 160.58], [619.54, 159.3], [601.6, 159.3]]}]\n```" - } - ] - }, - { - "id": 1837, - "data_source": "Mono3DRefer", - "scene_name": "001529", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[875.01, 239.37], [891.1, 240.51], [841.14, 245.46], [824.77, 244.13], [875.01, 148.68], [891.1, 148.06], [841.14, 145.38], [824.77, 146.1]]}, {\"category\": \"van\", \"corners_3d\": [[601.79, 179.66], [618.64, 179.66], [619.54, 179.6], [601.6, 179.6], [601.79, 160.58], [618.64, 160.58], [619.54, 159.3], [601.6, 159.3]]}]\n```" - } - ] - }, - { - "id": 1838, - "data_source": "Mono3DRefer", - "scene_name": "001529", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[875.01, 239.37], [891.1, 240.51], [841.14, 245.46], [824.77, 244.13], [875.01, 148.68], [891.1, 148.06], [841.14, 145.38], [824.77, 146.1]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1360.38, 530.51], [1164.55, 523.42], [1054.69, 448.72], [1205.4, 453.04], [1360.38, 98.48], [1164.55, 100.14], [1054.69, 117.65], [1205.4, 116.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[925.99, 204.74], [950.01, 205.21], [946.32, 205.68], [921.92, 205.2], [925.99, 145.65], [950.01, 144.98], [946.32, 144.3], [921.92, 144.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[582.3, 297.24], [545.2, 294.09], [581.4, 284.23], [616.12, 286.86], [582.3, 152.73], [545.2, 153.48], [581.4, 155.82], [616.12, 155.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[488.78, 288.21], [460.19, 284.94], [514.04, 275.82], [541.6, 278.53], [488.78, 149.26], [460.19, 150.21], [514.04, 152.85], [541.6, 152.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[530.52, 291.93], [493.86, 288.07], [539.93, 279.1], [574.95, 282.34], [530.52, 137.16], [493.86, 138.66], [539.93, 142.15], [574.95, 140.89]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[844.43, 227.71], [839.17, 225.93], [884.99, 226.33], [892.12, 228.14], [844.43, 151.98], [839.17, 153.05], [884.99, 152.81], [892.12, 151.71]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1088.68, 232.76], [1067.69, 230.49], [1095.58, 230.49], [1117.84, 232.76], [1088.68, 144.38], [1067.69, 145.95], [1095.58, 145.94], [1117.84, 144.37]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1129.04, 232.69], [1104.74, 230.37], [1140.91, 230.29], [1166.85, 232.6], [1129.04, 130.83], [1104.74, 133.03], [1140.91, 133.11], [1166.85, 130.91]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1107.62, 234.1], [1093.69, 232.6], [1123.76, 232.62], [1138.56, 234.13], [1107.62, 150.76], [1093.69, 151.6], [1123.76, 151.58], [1138.56, 150.75]]}, {\"category\": \"van\", \"corners_3d\": [[601.79, 179.66], [618.64, 179.66], [619.54, 179.6], [601.6, 179.6], [601.79, 160.58], [618.64, 160.58], [619.54, 159.3], [601.6, 159.3]]}]\n```" - } - ] - }, - { - "id": 1839, - "data_source": "Mono3DRefer", - "scene_name": "001532", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[621.08, 215.12], [647.59, 215.11], [652.52, 219.77], [623.09, 219.79], [621.08, 183.61], [647.59, 183.61], [652.52, 184.8], [623.09, 184.8]]}]\n```" - } - ] - }, - { - "id": 1840, - "data_source": "Mono3DRefer", - "scene_name": "001532", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[570.98, 213.97], [596.25, 214.0], [593.25, 218.35], [565.32, 218.32], [570.98, 192.04], [596.25, 192.05], [593.25, 194.08], [565.32, 194.07]]}]\n```" - } - ] - }, - { - "id": 1841, - "data_source": "Mono3DRefer", - "scene_name": "001532", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[523.14, 208.67], [495.43, 209.42], [442.85, 203.21], [466.84, 202.69], [523.14, 169.07], [495.43, 168.99], [442.85, 169.65], [466.84, 169.7]]}]\n```" - } - ] - }, - { - "id": 1842, - "data_source": "Mono3DRefer", - "scene_name": "001532", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[570.98, 213.97], [596.25, 214.0], [593.25, 218.35], [565.32, 218.32], [570.98, 192.04], [596.25, 192.05], [593.25, 194.08], [565.32, 194.07]]}, {\"category\": \"van\", \"corners_3d\": [[621.08, 215.12], [647.59, 215.11], [652.52, 219.77], [623.09, 219.79], [621.08, 183.61], [647.59, 183.61], [652.52, 184.8], [623.09, 184.8]]}]\n```" - } - ] - }, - { - "id": 1843, - "data_source": "Mono3DRefer", - "scene_name": "001532", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[621.08, 215.12], [647.59, 215.11], [652.52, 219.77], [623.09, 219.79], [621.08, 183.61], [647.59, 183.61], [652.52, 184.8], [623.09, 184.8]]}, {\"category\": \"tram\", \"corners_3d\": [[523.14, 208.67], [495.43, 209.42], [442.85, 203.21], [466.84, 202.69], [523.14, 169.07], [495.43, 168.99], [442.85, 169.65], [466.84, 169.7]]}]\n```" - } - ] - }, - { - "id": 1844, - "data_source": "Mono3DRefer", - "scene_name": "001532", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[570.98, 213.97], [596.25, 214.0], [593.25, 218.35], [565.32, 218.32], [570.98, 192.04], [596.25, 192.05], [593.25, 194.08], [565.32, 194.07]]}, {\"category\": \"tram\", \"corners_3d\": [[523.14, 208.67], [495.43, 209.42], [442.85, 203.21], [466.84, 202.69], [523.14, 169.07], [495.43, 168.99], [442.85, 169.65], [466.84, 169.7]]}]\n```" - } - ] - }, - { - "id": 1845, - "data_source": "Mono3DRefer", - "scene_name": "001532", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[570.98, 213.97], [596.25, 214.0], [593.25, 218.35], [565.32, 218.32], [570.98, 192.04], [596.25, 192.05], [593.25, 194.08], [565.32, 194.07]]}, {\"category\": \"van\", \"corners_3d\": [[621.08, 215.12], [647.59, 215.11], [652.52, 219.77], [623.09, 219.79], [621.08, 183.61], [647.59, 183.61], [652.52, 184.8], [623.09, 184.8]]}, {\"category\": \"tram\", \"corners_3d\": [[523.14, 208.67], [495.43, 209.42], [442.85, 203.21], [466.84, 202.69], [523.14, 169.07], [495.43, 168.99], [442.85, 169.65], [466.84, 169.7]]}]\n```" - } - ] - }, - { - "id": 1846, - "data_source": "Mono3DRefer", - "scene_name": "001533", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[584.14, 226.84], [640.57, 226.87], [645.86, 235.48], [577.78, 235.43], [584.14, 175.57], [640.57, 175.56], [645.86, 173.56], [577.78, 173.57]]}, {\"category\": \"car\", \"corners_3d\": [[160.17, 261.32], [244.08, 261.23], [111.52, 289.5], [-3.83, 289.65], [160.17, 187.62], [244.08, 187.61], [111.52, 188.51], [-3.83, 188.51]]}, {\"category\": \"car\", \"corners_3d\": [[298.51, 235.2], [355.37, 235.21], [307.07, 244.77], [239.33, 244.77], [298.51, 184.15], [355.37, 184.15], [307.07, 183.94], [239.33, 183.94]]}, {\"category\": \"car\", \"corners_3d\": [[419.9, 243.03], [486.53, 243.04], [452.81, 259.12], [367.65, 259.11], [419.9, 177.99], [486.53, 177.99], [452.81, 175.98], [367.65, 175.98]]}, {\"category\": \"car\", \"corners_3d\": [[383.2, 216.2], [421.39, 216.18], [393.84, 220.94], [349.75, 220.96], [383.2, 182.59], [421.39, 182.59], [393.84, 182.18], [349.75, 182.18]]}, {\"category\": \"car\", \"corners_3d\": [[475.94, 216.46], [516.91, 216.47], [504.23, 220.83], [457.55, 220.82], [475.94, 178.64], [516.91, 178.64], [504.23, 177.72], [457.55, 177.72]]}, {\"category\": \"car\", \"corners_3d\": [[502.0, 205.51], [529.93, 205.5], [523.31, 207.4], [492.75, 207.41], [502.0, 181.13], [529.93, 181.13], [523.31, 180.75], [492.75, 180.75]]}, {\"category\": \"car\", \"corners_3d\": [[523.91, 200.79], [547.6, 200.78], [542.55, 202.27], [516.61, 202.27], [523.91, 180.15], [547.6, 180.15], [542.55, 179.67], [516.61, 179.66]]}, {\"category\": \"car\", \"corners_3d\": [[343.87, 213.04], [376.74, 213.01], [346.79, 216.86], [309.34, 216.89], [343.87, 184.42], [376.74, 184.42], [346.79, 184.31], [309.34, 184.31]]}]\n```" - } - ] - }, - { - "id": 1847, - "data_source": "Mono3DRefer", - "scene_name": "001533", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[120.69, 338.03], [308.98, 338.06], [-889.68, 951.23], [-1831.28, 950.43], [120.69, 103.06], [308.98, 103.04], [-889.68, -226.63], [-1831.28, -226.2]]}]\n```" - } - ] - }, - { - "id": 1848, - "data_source": "Mono3DRefer", - "scene_name": "001533", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[584.14, 226.84], [640.57, 226.87], [645.86, 235.48], [577.78, 235.43], [584.14, 175.57], [640.57, 175.56], [645.86, 173.56], [577.78, 173.57]]}, {\"category\": \"van\", \"corners_3d\": [[120.69, 338.03], [308.98, 338.06], [-889.68, 951.23], [-1831.28, 950.43], [120.69, 103.06], [308.98, 103.04], [-889.68, -226.63], [-1831.28, -226.2]]}, {\"category\": \"car\", \"corners_3d\": [[160.17, 261.32], [244.08, 261.23], [111.52, 289.5], [-3.83, 289.65], [160.17, 187.62], [244.08, 187.61], [111.52, 188.51], [-3.83, 188.51]]}, {\"category\": \"car\", \"corners_3d\": [[298.51, 235.2], [355.37, 235.21], [307.07, 244.77], [239.33, 244.77], [298.51, 184.15], [355.37, 184.15], [307.07, 183.94], [239.33, 183.94]]}, {\"category\": \"car\", \"corners_3d\": [[419.9, 243.03], [486.53, 243.04], [452.81, 259.12], [367.65, 259.11], [419.9, 177.99], [486.53, 177.99], [452.81, 175.98], [367.65, 175.98]]}, {\"category\": \"car\", \"corners_3d\": [[383.2, 216.2], [421.39, 216.18], [393.84, 220.94], [349.75, 220.96], [383.2, 182.59], [421.39, 182.59], [393.84, 182.18], [349.75, 182.18]]}, {\"category\": \"car\", \"corners_3d\": [[475.94, 216.46], [516.91, 216.47], [504.23, 220.83], [457.55, 220.82], [475.94, 178.64], [516.91, 178.64], [504.23, 177.72], [457.55, 177.72]]}, {\"category\": \"car\", \"corners_3d\": [[502.0, 205.51], [529.93, 205.5], [523.31, 207.4], [492.75, 207.41], [502.0, 181.13], [529.93, 181.13], [523.31, 180.75], [492.75, 180.75]]}, {\"category\": \"car\", \"corners_3d\": [[523.91, 200.79], [547.6, 200.78], [542.55, 202.27], [516.61, 202.27], [523.91, 180.15], [547.6, 180.15], [542.55, 179.67], [516.61, 179.66]]}, {\"category\": \"car\", \"corners_3d\": [[343.87, 213.04], [376.74, 213.01], [346.79, 216.86], [309.34, 216.89], [343.87, 184.42], [376.74, 184.42], [346.79, 184.31], [309.34, 184.31]]}]\n```" - } - ] - }, - { - "id": 1849, - "data_source": "Mono3DRefer", - "scene_name": "001539", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[582.82, 253.45], [657.48, 253.46], [670.12, 274.98], [575.53, 274.97], [582.82, 182.46], [657.48, 182.46], [670.12, 185.02], [575.53, 185.02]]}, {\"category\": \"car\", \"corners_3d\": [[263.79, 237.79], [219.9, 237.75], [266.4, 230.14], [305.15, 230.17], [263.79, 204.59], [219.9, 204.57], [266.4, 200.85], [305.15, 200.87]]}, {\"category\": \"car\", \"corners_3d\": [[155.01, 243.95], [119.65, 242.78], [180.81, 236.58], [213.91, 237.54], [155.01, 209.15], [119.65, 208.56], [180.81, 205.39], [213.91, 205.88]]}, {\"category\": \"car\", \"corners_3d\": [[396.96, 223.22], [363.56, 223.22], [390.95, 217.6], [420.62, 217.6], [396.96, 196.51], [363.56, 196.51], [390.95, 193.87], [420.62, 193.87]]}, {\"category\": \"car\", \"corners_3d\": [[1076.57, 243.17], [1045.95, 237.8], [1191.76, 238.94], [1235.25, 244.51], [1076.57, 188.09], [1045.95, 186.92], [1191.76, 187.17], [1235.25, 188.38]]}, {\"category\": \"car\", \"corners_3d\": [[335.0, 222.35], [305.97, 222.37], [330.92, 218.2], [357.49, 218.18], [335.0, 199.04], [305.97, 199.05], [330.92, 196.84], [357.49, 196.83]]}]\n```" - } - ] - }, - { - "id": 1850, - "data_source": "Mono3DRefer", - "scene_name": "001541", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[56.07, 310.54], [-57.31, 309.9], [141.7, 270.25], [222.84, 270.58], [56.07, 162.6], [-57.31, 162.65], [141.7, 165.6], [222.84, 165.58]]}, {\"category\": \"car\", \"corners_3d\": [[313.91, 250.51], [255.16, 250.13], [325.27, 236.49], [373.9, 236.75], [313.91, 184.96], [255.16, 184.9], [325.27, 182.78], [373.9, 182.82]]}, {\"category\": \"car\", \"corners_3d\": [[424.92, 224.41], [384.35, 224.24], [418.08, 217.77], [453.62, 217.91], [424.92, 187.07], [384.35, 187.02], [418.08, 185.24], [453.62, 185.28]]}, {\"category\": \"car\", \"corners_3d\": [[472.8, 213.88], [440.55, 213.75], [461.85, 209.77], [491.02, 209.88], [472.8, 177.13], [440.55, 177.12], [461.85, 176.71], [491.02, 176.72]]}, {\"category\": \"car\", \"corners_3d\": [[497.6, 210.2], [472.21, 210.11], [486.64, 207.24], [510.1, 207.32], [497.6, 184.41], [472.21, 184.38], [486.64, 183.49], [510.1, 183.52]]}]\n```" - } - ] - }, - { - "id": 1851, - "data_source": "Mono3DRefer", - "scene_name": "001541", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[712.7, 184.79], [730.94, 184.82], [746.63, 188.7], [722.45, 188.65], [712.7, 159.47], [730.94, 159.43], [746.63, 155.08], [722.45, 155.14]]}]\n```" - } - ] - }, - { - "id": 1852, - "data_source": "Mono3DRefer", - "scene_name": "001541", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[56.07, 310.54], [-57.31, 309.9], [141.7, 270.25], [222.84, 270.58], [56.07, 162.6], [-57.31, 162.65], [141.7, 165.6], [222.84, 165.58]]}, {\"category\": \"car\", \"corners_3d\": [[313.91, 250.51], [255.16, 250.13], [325.27, 236.49], [373.9, 236.75], [313.91, 184.96], [255.16, 184.9], [325.27, 182.78], [373.9, 182.82]]}, {\"category\": \"car\", \"corners_3d\": [[424.92, 224.41], [384.35, 224.24], [418.08, 217.77], [453.62, 217.91], [424.92, 187.07], [384.35, 187.02], [418.08, 185.24], [453.62, 185.28]]}, {\"category\": \"tram\", \"corners_3d\": [[712.7, 184.79], [730.94, 184.82], [746.63, 188.7], [722.45, 188.65], [712.7, 159.47], [730.94, 159.43], [746.63, 155.08], [722.45, 155.14]]}, {\"category\": \"car\", \"corners_3d\": [[472.8, 213.88], [440.55, 213.75], [461.85, 209.77], [491.02, 209.88], [472.8, 177.13], [440.55, 177.12], [461.85, 176.71], [491.02, 176.72]]}, {\"category\": \"car\", \"corners_3d\": [[497.6, 210.2], [472.21, 210.11], [486.64, 207.24], [510.1, 207.32], [497.6, 184.41], [472.21, 184.38], [486.64, 183.49], [510.1, 183.52]]}]\n```" - } - ] - }, - { - "id": 1853, - "data_source": "Mono3DRefer", - "scene_name": "001548", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[589.05, 300.04], [590.21, 294.08], [675.0, 294.23], [678.29, 300.22], [589.05, 158.9], [590.21, 159.98], [675.0, 159.95], [678.29, 158.87]]}]\n```" - } - ] - }, - { - "id": 1854, - "data_source": "Mono3DRefer", - "scene_name": "001551", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[751.45, 218.81], [784.92, 221.04], [705.6, 225.98], [672.9, 223.29], [751.45, 180.15], [784.92, 180.5], [705.6, 181.28], [672.9, 180.85]]}, {\"category\": \"car\", \"corners_3d\": [[331.37, 299.98], [221.91, 298.61], [340.37, 264.57], [420.92, 265.3], [331.37, 187.56], [221.91, 187.4], [340.37, 183.46], [420.92, 183.55]]}, {\"category\": \"car\", \"corners_3d\": [[505.65, 233.14], [449.11, 232.86], [478.28, 224.25], [526.78, 224.45], [505.65, 180.83], [449.11, 180.79], [478.28, 179.66], [526.78, 179.68]]}, {\"category\": \"car\", \"corners_3d\": [[529.65, 225.72], [477.92, 225.53], [504.13, 217.27], [547.8, 217.4], [529.65, 172.54], [477.92, 172.54], [504.13, 172.59], [547.8, 172.59]]}, {\"category\": \"car\", \"corners_3d\": [[556.28, 210.42], [517.79, 210.56], [522.66, 206.04], [556.53, 205.94], [556.28, 179.07], [517.79, 179.09], [522.66, 178.35], [556.53, 178.33]]}]\n```" - } - ] - }, - { - "id": 1855, - "data_source": "Mono3DRefer", - "scene_name": "001552", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[394.25, 288.3], [490.5, 287.7], [439.86, 351.77], [289.06, 353.23], [394.25, 195.68], [490.5, 195.56], [439.86, 208.23], [289.06, 208.52]]}, {\"category\": \"car\", \"corners_3d\": [[714.22, 285.83], [818.1, 285.68], [919.66, 338.93], [766.86, 339.25], [714.22, 170.94], [818.1, 170.94], [919.66, 170.04], [766.86, 170.03]]}, {\"category\": \"car\", \"corners_3d\": [[693.94, 249.77], [757.99, 249.71], [810.21, 275.59], [724.62, 275.7], [693.94, 185.16], [757.99, 185.15], [810.21, 189.3], [724.62, 189.32]]}, {\"category\": \"car\", \"corners_3d\": [[532.23, 242.91], [478.82, 242.91], [499.9, 231.57], [544.66, 231.56], [532.23, 194.61], [478.82, 194.61], [499.9, 191.09], [544.66, 191.09]]}, {\"category\": \"car\", \"corners_3d\": [[657.93, 236.17], [711.76, 235.89], [745.56, 250.59], [679.3, 251.02], [657.93, 188.76], [711.76, 188.69], [745.56, 192.38], [679.3, 192.49]]}, {\"category\": \"car\", \"corners_3d\": [[568.01, 210.62], [542.02, 210.68], [544.97, 207.74], [568.95, 207.7], [568.01, 187.25], [542.02, 187.27], [544.97, 186.15], [568.95, 186.13]]}, {\"category\": \"car\", \"corners_3d\": [[629.88, 200.01], [652.73, 200.0], [656.94, 202.3], [632.16, 202.31], [629.88, 179.36], [652.73, 179.36], [656.94, 179.91], [632.16, 179.91]]}]\n```" - } - ] - }, - { - "id": 1856, - "data_source": "Mono3DRefer", - "scene_name": "001558", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[703.07, 298.84], [650.78, 300.67], [635.13, 290.81], [683.41, 289.26], [703.07, 131.42], [650.78, 130.66], [635.13, 134.75], [683.41, 135.39]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[767.14, 294.64], [719.24, 296.56], [702.79, 289.77], [748.2, 288.07], [767.14, 148.8], [719.24, 148.27], [702.79, 150.15], [748.2, 150.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[869.13, 269.5], [841.68, 271.13], [809.2, 265.48], [835.51, 264.05], [869.13, 151.41], [841.68, 150.88], [809.2, 152.73], [835.51, 153.19]]}]\n```" - } - ] - }, - { - "id": 1857, - "data_source": "Mono3DRefer", - "scene_name": "001559", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[581.18, 196.6], [628.47, 196.56], [641.43, 204.24], [578.8, 204.32], [581.18, 138.94], [628.47, 139.0], [641.43, 128.02], [578.8, 127.91]]}]\n```" - } - ] - }, - { - "id": 1858, - "data_source": "Mono3DRefer", - "scene_name": "001561", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[563.54, 196.68], [545.09, 196.69], [548.17, 195.21], [565.47, 195.2], [563.54, 179.23], [545.09, 179.23], [548.17, 178.83], [565.47, 178.83]]}, {\"category\": \"car\", \"corners_3d\": [[539.67, 212.05], [508.19, 212.03], [519.02, 208.11], [547.34, 208.12], [539.67, 178.71], [508.19, 178.7], [519.02, 178.12], [547.34, 178.12]]}]\n```" - } - ] - }, - { - "id": 1859, - "data_source": "Mono3DRefer", - "scene_name": "001563", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1008.76, 241.07], [948.91, 236.67], [1026.39, 229.03], [1083.31, 232.41], [1008.76, 180.6], [948.91, 180.1], [1026.39, 179.23], [1083.31, 179.62]]}, {\"category\": \"car\", \"corners_3d\": [[1104.55, 230.65], [1042.42, 227.14], [1098.82, 221.61], [1157.5, 224.42], [1104.55, 181.79], [1042.42, 181.25], [1098.82, 180.39], [1157.5, 180.83]]}, {\"category\": \"car\", \"corners_3d\": [[1192.86, 218.14], [1146.39, 216.22], [1182.09, 212.93], [1226.34, 214.56], [1192.86, 180.11], [1146.39, 179.81], [1182.09, 179.28], [1226.34, 179.54]]}]\n```" - } - ] - }, - { - "id": 1860, - "data_source": "Mono3DRefer", - "scene_name": "001564", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[334.14, 228.35], [381.16, 228.38], [336.45, 236.64], [280.49, 236.59], [334.14, 185.18], [381.16, 185.18], [336.45, 185.17], [280.49, 185.17]]}, {\"category\": \"car\", \"corners_3d\": [[401.68, 215.7], [435.61, 215.69], [413.6, 219.75], [375.14, 219.77], [401.68, 184.3], [435.61, 184.3], [413.6, 184.18], [375.14, 184.18]]}, {\"category\": \"car\", \"corners_3d\": [[520.54, 202.85], [546.87, 202.84], [541.85, 204.49], [513.06, 204.5], [520.54, 180.62], [546.87, 180.62], [541.85, 180.19], [513.06, 180.19]]}, {\"category\": \"car\", \"corners_3d\": [[534.38, 196.97], [555.68, 196.97], [552.74, 197.88], [529.78, 197.89], [534.38, 178.65], [555.68, 178.65], [552.74, 178.14], [529.78, 178.13]]}, {\"category\": \"car\", \"corners_3d\": [[589.39, 200.13], [613.74, 200.13], [614.25, 201.42], [587.8, 201.42], [589.39, 177.23], [613.74, 177.23], [614.25, 176.54], [587.8, 176.54]]}, {\"category\": \"car\", \"corners_3d\": [[619.51, 191.07], [632.85, 191.07], [633.69, 191.34], [619.73, 191.34], [619.51, 177.45], [632.85, 177.45], [633.69, 177.09], [619.73, 177.09]]}, {\"category\": \"car\", \"corners_3d\": [[697.56, 242.22], [760.95, 242.12], [803.99, 256.74], [724.37, 256.89], [697.56, 185.58], [760.95, 185.58], [803.99, 185.68], [724.37, 185.68]]}, {\"category\": \"car\", \"corners_3d\": [[356.1, 271.79], [449.9, 272.02], [383.18, 305.65], [253.44, 305.19], [356.1, 179.88], [449.9, 179.87], [383.18, 177.79], [253.44, 177.82]]}, {\"category\": \"car\", \"corners_3d\": [[640.47, 203.35], [670.22, 203.33], [679.65, 205.36], [646.57, 205.39], [640.47, 177.11], [670.22, 177.12], [679.65, 176.21], [646.57, 176.19]]}, {\"category\": \"car\", \"corners_3d\": [[135.44, 265.49], [221.79, 265.5], [74.95, 296.04], [-44.22, 296.02], [135.44, 185.7], [221.79, 185.7], [74.95, 185.89], [-44.22, 185.89]]}, {\"category\": \"car\", \"corners_3d\": [[448.08, 233.02], [504.63, 233.06], [476.4, 245.31], [405.42, 245.24], [448.08, 181.31], [504.63, 181.3], [476.4, 180.3], [405.42, 180.31]]}]\n```" - } - ] - }, - { - "id": 1861, - "data_source": "Mono3DRefer", - "scene_name": "001564", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[439.2, 205.09], [466.19, 205.09], [454.12, 206.78], [424.82, 206.78], [439.2, 173.12], [466.19, 173.12], [454.12, 172.09], [424.82, 172.09]]}, {\"category\": \"van\", \"corners_3d\": [[473.08, 198.87], [498.53, 198.86], [487.51, 200.34], [459.3, 200.34], [473.08, 173.17], [498.53, 173.18], [487.51, 171.88], [459.3, 171.87]]}]\n```" - } - ] - }, - { - "id": 1862, - "data_source": "Mono3DRefer", - "scene_name": "001564", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[334.14, 228.35], [381.16, 228.38], [336.45, 236.64], [280.49, 236.59], [334.14, 185.18], [381.16, 185.18], [336.45, 185.17], [280.49, 185.17]]}, {\"category\": \"car\", \"corners_3d\": [[401.68, 215.7], [435.61, 215.69], [413.6, 219.75], [375.14, 219.77], [401.68, 184.3], [435.61, 184.3], [413.6, 184.18], [375.14, 184.18]]}, {\"category\": \"van\", \"corners_3d\": [[439.2, 205.09], [466.19, 205.09], [454.12, 206.78], [424.82, 206.78], [439.2, 173.12], [466.19, 173.12], [454.12, 172.09], [424.82, 172.09]]}, {\"category\": \"car\", \"corners_3d\": [[520.54, 202.85], [546.87, 202.84], [541.85, 204.49], [513.06, 204.5], [520.54, 180.62], [546.87, 180.62], [541.85, 180.19], [513.06, 180.19]]}, {\"category\": \"van\", \"corners_3d\": [[473.08, 198.87], [498.53, 198.86], [487.51, 200.34], [459.3, 200.34], [473.08, 173.17], [498.53, 173.18], [487.51, 171.88], [459.3, 171.87]]}, {\"category\": \"car\", \"corners_3d\": [[534.38, 196.97], [555.68, 196.97], [552.74, 197.88], [529.78, 197.89], [534.38, 178.65], [555.68, 178.65], [552.74, 178.14], [529.78, 178.13]]}, {\"category\": \"car\", \"corners_3d\": [[589.39, 200.13], [613.74, 200.13], [614.25, 201.42], [587.8, 201.42], [589.39, 177.23], [613.74, 177.23], [614.25, 176.54], [587.8, 176.54]]}, {\"category\": \"car\", \"corners_3d\": [[619.51, 191.07], [632.85, 191.07], [633.69, 191.34], [619.73, 191.34], [619.51, 177.45], [632.85, 177.45], [633.69, 177.09], [619.73, 177.09]]}, {\"category\": \"car\", \"corners_3d\": [[697.56, 242.22], [760.95, 242.12], [803.99, 256.74], [724.37, 256.89], [697.56, 185.58], [760.95, 185.58], [803.99, 185.68], [724.37, 185.68]]}, {\"category\": \"car\", \"corners_3d\": [[356.1, 271.79], [449.9, 272.02], [383.18, 305.65], [253.44, 305.19], [356.1, 179.88], [449.9, 179.87], [383.18, 177.79], [253.44, 177.82]]}, {\"category\": \"car\", \"corners_3d\": [[640.47, 203.35], [670.22, 203.33], [679.65, 205.36], [646.57, 205.39], [640.47, 177.11], [670.22, 177.12], [679.65, 176.21], [646.57, 176.19]]}, {\"category\": \"car\", \"corners_3d\": [[135.44, 265.49], [221.79, 265.5], [74.95, 296.04], [-44.22, 296.02], [135.44, 185.7], [221.79, 185.7], [74.95, 185.89], [-44.22, 185.89]]}, {\"category\": \"car\", \"corners_3d\": [[448.08, 233.02], [504.63, 233.06], [476.4, 245.31], [405.42, 245.24], [448.08, 181.31], [504.63, 181.3], [476.4, 180.3], [405.42, 180.31]]}]\n```" - } - ] - }, - { - "id": 1863, - "data_source": "Mono3DRefer", - "scene_name": "001570", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[652.32, 242.38], [677.54, 242.32], [688.55, 249.55], [660.39, 249.63], [652.32, 163.41], [677.54, 163.43], [688.55, 161.43], [660.39, 161.41]]}, {\"category\": \"cyclist\", \"corners_3d\": [[551.18, 210.83], [535.94, 210.8], [541.39, 209.08], [555.76, 209.1], [551.18, 167.41], [535.94, 167.42], [541.39, 168.17], [555.76, 168.16]]}]\n```" - } - ] - }, - { - "id": 1864, - "data_source": "Mono3DRefer", - "scene_name": "001570", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[105.93, 340.62], [40.55, 340.63], [105.3, 322.21], [163.16, 322.2], [105.93, 166.03], [40.55, 166.03], [105.3, 167.7], [163.16, 167.7]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[703.79, 221.28], [722.1, 221.27], [727.7, 223.1], [708.57, 223.11], [703.79, 158.3], [722.1, 158.3], [727.7, 157.31], [708.57, 157.3]]}]\n```" - } - ] - }, - { - "id": 1865, - "data_source": "Mono3DRefer", - "scene_name": "001570", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1342.55, 298.84], [1554.28, 332.45], [1086.06, 333.47], [977.33, 299.46], [1342.55, 171.59], [1554.28, 169.06], [1086.06, 168.99], [977.33, 171.55]]}, {\"category\": \"car\", \"corners_3d\": [[835.1, 208.32], [850.15, 210.19], [721.11, 210.13], [714.2, 208.27], [835.1, 169.81], [850.15, 169.09], [721.11, 169.11], [714.2, 169.83]]}]\n```" - } - ] - }, - { - "id": 1866, - "data_source": "Mono3DRefer", - "scene_name": "001570", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[105.93, 340.62], [40.55, 340.63], [105.3, 322.21], [163.16, 322.2], [105.93, 166.03], [40.55, 166.03], [105.3, 167.7], [163.16, 167.7]]}, {\"category\": \"cyclist\", \"corners_3d\": [[652.32, 242.38], [677.54, 242.32], [688.55, 249.55], [660.39, 249.63], [652.32, 163.41], [677.54, 163.43], [688.55, 161.43], [660.39, 161.41]]}, {\"category\": \"cyclist\", \"corners_3d\": [[551.18, 210.83], [535.94, 210.8], [541.39, 209.08], [555.76, 209.1], [551.18, 167.41], [535.94, 167.42], [541.39, 168.17], [555.76, 168.16]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[703.79, 221.28], [722.1, 221.27], [727.7, 223.1], [708.57, 223.11], [703.79, 158.3], [722.1, 158.3], [727.7, 157.31], [708.57, 157.3]]}]\n```" - } - ] - }, - { - "id": 1867, - "data_source": "Mono3DRefer", - "scene_name": "001570", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[652.32, 242.38], [677.54, 242.32], [688.55, 249.55], [660.39, 249.63], [652.32, 163.41], [677.54, 163.43], [688.55, 161.43], [660.39, 161.41]]}, {\"category\": \"car\", \"corners_3d\": [[1342.55, 298.84], [1554.28, 332.45], [1086.06, 333.47], [977.33, 299.46], [1342.55, 171.59], [1554.28, 169.06], [1086.06, 168.99], [977.33, 171.55]]}, {\"category\": \"cyclist\", \"corners_3d\": [[551.18, 210.83], [535.94, 210.8], [541.39, 209.08], [555.76, 209.1], [551.18, 167.41], [535.94, 167.42], [541.39, 168.17], [555.76, 168.16]]}, {\"category\": \"car\", \"corners_3d\": [[835.1, 208.32], [850.15, 210.19], [721.11, 210.13], [714.2, 208.27], [835.1, 169.81], [850.15, 169.09], [721.11, 169.11], [714.2, 169.83]]}]\n```" - } - ] - }, - { - "id": 1868, - "data_source": "Mono3DRefer", - "scene_name": "001570", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[105.93, 340.62], [40.55, 340.63], [105.3, 322.21], [163.16, 322.2], [105.93, 166.03], [40.55, 166.03], [105.3, 167.7], [163.16, 167.7]]}, {\"category\": \"car\", \"corners_3d\": [[1342.55, 298.84], [1554.28, 332.45], [1086.06, 333.47], [977.33, 299.46], [1342.55, 171.59], [1554.28, 169.06], [1086.06, 168.99], [977.33, 171.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[703.79, 221.28], [722.1, 221.27], [727.7, 223.1], [708.57, 223.11], [703.79, 158.3], [722.1, 158.3], [727.7, 157.31], [708.57, 157.3]]}, {\"category\": \"car\", \"corners_3d\": [[835.1, 208.32], [850.15, 210.19], [721.11, 210.13], [714.2, 208.27], [835.1, 169.81], [850.15, 169.09], [721.11, 169.11], [714.2, 169.83]]}]\n```" - } - ] - }, - { - "id": 1869, - "data_source": "Mono3DRefer", - "scene_name": "001570", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[105.93, 340.62], [40.55, 340.63], [105.3, 322.21], [163.16, 322.2], [105.93, 166.03], [40.55, 166.03], [105.3, 167.7], [163.16, 167.7]]}, {\"category\": \"cyclist\", \"corners_3d\": [[652.32, 242.38], [677.54, 242.32], [688.55, 249.55], [660.39, 249.63], [652.32, 163.41], [677.54, 163.43], [688.55, 161.43], [660.39, 161.41]]}, {\"category\": \"car\", \"corners_3d\": [[1342.55, 298.84], [1554.28, 332.45], [1086.06, 333.47], [977.33, 299.46], [1342.55, 171.59], [1554.28, 169.06], [1086.06, 168.99], [977.33, 171.55]]}, {\"category\": \"cyclist\", \"corners_3d\": [[551.18, 210.83], [535.94, 210.8], [541.39, 209.08], [555.76, 209.1], [551.18, 167.41], [535.94, 167.42], [541.39, 168.17], [555.76, 168.16]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[703.79, 221.28], [722.1, 221.27], [727.7, 223.1], [708.57, 223.11], [703.79, 158.3], [722.1, 158.3], [727.7, 157.31], [708.57, 157.3]]}, {\"category\": \"car\", \"corners_3d\": [[835.1, 208.32], [850.15, 210.19], [721.11, 210.13], [714.2, 208.27], [835.1, 169.81], [850.15, 169.09], [721.11, 169.11], [714.2, 169.83]]}]\n```" - } - ] - }, - { - "id": 1870, - "data_source": "Mono3DRefer", - "scene_name": "001574", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[595.7, 314.36], [599.67, 323.96], [469.32, 326.34], [474.14, 316.44], [595.7, 158.94], [599.67, 157.39], [469.32, 157.01], [474.14, 158.6]]}]\n```" - } - ] - }, - { - "id": 1871, - "data_source": "Mono3DRefer", - "scene_name": "001575", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[531.49, 228.78], [492.04, 229.15], [495.47, 222.71], [530.41, 222.42], [531.49, 188.62], [492.04, 188.73], [495.47, 186.91], [530.41, 186.83]]}, {\"category\": \"car\", \"corners_3d\": [[505.8, 213.85], [477.86, 214.31], [470.48, 210.68], [496.09, 210.3], [505.8, 189.93], [477.86, 190.12], [470.48, 188.61], [496.09, 188.45]]}]\n```" - } - ] - }, - { - "id": 1872, - "data_source": "Mono3DRefer", - "scene_name": "001577", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[375.87, 203.96], [359.97, 203.4], [409.71, 201.27], [425.32, 201.75], [375.87, 180.59], [359.97, 180.45], [409.71, 179.92], [425.32, 180.04]]}, {\"category\": \"car\", \"corners_3d\": [[495.85, 194.7], [491.51, 194.13], [536.35, 193.65], [541.76, 194.19], [495.85, 172.73], [491.51, 172.73], [536.35, 172.73], [541.76, 172.73]]}, {\"category\": \"car\", \"corners_3d\": [[446.92, 197.16], [443.71, 196.55], [489.96, 195.99], [494.26, 196.58], [446.92, 178.93], [443.71, 178.77], [489.96, 178.63], [494.26, 178.78]]}]\n```" - } - ] - }, - { - "id": 1873, - "data_source": "Mono3DRefer", - "scene_name": "001578", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[868.76, 234.86], [814.0, 234.87], [781.85, 225.14], [828.02, 225.14], [868.76, 182.27], [814.0, 182.27], [781.85, 180.79], [828.02, 180.79]]}, {\"category\": \"car\", \"corners_3d\": [[660.34, 206.64], [655.55, 205.06], [740.59, 204.79], [749.48, 206.34], [660.34, 174.14], [655.55, 174.08], [740.59, 174.07], [749.48, 174.13]]}, {\"category\": \"car\", \"corners_3d\": [[578.07, 200.79], [603.85, 200.78], [603.95, 203.55], [575.61, 203.56], [578.07, 177.01], [603.85, 177.0], [603.95, 177.42], [575.61, 177.42]]}, {\"category\": \"car\", \"corners_3d\": [[1050.31, 334.16], [896.37, 333.51], [799.06, 276.23], [897.72, 276.5], [1050.31, 189.9], [896.37, 189.83], [799.06, 183.78], [897.72, 183.81]]}]\n```" - } - ] - }, - { - "id": 1874, - "data_source": "Mono3DRefer", - "scene_name": "001580", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[583.1, 221.9], [632.73, 221.86], [637.48, 229.68], [579.94, 229.72], [583.1, 181.95], [632.73, 181.94], [637.48, 183.39], [579.94, 183.4]]}, {\"category\": \"car\", \"corners_3d\": [[-513.37, 414.52], [-192.05, 357.56], [116.59, 399.35], [-181.53, 491.43], [-513.37, 175.68], [-192.05, 175.01], [116.59, 175.5], [-181.53, 176.58]]}, {\"category\": \"car\", \"corners_3d\": [[272.65, 245.88], [333.34, 240.62], [440.06, 250.23], [379.56, 257.16], [272.65, 184.3], [333.34, 183.48], [440.06, 184.99], [379.56, 186.07]]}, {\"category\": \"car\", \"corners_3d\": [[341.13, 237.03], [389.53, 233.29], [474.95, 240.59], [426.26, 245.33], [341.13, 187.99], [389.53, 187.1], [474.95, 188.83], [426.26, 189.94]]}, {\"category\": \"car\", \"corners_3d\": [[488.13, 222.0], [451.5, 224.16], [401.38, 219.66], [436.85, 217.86], [488.13, 181.81], [451.5, 182.2], [401.38, 181.38], [436.85, 181.05]]}, {\"category\": \"car\", \"corners_3d\": [[519.88, 220.65], [566.62, 220.66], [559.75, 228.2], [505.65, 228.19], [519.88, 187.08], [566.62, 187.08], [559.75, 189.32], [505.65, 189.32]]}, {\"category\": \"car\", \"corners_3d\": [[670.23, 225.83], [724.62, 225.75], [751.67, 236.85], [685.89, 236.96], [670.23, 179.51], [724.62, 179.5], [751.67, 180.89], [685.89, 180.9]]}, {\"category\": \"car\", \"corners_3d\": [[631.07, 198.53], [652.77, 198.51], [657.6, 200.43], [634.27, 200.46], [631.07, 179.16], [652.77, 179.16], [657.6, 179.63], [634.27, 179.64]]}, {\"category\": \"car\", \"corners_3d\": [[185.06, 260.56], [105.22, 260.66], [199.97, 243.91], [264.5, 243.84], [185.06, 193.27], [105.22, 193.3], [199.97, 189.39], [264.5, 189.38]]}, {\"category\": \"car\", \"corners_3d\": [[591.2, 209.9], [625.96, 209.89], [628.64, 214.19], [589.84, 214.21], [591.2, 178.26], [625.96, 178.26], [628.64, 178.88], [589.84, 178.89]]}, {\"category\": \"car\", \"corners_3d\": [[138.19, 231.26], [88.52, 231.42], [153.53, 223.67], [196.49, 223.55], [138.19, 186.69], [88.52, 186.72], [153.53, 184.89], [196.49, 184.86]]}]\n```" - } - ] - }, - { - "id": 1875, - "data_source": "Mono3DRefer", - "scene_name": "001580", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[697.43, 247.87], [789.28, 247.68], [891.27, 287.11], [751.22, 287.54], [697.43, 137.82], [789.28, 137.91], [891.27, 119.5], [751.22, 119.3]]}, {\"category\": \"van\", \"corners_3d\": [[288.88, 240.83], [213.56, 240.76], [293.47, 227.28], [353.9, 227.33], [288.88, 159.71], [213.56, 159.72], [293.47, 162.33], [353.9, 162.32]]}]\n```" - } - ] - }, - { - "id": 1876, - "data_source": "Mono3DRefer", - "scene_name": "001580", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[28.37, 233.59], [7.72, 233.65], [47.66, 229.46], [66.85, 229.41], [28.37, 183.92], [7.72, 183.93], [47.66, 183.16], [66.85, 183.16]]}]\n```" - } - ] - }, - { - "id": 1877, - "data_source": "Mono3DRefer", - "scene_name": "001580", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[583.1, 221.9], [632.73, 221.86], [637.48, 229.68], [579.94, 229.72], [583.1, 181.95], [632.73, 181.94], [637.48, 183.39], [579.94, 183.4]]}, {\"category\": \"car\", \"corners_3d\": [[-513.37, 414.52], [-192.05, 357.56], [116.59, 399.35], [-181.53, 491.43], [-513.37, 175.68], [-192.05, 175.01], [116.59, 175.5], [-181.53, 176.58]]}, {\"category\": \"car\", \"corners_3d\": [[272.65, 245.88], [333.34, 240.62], [440.06, 250.23], [379.56, 257.16], [272.65, 184.3], [333.34, 183.48], [440.06, 184.99], [379.56, 186.07]]}, {\"category\": \"car\", \"corners_3d\": [[341.13, 237.03], [389.53, 233.29], [474.95, 240.59], [426.26, 245.33], [341.13, 187.99], [389.53, 187.1], [474.95, 188.83], [426.26, 189.94]]}, {\"category\": \"car\", \"corners_3d\": [[488.13, 222.0], [451.5, 224.16], [401.38, 219.66], [436.85, 217.86], [488.13, 181.81], [451.5, 182.2], [401.38, 181.38], [436.85, 181.05]]}, {\"category\": \"car\", \"corners_3d\": [[519.88, 220.65], [566.62, 220.66], [559.75, 228.2], [505.65, 228.19], [519.88, 187.08], [566.62, 187.08], [559.75, 189.32], [505.65, 189.32]]}, {\"category\": \"van\", \"corners_3d\": [[697.43, 247.87], [789.28, 247.68], [891.27, 287.11], [751.22, 287.54], [697.43, 137.82], [789.28, 137.91], [891.27, 119.5], [751.22, 119.3]]}, {\"category\": \"car\", \"corners_3d\": [[670.23, 225.83], [724.62, 225.75], [751.67, 236.85], [685.89, 236.96], [670.23, 179.51], [724.62, 179.5], [751.67, 180.89], [685.89, 180.9]]}, {\"category\": \"car\", \"corners_3d\": [[631.07, 198.53], [652.77, 198.51], [657.6, 200.43], [634.27, 200.46], [631.07, 179.16], [652.77, 179.16], [657.6, 179.63], [634.27, 179.64]]}, {\"category\": \"car\", \"corners_3d\": [[185.06, 260.56], [105.22, 260.66], [199.97, 243.91], [264.5, 243.84], [185.06, 193.27], [105.22, 193.3], [199.97, 189.39], [264.5, 189.38]]}, {\"category\": \"van\", \"corners_3d\": [[288.88, 240.83], [213.56, 240.76], [293.47, 227.28], [353.9, 227.33], [288.88, 159.71], [213.56, 159.72], [293.47, 162.33], [353.9, 162.32]]}, {\"category\": \"car\", \"corners_3d\": [[591.2, 209.9], [625.96, 209.89], [628.64, 214.19], [589.84, 214.21], [591.2, 178.26], [625.96, 178.26], [628.64, 178.88], [589.84, 178.89]]}, {\"category\": \"car\", \"corners_3d\": [[138.19, 231.26], [88.52, 231.42], [153.53, 223.67], [196.49, 223.55], [138.19, 186.69], [88.52, 186.72], [153.53, 184.89], [196.49, 184.86]]}]\n```" - } - ] - }, - { - "id": 1878, - "data_source": "Mono3DRefer", - "scene_name": "001580", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[583.1, 221.9], [632.73, 221.86], [637.48, 229.68], [579.94, 229.72], [583.1, 181.95], [632.73, 181.94], [637.48, 183.39], [579.94, 183.4]]}, {\"category\": \"car\", \"corners_3d\": [[-513.37, 414.52], [-192.05, 357.56], [116.59, 399.35], [-181.53, 491.43], [-513.37, 175.68], [-192.05, 175.01], [116.59, 175.5], [-181.53, 176.58]]}, {\"category\": \"car\", \"corners_3d\": [[272.65, 245.88], [333.34, 240.62], [440.06, 250.23], [379.56, 257.16], [272.65, 184.3], [333.34, 183.48], [440.06, 184.99], [379.56, 186.07]]}, {\"category\": \"car\", \"corners_3d\": [[341.13, 237.03], [389.53, 233.29], [474.95, 240.59], [426.26, 245.33], [341.13, 187.99], [389.53, 187.1], [474.95, 188.83], [426.26, 189.94]]}, {\"category\": \"car\", \"corners_3d\": [[488.13, 222.0], [451.5, 224.16], [401.38, 219.66], [436.85, 217.86], [488.13, 181.81], [451.5, 182.2], [401.38, 181.38], [436.85, 181.05]]}, {\"category\": \"car\", \"corners_3d\": [[519.88, 220.65], [566.62, 220.66], [559.75, 228.2], [505.65, 228.19], [519.88, 187.08], [566.62, 187.08], [559.75, 189.32], [505.65, 189.32]]}, {\"category\": \"car\", \"corners_3d\": [[670.23, 225.83], [724.62, 225.75], [751.67, 236.85], [685.89, 236.96], [670.23, 179.51], [724.62, 179.5], [751.67, 180.89], [685.89, 180.9]]}, {\"category\": \"car\", \"corners_3d\": [[631.07, 198.53], [652.77, 198.51], [657.6, 200.43], [634.27, 200.46], [631.07, 179.16], [652.77, 179.16], [657.6, 179.63], [634.27, 179.64]]}, {\"category\": \"car\", \"corners_3d\": [[185.06, 260.56], [105.22, 260.66], [199.97, 243.91], [264.5, 243.84], [185.06, 193.27], [105.22, 193.3], [199.97, 189.39], [264.5, 189.38]]}, {\"category\": \"car\", \"corners_3d\": [[591.2, 209.9], [625.96, 209.89], [628.64, 214.19], [589.84, 214.21], [591.2, 178.26], [625.96, 178.26], [628.64, 178.88], [589.84, 178.89]]}, {\"category\": \"car\", \"corners_3d\": [[138.19, 231.26], [88.52, 231.42], [153.53, 223.67], [196.49, 223.55], [138.19, 186.69], [88.52, 186.72], [153.53, 184.89], [196.49, 184.86]]}, {\"category\": \"cyclist\", \"corners_3d\": [[28.37, 233.59], [7.72, 233.65], [47.66, 229.46], [66.85, 229.41], [28.37, 183.92], [7.72, 183.93], [47.66, 183.16], [66.85, 183.16]]}]\n```" - } - ] - }, - { - "id": 1879, - "data_source": "Mono3DRefer", - "scene_name": "001580", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[697.43, 247.87], [789.28, 247.68], [891.27, 287.11], [751.22, 287.54], [697.43, 137.82], [789.28, 137.91], [891.27, 119.5], [751.22, 119.3]]}, {\"category\": \"van\", \"corners_3d\": [[288.88, 240.83], [213.56, 240.76], [293.47, 227.28], [353.9, 227.33], [288.88, 159.71], [213.56, 159.72], [293.47, 162.33], [353.9, 162.32]]}, {\"category\": \"cyclist\", \"corners_3d\": [[28.37, 233.59], [7.72, 233.65], [47.66, 229.46], [66.85, 229.41], [28.37, 183.92], [7.72, 183.93], [47.66, 183.16], [66.85, 183.16]]}]\n```" - } - ] - }, - { - "id": 1880, - "data_source": "Mono3DRefer", - "scene_name": "001580", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[583.1, 221.9], [632.73, 221.86], [637.48, 229.68], [579.94, 229.72], [583.1, 181.95], [632.73, 181.94], [637.48, 183.39], [579.94, 183.4]]}, {\"category\": \"car\", \"corners_3d\": [[-513.37, 414.52], [-192.05, 357.56], [116.59, 399.35], [-181.53, 491.43], [-513.37, 175.68], [-192.05, 175.01], [116.59, 175.5], [-181.53, 176.58]]}, {\"category\": \"car\", \"corners_3d\": [[272.65, 245.88], [333.34, 240.62], [440.06, 250.23], [379.56, 257.16], [272.65, 184.3], [333.34, 183.48], [440.06, 184.99], [379.56, 186.07]]}, {\"category\": \"car\", \"corners_3d\": [[341.13, 237.03], [389.53, 233.29], [474.95, 240.59], [426.26, 245.33], [341.13, 187.99], [389.53, 187.1], [474.95, 188.83], [426.26, 189.94]]}, {\"category\": \"car\", \"corners_3d\": [[488.13, 222.0], [451.5, 224.16], [401.38, 219.66], [436.85, 217.86], [488.13, 181.81], [451.5, 182.2], [401.38, 181.38], [436.85, 181.05]]}, {\"category\": \"car\", \"corners_3d\": [[519.88, 220.65], [566.62, 220.66], [559.75, 228.2], [505.65, 228.19], [519.88, 187.08], [566.62, 187.08], [559.75, 189.32], [505.65, 189.32]]}, {\"category\": \"van\", \"corners_3d\": [[697.43, 247.87], [789.28, 247.68], [891.27, 287.11], [751.22, 287.54], [697.43, 137.82], [789.28, 137.91], [891.27, 119.5], [751.22, 119.3]]}, {\"category\": \"car\", \"corners_3d\": [[670.23, 225.83], [724.62, 225.75], [751.67, 236.85], [685.89, 236.96], [670.23, 179.51], [724.62, 179.5], [751.67, 180.89], [685.89, 180.9]]}, {\"category\": \"car\", \"corners_3d\": [[631.07, 198.53], [652.77, 198.51], [657.6, 200.43], [634.27, 200.46], [631.07, 179.16], [652.77, 179.16], [657.6, 179.63], [634.27, 179.64]]}, {\"category\": \"car\", \"corners_3d\": [[185.06, 260.56], [105.22, 260.66], [199.97, 243.91], [264.5, 243.84], [185.06, 193.27], [105.22, 193.3], [199.97, 189.39], [264.5, 189.38]]}, {\"category\": \"van\", \"corners_3d\": [[288.88, 240.83], [213.56, 240.76], [293.47, 227.28], [353.9, 227.33], [288.88, 159.71], [213.56, 159.72], [293.47, 162.33], [353.9, 162.32]]}, {\"category\": \"car\", \"corners_3d\": [[591.2, 209.9], [625.96, 209.89], [628.64, 214.19], [589.84, 214.21], [591.2, 178.26], [625.96, 178.26], [628.64, 178.88], [589.84, 178.89]]}, {\"category\": \"car\", \"corners_3d\": [[138.19, 231.26], [88.52, 231.42], [153.53, 223.67], [196.49, 223.55], [138.19, 186.69], [88.52, 186.72], [153.53, 184.89], [196.49, 184.86]]}, {\"category\": \"cyclist\", \"corners_3d\": [[28.37, 233.59], [7.72, 233.65], [47.66, 229.46], [66.85, 229.41], [28.37, 183.92], [7.72, 183.93], [47.66, 183.16], [66.85, 183.16]]}]\n```" - } - ] - }, - { - "id": 1881, - "data_source": "Mono3DRefer", - "scene_name": "001582", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[397.76, 353.41], [188.9, 354.49], [346.42, 283.96], [473.9, 283.56], [397.76, 145.89], [188.9, 145.73], [346.42, 156.26], [473.9, 156.32]]}, {\"category\": \"van\", \"corners_3d\": [[748.94, 296.31], [867.64, 296.12], [1030.0, 371.72], [838.71, 372.21], [748.94, 166.98], [867.64, 166.99], [1030.0, 163.4], [838.71, 163.37]]}, {\"category\": \"van\", \"corners_3d\": [[553.42, 228.69], [494.8, 228.7], [514.45, 219.09], [562.98, 219.08], [553.42, 164.91], [494.8, 164.91], [514.45, 166.28], [562.98, 166.28]]}]\n```" - } - ] - }, - { - "id": 1882, - "data_source": "Mono3DRefer", - "scene_name": "001582", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[533.68, 254.76], [454.5, 255.12], [483.46, 236.16], [544.34, 235.94], [533.68, 184.4], [454.5, 184.45], [483.46, 181.78], [544.34, 181.75]]}, {\"category\": \"car\", \"corners_3d\": [[569.52, 216.41], [520.09, 216.15], [542.28, 209.91], [584.67, 210.1], [569.52, 169.54], [520.09, 169.56], [542.28, 170.03], [584.67, 170.02]]}, {\"category\": \"car\", \"corners_3d\": [[651.45, 198.44], [671.7, 199.11], [641.25, 200.34], [620.89, 199.6], [651.45, 172.01], [671.7, 171.98], [641.25, 171.94], [620.89, 171.97]]}]\n```" - } - ] - }, - { - "id": 1883, - "data_source": "Mono3DRefer", - "scene_name": "001582", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[397.76, 353.41], [188.9, 354.49], [346.42, 283.96], [473.9, 283.56], [397.76, 145.89], [188.9, 145.73], [346.42, 156.26], [473.9, 156.32]]}, {\"category\": \"van\", \"corners_3d\": [[748.94, 296.31], [867.64, 296.12], [1030.0, 371.72], [838.71, 372.21], [748.94, 166.98], [867.64, 166.99], [1030.0, 163.4], [838.71, 163.37]]}, {\"category\": \"car\", \"corners_3d\": [[533.68, 254.76], [454.5, 255.12], [483.46, 236.16], [544.34, 235.94], [533.68, 184.4], [454.5, 184.45], [483.46, 181.78], [544.34, 181.75]]}, {\"category\": \"van\", \"corners_3d\": [[553.42, 228.69], [494.8, 228.7], [514.45, 219.09], [562.98, 219.08], [553.42, 164.91], [494.8, 164.91], [514.45, 166.28], [562.98, 166.28]]}, {\"category\": \"car\", \"corners_3d\": [[569.52, 216.41], [520.09, 216.15], [542.28, 209.91], [584.67, 210.1], [569.52, 169.54], [520.09, 169.56], [542.28, 170.03], [584.67, 170.02]]}, {\"category\": \"car\", \"corners_3d\": [[651.45, 198.44], [671.7, 199.11], [641.25, 200.34], [620.89, 199.6], [651.45, 172.01], [671.7, 171.98], [641.25, 171.94], [620.89, 171.97]]}]\n```" - } - ] - }, - { - "id": 1884, - "data_source": "Mono3DRefer", - "scene_name": "001583", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[36.88, 285.42], [115.93, 270.48], [323.63, 271.38], [275.88, 286.63], [36.88, 180.52], [115.93, 179.5], [323.63, 179.57], [275.88, 180.6]]}, {\"category\": \"car\", \"corners_3d\": [[431.26, 247.06], [408.8, 255.4], [186.94, 254.63], [231.42, 246.44], [431.26, 160.39], [408.8, 158.99], [186.94, 159.12], [231.42, 160.5]]}, {\"category\": \"car\", \"corners_3d\": [[154.23, 270.68], [206.4, 259.68], [375.83, 259.91], [345.05, 270.98], [154.23, 174.57], [206.4, 174.38], [375.83, 174.38], [345.05, 174.58]]}, {\"category\": \"car\", \"corners_3d\": [[701.23, 244.66], [775.95, 244.36], [825.76, 262.66], [732.09, 263.12], [701.23, 179.2], [775.95, 179.18], [825.76, 180.79], [732.09, 180.83]]}, {\"category\": \"car\", \"corners_3d\": [[254.71, 242.73], [285.62, 236.66], [422.94, 236.68], [405.08, 242.75], [254.71, 163.04], [285.62, 163.89], [422.94, 163.89], [405.08, 163.03]]}, {\"category\": \"car\", \"corners_3d\": [[457.46, 232.35], [437.21, 237.24], [311.36, 235.73], [340.52, 231.06], [457.46, 176.03], [437.21, 176.29], [311.36, 176.21], [340.52, 175.96]]}, {\"category\": \"car\", \"corners_3d\": [[316.84, 228.76], [339.52, 224.83], [471.53, 225.23], [458.73, 229.22], [316.84, 171.85], [339.52, 171.92], [471.53, 171.91], [458.73, 171.84]]}, {\"category\": \"car\", \"corners_3d\": [[506.81, 208.64], [498.77, 210.52], [400.11, 210.16], [412.94, 208.31], [506.81, 171.91], [498.77, 171.86], [400.11, 171.87], [412.94, 171.92]]}, {\"category\": \"car\", \"corners_3d\": [[434.03, 204.61], [442.18, 203.2], [519.86, 203.24], [515.31, 204.65], [434.03, 164.85], [442.18, 165.21], [519.86, 165.2], [515.31, 164.84]]}, {\"category\": \"car\", \"corners_3d\": [[649.99, 206.38], [684.1, 206.35], [694.67, 210.36], [656.49, 210.4], [649.99, 175.21], [684.1, 175.21], [694.67, 175.49], [656.49, 175.49]]}, {\"category\": \"car\", \"corners_3d\": [[678.48, 227.21], [734.18, 227.21], [759.95, 238.5], [692.67, 238.49], [678.48, 175.84], [734.18, 175.84], [759.95, 176.46], [692.67, 176.46]]}, {\"category\": \"car\", \"corners_3d\": [[638.21, 197.41], [666.51, 197.41], [671.57, 199.61], [640.73, 199.61], [638.21, 170.41], [666.51, 170.41], [671.57, 170.19], [640.73, 170.19]]}, {\"category\": \"car\", \"corners_3d\": [[736.76, 278.8], [847.96, 278.81], [1001.3, 347.13], [818.39, 347.09], [736.76, 164.38], [847.96, 164.38], [1001.3, 158.91], [818.39, 158.92]]}, {\"category\": \"car\", \"corners_3d\": [[186.44, 347.15], [56.61, 403.51], [-586.54, 409.54], [-293.92, 350.57], [186.44, 190.35], [56.61, 196.01], [-586.54, 196.62], [-293.92, 190.7]]}, {\"category\": \"car\", \"corners_3d\": [[-155.05, 331.96], [-17.98, 304.76], [271.3, 307.62], [191.85, 336.14], [-155.05, 194.1], [-17.98, 190.47], [271.3, 190.85], [191.85, 194.66]]}]\n```" - } - ] - }, - { - "id": 1885, - "data_source": "Mono3DRefer", - "scene_name": "001584", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[786.35, 321.42], [937.25, 322.07], [1223.93, 458.87], [933.47, 456.52], [786.35, 192.88], [937.25, 192.96], [1223.93, 211.4], [933.47, 211.09]]}, {\"category\": \"car\", \"corners_3d\": [[520.7, 220.75], [477.32, 220.72], [493.92, 215.0], [532.12, 215.02], [520.7, 185.44], [477.32, 185.43], [493.92, 183.93], [532.12, 183.93]]}, {\"category\": \"car\", \"corners_3d\": [[266.93, 336.01], [133.25, 335.43], [287.13, 284.39], [379.13, 284.66], [266.93, 210.68], [133.25, 210.54], [287.13, 198.71], [379.13, 198.77]]}, {\"category\": \"car\", \"corners_3d\": [[396.64, 274.83], [309.51, 274.34], [382.74, 251.71], [450.64, 252.0], [396.64, 195.8], [309.51, 195.69], [382.74, 190.6], [450.64, 190.66]]}]\n```" - } - ] - }, - { - "id": 1886, - "data_source": "Mono3DRefer", - "scene_name": "001588", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[308.09, 208.61], [298.95, 208.54], [312.71, 207.33], [321.56, 207.39], [308.09, 180.72], [298.95, 180.71], [312.71, 180.44], [321.56, 180.46]]}]\n```" - } - ] - }, - { - "id": 1887, - "data_source": "Mono3DRefer", - "scene_name": "001588", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-57.12, 256.2], [-134.52, 256.11], [-6.82, 241.95], [57.51, 242.01], [-57.12, 196.28], [-134.52, 196.26], [-6.82, 192.28], [57.51, 192.29]]}, {\"category\": \"car\", \"corners_3d\": [[344.11, 212.07], [314.33, 211.93], [341.52, 208.92], [369.08, 209.03], [344.11, 186.33], [314.33, 186.29], [341.52, 185.25], [369.08, 185.29]]}]\n```" - } - ] - }, - { - "id": 1888, - "data_source": "Mono3DRefer", - "scene_name": "001588", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[358.32, 202.03], [332.94, 202.01], [357.42, 199.55], [380.68, 199.56], [358.32, 172.98], [332.94, 172.98], [357.42, 172.97], [380.68, 172.97]]}]\n```" - } - ] - }, - { - "id": 1889, - "data_source": "Mono3DRefer", - "scene_name": "001588", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-57.12, 256.2], [-134.52, 256.11], [-6.82, 241.95], [57.51, 242.01], [-57.12, 196.28], [-134.52, 196.26], [-6.82, 192.28], [57.51, 192.29]]}, {\"category\": \"cyclist\", \"corners_3d\": [[308.09, 208.61], [298.95, 208.54], [312.71, 207.33], [321.56, 207.39], [308.09, 180.72], [298.95, 180.71], [312.71, 180.44], [321.56, 180.46]]}, {\"category\": \"car\", \"corners_3d\": [[344.11, 212.07], [314.33, 211.93], [341.52, 208.92], [369.08, 209.03], [344.11, 186.33], [314.33, 186.29], [341.52, 185.25], [369.08, 185.29]]}]\n```" - } - ] - }, - { - "id": 1890, - "data_source": "Mono3DRefer", - "scene_name": "001588", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[308.09, 208.61], [298.95, 208.54], [312.71, 207.33], [321.56, 207.39], [308.09, 180.72], [298.95, 180.71], [312.71, 180.44], [321.56, 180.46]]}, {\"category\": \"van\", \"corners_3d\": [[358.32, 202.03], [332.94, 202.01], [357.42, 199.55], [380.68, 199.56], [358.32, 172.98], [332.94, 172.98], [357.42, 172.97], [380.68, 172.97]]}]\n```" - } - ] - }, - { - "id": 1891, - "data_source": "Mono3DRefer", - "scene_name": "001588", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-57.12, 256.2], [-134.52, 256.11], [-6.82, 241.95], [57.51, 242.01], [-57.12, 196.28], [-134.52, 196.26], [-6.82, 192.28], [57.51, 192.29]]}, {\"category\": \"car\", \"corners_3d\": [[344.11, 212.07], [314.33, 211.93], [341.52, 208.92], [369.08, 209.03], [344.11, 186.33], [314.33, 186.29], [341.52, 185.25], [369.08, 185.29]]}, {\"category\": \"van\", \"corners_3d\": [[358.32, 202.03], [332.94, 202.01], [357.42, 199.55], [380.68, 199.56], [358.32, 172.98], [332.94, 172.98], [357.42, 172.97], [380.68, 172.97]]}]\n```" - } - ] - }, - { - "id": 1892, - "data_source": "Mono3DRefer", - "scene_name": "001588", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-57.12, 256.2], [-134.52, 256.11], [-6.82, 241.95], [57.51, 242.01], [-57.12, 196.28], [-134.52, 196.26], [-6.82, 192.28], [57.51, 192.29]]}, {\"category\": \"cyclist\", \"corners_3d\": [[308.09, 208.61], [298.95, 208.54], [312.71, 207.33], [321.56, 207.39], [308.09, 180.72], [298.95, 180.71], [312.71, 180.44], [321.56, 180.46]]}, {\"category\": \"car\", \"corners_3d\": [[344.11, 212.07], [314.33, 211.93], [341.52, 208.92], [369.08, 209.03], [344.11, 186.33], [314.33, 186.29], [341.52, 185.25], [369.08, 185.29]]}, {\"category\": \"van\", \"corners_3d\": [[358.32, 202.03], [332.94, 202.01], [357.42, 199.55], [380.68, 199.56], [358.32, 172.98], [332.94, 172.98], [357.42, 172.97], [380.68, 172.97]]}]\n```" - } - ] - }, - { - "id": 1893, - "data_source": "Mono3DRefer", - "scene_name": "001593", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-13.42, 812.58], [-610.2, 813.0], [154.66, 411.4], [376.95, 411.35], [-13.42, 286.07], [-610.2, 286.14], [154.66, 215.07], [376.95, 215.06]]}, {\"category\": \"car\", \"corners_3d\": [[906.67, 440.3], [1131.19, 434.37], [2322.56, 979.39], [1683.28, 1038.68], [906.67, 203.37], [1131.19, 202.69], [2322.56, 264.88], [1683.28, 271.65]]}, {\"category\": \"car\", \"corners_3d\": [[733.49, 303.07], [847.39, 303.08], [989.06, 380.85], [807.13, 380.81], [733.49, 197.74], [847.39, 197.74], [989.06, 212.6], [807.13, 212.59]]}, {\"category\": \"car\", \"corners_3d\": [[535.72, 262.94], [470.85, 262.54], [501.35, 246.84], [554.93, 247.11], [535.72, 196.95], [470.85, 196.84], [501.35, 192.64], [554.93, 192.72]]}, {\"category\": \"car\", \"corners_3d\": [[573.49, 241.52], [520.01, 241.32], [538.72, 230.45], [583.73, 230.59], [573.49, 179.65], [520.01, 179.63], [538.72, 178.55], [583.73, 178.56]]}, {\"category\": \"car\", \"corners_3d\": [[602.98, 227.63], [560.08, 227.37], [574.06, 220.24], [611.37, 220.43], [602.98, 180.78], [560.08, 180.74], [574.06, 179.71], [611.37, 179.74]]}, {\"category\": \"car\", \"corners_3d\": [[646.2, 216.83], [620.26, 216.37], [636.69, 212.86], [660.68, 213.25], [646.2, 191.86], [620.26, 191.66], [636.69, 190.15], [660.68, 190.32]]}]\n```" - } - ] - }, - { - "id": 1894, - "data_source": "Mono3DRefer", - "scene_name": "001598", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[551.5, 194.82], [530.06, 194.82], [535.11, 193.28], [555.04, 193.27], [551.5, 172.16], [530.06, 172.16], [535.11, 172.21], [555.04, 172.21]]}]\n```" - } - ] - }, - { - "id": 1895, - "data_source": "Mono3DRefer", - "scene_name": "001604", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[595.04, 212.4], [632.77, 212.42], [634.82, 217.67], [592.08, 217.64], [595.04, 176.73], [632.77, 176.74], [634.82, 177.25], [592.08, 177.25]]}, {\"category\": \"car\", \"corners_3d\": [[354.33, 220.68], [316.16, 220.53], [354.51, 215.09], [388.43, 215.21], [354.33, 185.06], [316.16, 185.02], [354.51, 183.63], [388.43, 183.66]]}, {\"category\": \"car\", \"corners_3d\": [[399.24, 214.71], [362.42, 214.73], [390.33, 209.85], [422.84, 209.83], [399.24, 181.05], [362.42, 181.05], [390.33, 180.1], [422.84, 180.09]]}, {\"category\": \"car\", \"corners_3d\": [[723.14, 206.77], [750.96, 207.97], [700.1, 210.79], [672.01, 209.4], [723.14, 173.92], [750.96, 173.96], [700.1, 174.05], [672.01, 174.0]]}, {\"category\": \"car\", \"corners_3d\": [[718.19, 205.43], [745.17, 206.46], [696.45, 209.36], [668.84, 208.14], [718.19, 176.34], [745.17, 176.45], [696.45, 176.76], [668.84, 176.63]]}, {\"category\": \"car\", \"corners_3d\": [[700.97, 201.57], [724.5, 202.44], [673.85, 204.76], [650.13, 203.75], [700.97, 176.75], [724.5, 176.87], [673.85, 177.18], [650.13, 177.04]]}]\n```" - } - ] - }, - { - "id": 1896, - "data_source": "Mono3DRefer", - "scene_name": "001606", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[610.41, 215.35], [569.82, 215.23], [579.88, 209.57], [615.05, 209.65], [610.41, 178.85], [569.82, 178.83], [579.88, 178.03], [615.05, 178.04]]}, {\"category\": \"car\", \"corners_3d\": [[626.69, 200.47], [594.48, 200.25], [608.77, 197.56], [637.9, 197.75], [626.69, 169.78], [594.48, 169.8], [608.77, 170.1], [637.9, 170.08]]}, {\"category\": \"car\", \"corners_3d\": [[309.79, 235.42], [336.11, 229.94], [496.11, 229.96], [485.12, 235.44], [309.79, 174.9], [336.11, 174.72], [496.11, 174.72], [485.12, 174.9]]}, {\"category\": \"car\", \"corners_3d\": [[2220.61, 910.02], [1527.11, 910.59], [892.64, 400.78], [1107.17, 400.73], [2220.61, 267.56], [1527.11, 267.63], [892.64, 202.14], [1107.17, 202.13]]}]\n```" - } - ] - }, - { - "id": 1897, - "data_source": "Mono3DRefer", - "scene_name": "001606", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[709.73, 227.2], [760.12, 227.47], [778.98, 237.84], [718.97, 237.46], [709.73, 173.15], [760.12, 173.15], [778.98, 173.21], [718.97, 173.2]]}, {\"category\": \"van\", \"corners_3d\": [[617.38, 208.08], [582.32, 207.96], [590.86, 204.08], [622.06, 204.18], [617.38, 169.74], [582.32, 169.75], [590.86, 170.09], [622.06, 170.08]]}, {\"category\": \"van\", \"corners_3d\": [[577.71, 234.92], [512.83, 234.53], [537.19, 224.29], [591.37, 224.56], [577.71, 168.92], [512.83, 168.94], [537.19, 169.59], [591.37, 169.57]]}]\n```" - } - ] - }, - { - "id": 1898, - "data_source": "Mono3DRefer", - "scene_name": "001606", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[709.73, 227.2], [760.12, 227.47], [778.98, 237.84], [718.97, 237.46], [709.73, 173.15], [760.12, 173.15], [778.98, 173.21], [718.97, 173.2]]}, {\"category\": \"car\", \"corners_3d\": [[610.41, 215.35], [569.82, 215.23], [579.88, 209.57], [615.05, 209.65], [610.41, 178.85], [569.82, 178.83], [579.88, 178.03], [615.05, 178.04]]}, {\"category\": \"van\", \"corners_3d\": [[617.38, 208.08], [582.32, 207.96], [590.86, 204.08], [622.06, 204.18], [617.38, 169.74], [582.32, 169.75], [590.86, 170.09], [622.06, 170.08]]}, {\"category\": \"car\", \"corners_3d\": [[626.69, 200.47], [594.48, 200.25], [608.77, 197.56], [637.9, 197.75], [626.69, 169.78], [594.48, 169.8], [608.77, 170.1], [637.9, 170.08]]}, {\"category\": \"car\", \"corners_3d\": [[309.79, 235.42], [336.11, 229.94], [496.11, 229.96], [485.12, 235.44], [309.79, 174.9], [336.11, 174.72], [496.11, 174.72], [485.12, 174.9]]}, {\"category\": \"car\", \"corners_3d\": [[2220.61, 910.02], [1527.11, 910.59], [892.64, 400.78], [1107.17, 400.73], [2220.61, 267.56], [1527.11, 267.63], [892.64, 202.14], [1107.17, 202.13]]}, {\"category\": \"van\", \"corners_3d\": [[577.71, 234.92], [512.83, 234.53], [537.19, 224.29], [591.37, 224.56], [577.71, 168.92], [512.83, 168.94], [537.19, 169.59], [591.37, 169.57]]}]\n```" - } - ] - }, - { - "id": 1899, - "data_source": "Mono3DRefer", - "scene_name": "001609", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[591.63, 318.21], [593.38, 328.74], [490.4, 329.82], [496.01, 319.14], [591.63, 144.94], [593.38, 142.22], [490.4, 141.95], [496.01, 144.7]]}]\n```" - } - ] - }, - { - "id": 1900, - "data_source": "Mono3DRefer", - "scene_name": "001610", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[798.1, 324.18], [957.52, 323.2], [1279.74, 454.47], [983.38, 457.94], [798.1, 183.23], [957.52, 183.16], [1279.74, 192.16], [983.38, 192.39]]}, {\"category\": \"car\", \"corners_3d\": [[340.42, 274.37], [435.97, 273.85], [387.67, 307.43], [259.8, 308.35], [340.42, 201.62], [435.97, 201.47], [387.67, 210.99], [259.8, 211.25]]}, {\"category\": \"car\", \"corners_3d\": [[684.33, 240.07], [746.73, 239.9], [783.52, 255.49], [706.68, 255.75], [684.33, 180.54], [746.73, 180.52], [783.52, 182.3], [706.68, 182.33]]}, {\"category\": \"car\", \"corners_3d\": [[443.55, 235.88], [498.89, 235.89], [474.92, 249.47], [407.65, 249.47], [443.55, 182.99], [498.89, 182.99], [474.92, 185.18], [407.65, 185.18]]}, {\"category\": \"car\", \"corners_3d\": [[523.06, 219.0], [487.14, 219.05], [498.13, 214.32], [530.37, 214.28], [523.06, 184.06], [487.14, 184.07], [498.13, 182.92], [530.37, 182.92]]}, {\"category\": \"car\", \"corners_3d\": [[542.9, 206.59], [518.29, 206.6], [526.57, 203.51], [548.94, 203.51], [542.9, 185.08], [518.29, 185.08], [526.57, 183.96], [548.94, 183.96]]}, {\"category\": \"car\", \"corners_3d\": [[633.16, 192.8], [651.09, 192.82], [651.38, 193.99], [632.39, 193.97], [633.16, 178.11], [651.09, 178.11], [651.38, 178.42], [632.39, 178.41]]}, {\"category\": \"car\", \"corners_3d\": [[648.6, 210.34], [684.26, 210.29], [695.77, 214.78], [655.84, 214.85], [648.6, 175.85], [684.26, 175.84], [695.77, 176.2], [655.84, 176.21]]}]\n```" - } - ] - }, - { - "id": 1901, - "data_source": "Mono3DRefer", - "scene_name": "001610", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[577.06, 200.16], [569.55, 200.16], [571.09, 199.26], [578.36, 199.26], [577.06, 178.1], [569.55, 178.1], [571.09, 177.92], [578.36, 177.92]]}, {\"category\": \"cyclist\", \"corners_3d\": [[588.95, 197.97], [581.13, 197.97], [581.94, 197.24], [589.53, 197.24], [588.95, 177.12], [581.13, 177.12], [581.94, 176.99], [589.53, 176.99]]}]\n```" - } - ] - }, - { - "id": 1902, - "data_source": "Mono3DRefer", - "scene_name": "001610", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[798.1, 324.18], [957.52, 323.2], [1279.74, 454.47], [983.38, 457.94], [798.1, 183.23], [957.52, 183.16], [1279.74, 192.16], [983.38, 192.39]]}, {\"category\": \"car\", \"corners_3d\": [[340.42, 274.37], [435.97, 273.85], [387.67, 307.43], [259.8, 308.35], [340.42, 201.62], [435.97, 201.47], [387.67, 210.99], [259.8, 211.25]]}, {\"category\": \"car\", \"corners_3d\": [[684.33, 240.07], [746.73, 239.9], [783.52, 255.49], [706.68, 255.75], [684.33, 180.54], [746.73, 180.52], [783.52, 182.3], [706.68, 182.33]]}, {\"category\": \"car\", \"corners_3d\": [[443.55, 235.88], [498.89, 235.89], [474.92, 249.47], [407.65, 249.47], [443.55, 182.99], [498.89, 182.99], [474.92, 185.18], [407.65, 185.18]]}, {\"category\": \"car\", \"corners_3d\": [[523.06, 219.0], [487.14, 219.05], [498.13, 214.32], [530.37, 214.28], [523.06, 184.06], [487.14, 184.07], [498.13, 182.92], [530.37, 182.92]]}, {\"category\": \"cyclist\", \"corners_3d\": [[577.06, 200.16], [569.55, 200.16], [571.09, 199.26], [578.36, 199.26], [577.06, 178.1], [569.55, 178.1], [571.09, 177.92], [578.36, 177.92]]}, {\"category\": \"cyclist\", \"corners_3d\": [[588.95, 197.97], [581.13, 197.97], [581.94, 197.24], [589.53, 197.24], [588.95, 177.12], [581.13, 177.12], [581.94, 176.99], [589.53, 176.99]]}, {\"category\": \"car\", \"corners_3d\": [[542.9, 206.59], [518.29, 206.6], [526.57, 203.51], [548.94, 203.51], [542.9, 185.08], [518.29, 185.08], [526.57, 183.96], [548.94, 183.96]]}, {\"category\": \"car\", \"corners_3d\": [[633.16, 192.8], [651.09, 192.82], [651.38, 193.99], [632.39, 193.97], [633.16, 178.11], [651.09, 178.11], [651.38, 178.42], [632.39, 178.41]]}, {\"category\": \"car\", \"corners_3d\": [[648.6, 210.34], [684.26, 210.29], [695.77, 214.78], [655.84, 214.85], [648.6, 175.85], [684.26, 175.84], [695.77, 176.2], [655.84, 176.21]]}]\n```" - } - ] - }, - { - "id": 1903, - "data_source": "Mono3DRefer", - "scene_name": "001619", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[728.15, 260.79], [817.56, 260.69], [929.37, 306.44], [793.49, 306.67], [728.15, 180.78], [817.56, 180.77], [929.37, 184.89], [793.49, 184.91]]}]\n```" - } - ] - }, - { - "id": 1904, - "data_source": "Mono3DRefer", - "scene_name": "001626", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[448.83, 209.09], [453.29, 207.78], [517.5, 207.63], [515.46, 208.92], [448.83, 186.23], [453.29, 185.74], [517.5, 185.68], [515.46, 186.16]]}]\n```" - } - ] - }, - { - "id": 1905, - "data_source": "Mono3DRefer", - "scene_name": "001626", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[30.93, 543.64], [-460.48, 538.71], [305.22, 280.26], [451.12, 280.68], [30.93, -204.51], [-460.48, -199.5], [305.22, 63.54], [451.12, 63.11]]}]\n```" - } - ] - }, - { - "id": 1906, - "data_source": "Mono3DRefer", - "scene_name": "001626", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[30.93, 543.64], [-460.48, 538.71], [305.22, 280.26], [451.12, 280.68], [30.93, -204.51], [-460.48, -199.5], [305.22, 63.54], [451.12, 63.11]]}, {\"category\": \"car\", \"corners_3d\": [[448.83, 209.09], [453.29, 207.78], [517.5, 207.63], [515.46, 208.92], [448.83, 186.23], [453.29, 185.74], [517.5, 185.68], [515.46, 186.16]]}]\n```" - } - ] - }, - { - "id": 1907, - "data_source": "Mono3DRefer", - "scene_name": "001629", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[370.07, 320.53], [195.47, 320.91], [374.87, 255.54], [472.23, 255.42], [370.07, 58.21], [195.47, 57.91], [374.87, 108.66], [472.23, 108.76]]}]\n```" - } - ] - }, - { - "id": 1908, - "data_source": "Mono3DRefer", - "scene_name": "001629", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[599.11, 252.33], [617.51, 250.84], [629.29, 252.53], [610.72, 254.08], [599.11, 174.69], [617.51, 174.65], [629.29, 174.69], [610.72, 174.73]]}]\n```" - } - ] - }, - { - "id": 1909, - "data_source": "Mono3DRefer", - "scene_name": "001629", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[617.53, 244.07], [683.44, 243.95], [703.93, 260.55], [622.63, 260.74], [617.53, 181.04], [683.44, 181.03], [703.93, 182.94], [622.63, 182.96]]}]\n```" - } - ] - }, - { - "id": 1910, - "data_source": "Mono3DRefer", - "scene_name": "001629", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[370.07, 320.53], [195.47, 320.91], [374.87, 255.54], [472.23, 255.42], [370.07, 58.21], [195.47, 57.91], [374.87, 108.66], [472.23, 108.76]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[599.11, 252.33], [617.51, 250.84], [629.29, 252.53], [610.72, 254.08], [599.11, 174.69], [617.51, 174.65], [629.29, 174.69], [610.72, 174.73]]}]\n```" - } - ] - }, - { - "id": 1911, - "data_source": "Mono3DRefer", - "scene_name": "001629", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[370.07, 320.53], [195.47, 320.91], [374.87, 255.54], [472.23, 255.42], [370.07, 58.21], [195.47, 57.91], [374.87, 108.66], [472.23, 108.76]]}, {\"category\": \"car\", \"corners_3d\": [[617.53, 244.07], [683.44, 243.95], [703.93, 260.55], [622.63, 260.74], [617.53, 181.04], [683.44, 181.03], [703.93, 182.94], [622.63, 182.96]]}]\n```" - } - ] - }, - { - "id": 1912, - "data_source": "Mono3DRefer", - "scene_name": "001629", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[617.53, 244.07], [683.44, 243.95], [703.93, 260.55], [622.63, 260.74], [617.53, 181.04], [683.44, 181.03], [703.93, 182.94], [622.63, 182.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[599.11, 252.33], [617.51, 250.84], [629.29, 252.53], [610.72, 254.08], [599.11, 174.69], [617.51, 174.65], [629.29, 174.69], [610.72, 174.73]]}]\n```" - } - ] - }, - { - "id": 1913, - "data_source": "Mono3DRefer", - "scene_name": "001629", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[370.07, 320.53], [195.47, 320.91], [374.87, 255.54], [472.23, 255.42], [370.07, 58.21], [195.47, 57.91], [374.87, 108.66], [472.23, 108.76]]}, {\"category\": \"car\", \"corners_3d\": [[617.53, 244.07], [683.44, 243.95], [703.93, 260.55], [622.63, 260.74], [617.53, 181.04], [683.44, 181.03], [703.93, 182.94], [622.63, 182.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[599.11, 252.33], [617.51, 250.84], [629.29, 252.53], [610.72, 254.08], [599.11, 174.69], [617.51, 174.65], [629.29, 174.69], [610.72, 174.73]]}]\n```" - } - ] - }, - { - "id": 1914, - "data_source": "Mono3DRefer", - "scene_name": "001634", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[989.84, 424.24], [1269.39, 430.9], [3112.08, 1212.11], [1890.19, 1111.94], [989.84, 187.88], [1269.39, 188.28], [3112.08, 234.98], [1890.19, 228.99]]}, {\"category\": \"car\", \"corners_3d\": [[-84.74, 445.99], [-288.61, 444.48], [74.21, 336.42], [197.92, 336.97], [-84.74, 241.1], [-288.61, 240.72], [74.21, 213.72], [197.92, 213.86]]}, {\"category\": \"car\", \"corners_3d\": [[470.83, 226.62], [431.1, 226.66], [450.05, 220.68], [485.36, 220.66], [470.83, 190.08], [431.1, 190.1], [450.05, 188.18], [485.36, 188.17]]}, {\"category\": \"car\", \"corners_3d\": [[650.88, 206.04], [686.74, 206.04], [695.95, 210.03], [655.77, 210.03], [650.88, 175.1], [686.74, 175.1], [695.95, 175.37], [655.77, 175.37]]}, {\"category\": \"car\", \"corners_3d\": [[491.74, 217.16], [459.01, 217.14], [475.97, 212.36], [505.18, 212.38], [491.74, 187.62], [459.01, 187.61], [475.97, 186.02], [505.18, 186.02]]}, {\"category\": \"car\", \"corners_3d\": [[518.43, 210.21], [489.75, 210.21], [500.49, 206.85], [526.59, 206.85], [518.43, 179.97], [489.75, 179.97], [500.49, 179.33], [526.59, 179.33]]}, {\"category\": \"car\", \"corners_3d\": [[121.06, 230.35], [153.9, 230.36], [103.1, 236.76], [66.6, 236.76], [121.06, 204.25], [153.9, 204.25], [103.1, 207.74], [66.6, 207.74]]}, {\"category\": \"car\", \"corners_3d\": [[168.18, 227.93], [197.35, 227.93], [151.3, 234.08], [118.88, 234.08], [168.18, 202.4], [197.35, 202.4], [151.3, 205.69], [118.88, 205.69]]}, {\"category\": \"car\", \"corners_3d\": [[212.35, 225.43], [243.43, 225.45], [194.86, 232.28], [159.77, 232.25], [212.35, 198.03], [243.43, 198.05], [194.86, 201.32], [159.77, 201.3]]}, {\"category\": \"car\", \"corners_3d\": [[260.89, 225.93], [289.43, 225.93], [248.62, 232.69], [216.44, 232.69], [260.89, 198.47], [289.43, 198.47], [248.62, 201.74], [216.44, 201.73]]}]\n```" - } - ] - }, - { - "id": 1915, - "data_source": "Mono3DRefer", - "scene_name": "001639", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[702.74, 200.34], [733.29, 200.31], [750.14, 203.5], [716.06, 203.55], [702.74, 165.38], [733.29, 165.39], [750.14, 164.52], [716.06, 164.51]]}]\n```" - } - ] - }, - { - "id": 1916, - "data_source": "Mono3DRefer", - "scene_name": "001639", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[795.74, 199.91], [811.56, 199.44], [860.79, 201.38], [844.72, 201.92], [795.74, 174.72], [811.56, 174.69], [860.79, 174.82], [844.72, 174.86]]}]\n```" - } - ] - }, - { - "id": 1917, - "data_source": "Mono3DRefer", - "scene_name": "001639", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[702.74, 200.34], [733.29, 200.31], [750.14, 203.5], [716.06, 203.55], [702.74, 165.38], [733.29, 165.39], [750.14, 164.52], [716.06, 164.51]]}, {\"category\": \"car\", \"corners_3d\": [[795.74, 199.91], [811.56, 199.44], [860.79, 201.38], [844.72, 201.92], [795.74, 174.72], [811.56, 174.69], [860.79, 174.82], [844.72, 174.86]]}]\n```" - } - ] - }, - { - "id": 1918, - "data_source": "Mono3DRefer", - "scene_name": "001646", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[734.82, 212.42], [768.2, 214.02], [703.56, 218.56], [669.43, 216.6], [734.82, 179.24], [768.2, 179.5], [703.56, 180.24], [669.43, 179.92]]}, {\"category\": \"car\", \"corners_3d\": [[715.65, 210.84], [745.34, 212.19], [691.02, 215.88], [660.69, 214.27], [715.65, 176.9], [745.34, 177.05], [691.02, 177.44], [660.69, 177.27]]}, {\"category\": \"car\", \"corners_3d\": [[596.46, 213.69], [633.49, 213.7], [636.52, 218.99], [594.69, 218.99], [596.46, 178.68], [633.49, 178.68], [636.52, 179.44], [594.69, 179.44]]}, {\"category\": \"car\", \"corners_3d\": [[214.32, 248.93], [154.95, 248.61], [237.16, 235.9], [286.83, 236.12], [214.32, 193.16], [154.95, 193.07], [237.16, 189.68], [286.83, 189.74]]}, {\"category\": \"car\", \"corners_3d\": [[310.36, 232.95], [257.41, 233.0], [312.28, 223.42], [356.76, 223.38], [310.36, 184.62], [257.41, 184.63], [312.28, 182.75], [356.76, 182.74]]}, {\"category\": \"car\", \"corners_3d\": [[777.58, 222.17], [820.95, 224.72], [752.47, 231.02], [707.86, 227.84], [777.58, 175.0], [820.95, 175.11], [752.47, 175.39], [707.86, 175.25]]}, {\"category\": \"car\", \"corners_3d\": [[766.84, 217.67], [805.34, 219.78], [735.76, 225.27], [696.46, 222.65], [766.84, 177.41], [805.34, 177.62], [735.76, 178.18], [696.46, 177.91]]}]\n```" - } - ] - }, - { - "id": 1919, - "data_source": "Mono3DRefer", - "scene_name": "001646", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[660.38, 190.8], [673.09, 191.08], [645.12, 191.88], [632.31, 191.57], [660.38, 173.93], [673.09, 173.94], [645.12, 173.99], [632.31, 173.97]]}]\n```" - } - ] - }, - { - "id": 1920, - "data_source": "Mono3DRefer", - "scene_name": "001646", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[734.82, 212.42], [768.2, 214.02], [703.56, 218.56], [669.43, 216.6], [734.82, 179.24], [768.2, 179.5], [703.56, 180.24], [669.43, 179.92]]}, {\"category\": \"car\", \"corners_3d\": [[715.65, 210.84], [745.34, 212.19], [691.02, 215.88], [660.69, 214.27], [715.65, 176.9], [745.34, 177.05], [691.02, 177.44], [660.69, 177.27]]}, {\"category\": \"van\", \"corners_3d\": [[660.38, 190.8], [673.09, 191.08], [645.12, 191.88], [632.31, 191.57], [660.38, 173.93], [673.09, 173.94], [645.12, 173.99], [632.31, 173.97]]}, {\"category\": \"car\", \"corners_3d\": [[596.46, 213.69], [633.49, 213.7], [636.52, 218.99], [594.69, 218.99], [596.46, 178.68], [633.49, 178.68], [636.52, 179.44], [594.69, 179.44]]}, {\"category\": \"car\", \"corners_3d\": [[214.32, 248.93], [154.95, 248.61], [237.16, 235.9], [286.83, 236.12], [214.32, 193.16], [154.95, 193.07], [237.16, 189.68], [286.83, 189.74]]}, {\"category\": \"car\", \"corners_3d\": [[310.36, 232.95], [257.41, 233.0], [312.28, 223.42], [356.76, 223.38], [310.36, 184.62], [257.41, 184.63], [312.28, 182.75], [356.76, 182.74]]}, {\"category\": \"car\", \"corners_3d\": [[777.58, 222.17], [820.95, 224.72], [752.47, 231.02], [707.86, 227.84], [777.58, 175.0], [820.95, 175.11], [752.47, 175.39], [707.86, 175.25]]}, {\"category\": \"car\", \"corners_3d\": [[766.84, 217.67], [805.34, 219.78], [735.76, 225.27], [696.46, 222.65], [766.84, 177.41], [805.34, 177.62], [735.76, 178.18], [696.46, 177.91]]}]\n```" - } - ] - }, - { - "id": 1921, - "data_source": "Mono3DRefer", - "scene_name": "001652", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[670.46, 214.68], [703.49, 214.99], [700.27, 220.85], [662.71, 220.45], [670.46, 175.74], [703.49, 175.76], [700.27, 176.16], [662.71, 176.14]]}]\n```" - } - ] - }, - { - "id": 1922, - "data_source": "Mono3DRefer", - "scene_name": "001652", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[592.86, 203.42], [568.91, 203.91], [518.84, 199.02], [539.74, 198.66], [592.86, 168.4], [568.91, 168.32], [518.84, 169.04], [539.74, 169.09]]}]\n```" - } - ] - }, - { - "id": 1923, - "data_source": "Mono3DRefer", - "scene_name": "001652", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[616.85, 211.45], [646.27, 211.69], [637.15, 216.55], [604.14, 216.25], [616.85, 185.83], [646.27, 185.92], [637.15, 187.55], [604.14, 187.45]]}]\n```" - } - ] - }, - { - "id": 1924, - "data_source": "Mono3DRefer", - "scene_name": "001652", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[670.46, 214.68], [703.49, 214.99], [700.27, 220.85], [662.71, 220.45], [670.46, 175.74], [703.49, 175.76], [700.27, 176.16], [662.71, 176.14]]}, {\"category\": \"tram\", \"corners_3d\": [[592.86, 203.42], [568.91, 203.91], [518.84, 199.02], [539.74, 198.66], [592.86, 168.4], [568.91, 168.32], [518.84, 169.04], [539.74, 169.09]]}]\n```" - } - ] - }, - { - "id": 1925, - "data_source": "Mono3DRefer", - "scene_name": "001652", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[616.85, 211.45], [646.27, 211.69], [637.15, 216.55], [604.14, 216.25], [616.85, 185.83], [646.27, 185.92], [637.15, 187.55], [604.14, 187.45]]}, {\"category\": \"van\", \"corners_3d\": [[670.46, 214.68], [703.49, 214.99], [700.27, 220.85], [662.71, 220.45], [670.46, 175.74], [703.49, 175.76], [700.27, 176.16], [662.71, 176.14]]}]\n```" - } - ] - }, - { - "id": 1926, - "data_source": "Mono3DRefer", - "scene_name": "001652", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[616.85, 211.45], [646.27, 211.69], [637.15, 216.55], [604.14, 216.25], [616.85, 185.83], [646.27, 185.92], [637.15, 187.55], [604.14, 187.45]]}, {\"category\": \"tram\", \"corners_3d\": [[592.86, 203.42], [568.91, 203.91], [518.84, 199.02], [539.74, 198.66], [592.86, 168.4], [568.91, 168.32], [518.84, 169.04], [539.74, 169.09]]}]\n```" - } - ] - }, - { - "id": 1927, - "data_source": "Mono3DRefer", - "scene_name": "001652", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[616.85, 211.45], [646.27, 211.69], [637.15, 216.55], [604.14, 216.25], [616.85, 185.83], [646.27, 185.92], [637.15, 187.55], [604.14, 187.45]]}, {\"category\": \"van\", \"corners_3d\": [[670.46, 214.68], [703.49, 214.99], [700.27, 220.85], [662.71, 220.45], [670.46, 175.74], [703.49, 175.76], [700.27, 176.16], [662.71, 176.14]]}, {\"category\": \"tram\", \"corners_3d\": [[592.86, 203.42], [568.91, 203.91], [518.84, 199.02], [539.74, 198.66], [592.86, 168.4], [568.91, 168.32], [518.84, 169.04], [539.74, 169.09]]}]\n```" - } - ] - }, - { - "id": 1928, - "data_source": "Mono3DRefer", - "scene_name": "001656", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[855.14, 395.04], [1075.97, 393.73], [2272.96, 944.22], [1515.37, 960.49], [855.14, 199.7], [1075.97, 199.54], [2272.96, 266.05], [1515.37, 268.02]]}, {\"category\": \"car\", \"corners_3d\": [[999.89, 358.2], [815.06, 359.69], [728.3, 288.51], [843.49, 287.94], [999.89, 192.63], [815.06, 192.79], [728.3, 185.2], [843.49, 185.14]]}, {\"category\": \"car\", \"corners_3d\": [[580.64, 258.14], [519.08, 252.69], [629.43, 238.38], [685.45, 242.01], [580.64, 178.42], [519.08, 178.07], [629.43, 177.13], [685.45, 177.37]]}]\n```" - } - ] - }, - { - "id": 1929, - "data_source": "Mono3DRefer", - "scene_name": "001662", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[596.04, 208.48], [603.81, 208.18], [608.89, 208.5], [601.1, 208.8], [596.04, 175.41], [603.81, 175.38], [608.89, 175.41], [601.1, 175.43]]}]\n```" - } - ] - }, - { - "id": 1930, - "data_source": "Mono3DRefer", - "scene_name": "001662", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1717.8, 522.5], [1344.17, 536.91], [876.56, 320.26], [1039.17, 317.84], [1717.8, 102.35], [1344.17, 99.45], [876.56, 143.13], [1039.17, 143.62]]}, {\"category\": \"car\", \"corners_3d\": [[106.99, 473.73], [-144.43, 476.89], [208.53, 330.24], [337.43, 329.39], [106.99, 236.58], [-144.43, 237.25], [208.53, 206.19], [337.43, 206.01]]}, {\"category\": \"car\", \"corners_3d\": [[439.59, 275.23], [352.24, 275.73], [401.91, 253.3], [470.1, 253.0], [439.59, 204.17], [352.24, 204.32], [401.91, 197.46], [470.1, 197.37]]}, {\"category\": \"car\", \"corners_3d\": [[488.43, 242.43], [425.22, 242.62], [452.97, 230.68], [505.32, 230.55], [488.43, 180.45], [425.22, 180.47], [452.97, 179.16], [505.32, 179.15]]}, {\"category\": \"car\", \"corners_3d\": [[603.75, 204.87], [632.68, 204.82], [637.34, 207.72], [605.79, 207.78], [603.75, 177.21], [632.68, 177.2], [637.34, 177.59], [605.79, 177.6]]}]\n```" - } - ] - }, - { - "id": 1931, - "data_source": "Mono3DRefer", - "scene_name": "001662", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[561.65, 204.14], [527.79, 204.06], [544.46, 199.88], [573.82, 199.95], [561.65, 152.51], [527.79, 152.57], [544.46, 155.29], [573.82, 155.24]]}]\n```" - } - ] - }, - { - "id": 1932, - "data_source": "Mono3DRefer", - "scene_name": "001662", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1717.8, 522.5], [1344.17, 536.91], [876.56, 320.26], [1039.17, 317.84], [1717.8, 102.35], [1344.17, 99.45], [876.56, 143.13], [1039.17, 143.62]]}, {\"category\": \"car\", \"corners_3d\": [[106.99, 473.73], [-144.43, 476.89], [208.53, 330.24], [337.43, 329.39], [106.99, 236.58], [-144.43, 237.25], [208.53, 206.19], [337.43, 206.01]]}, {\"category\": \"car\", \"corners_3d\": [[439.59, 275.23], [352.24, 275.73], [401.91, 253.3], [470.1, 253.0], [439.59, 204.17], [352.24, 204.32], [401.91, 197.46], [470.1, 197.37]]}, {\"category\": \"car\", \"corners_3d\": [[488.43, 242.43], [425.22, 242.62], [452.97, 230.68], [505.32, 230.55], [488.43, 180.45], [425.22, 180.47], [452.97, 179.16], [505.32, 179.15]]}, {\"category\": \"car\", \"corners_3d\": [[603.75, 204.87], [632.68, 204.82], [637.34, 207.72], [605.79, 207.78], [603.75, 177.21], [632.68, 177.2], [637.34, 177.59], [605.79, 177.6]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[596.04, 208.48], [603.81, 208.18], [608.89, 208.5], [601.1, 208.8], [596.04, 175.41], [603.81, 175.38], [608.89, 175.41], [601.1, 175.43]]}]\n```" - } - ] - }, - { - "id": 1933, - "data_source": "Mono3DRefer", - "scene_name": "001662", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[561.65, 204.14], [527.79, 204.06], [544.46, 199.88], [573.82, 199.95], [561.65, 152.51], [527.79, 152.57], [544.46, 155.29], [573.82, 155.24]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[596.04, 208.48], [603.81, 208.18], [608.89, 208.5], [601.1, 208.8], [596.04, 175.41], [603.81, 175.38], [608.89, 175.41], [601.1, 175.43]]}]\n```" - } - ] - }, - { - "id": 1934, - "data_source": "Mono3DRefer", - "scene_name": "001662", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1717.8, 522.5], [1344.17, 536.91], [876.56, 320.26], [1039.17, 317.84], [1717.8, 102.35], [1344.17, 99.45], [876.56, 143.13], [1039.17, 143.62]]}, {\"category\": \"car\", \"corners_3d\": [[106.99, 473.73], [-144.43, 476.89], [208.53, 330.24], [337.43, 329.39], [106.99, 236.58], [-144.43, 237.25], [208.53, 206.19], [337.43, 206.01]]}, {\"category\": \"car\", \"corners_3d\": [[439.59, 275.23], [352.24, 275.73], [401.91, 253.3], [470.1, 253.0], [439.59, 204.17], [352.24, 204.32], [401.91, 197.46], [470.1, 197.37]]}, {\"category\": \"car\", \"corners_3d\": [[488.43, 242.43], [425.22, 242.62], [452.97, 230.68], [505.32, 230.55], [488.43, 180.45], [425.22, 180.47], [452.97, 179.16], [505.32, 179.15]]}, {\"category\": \"truck\", \"corners_3d\": [[561.65, 204.14], [527.79, 204.06], [544.46, 199.88], [573.82, 199.95], [561.65, 152.51], [527.79, 152.57], [544.46, 155.29], [573.82, 155.24]]}, {\"category\": \"car\", \"corners_3d\": [[603.75, 204.87], [632.68, 204.82], [637.34, 207.72], [605.79, 207.78], [603.75, 177.21], [632.68, 177.2], [637.34, 177.59], [605.79, 177.6]]}]\n```" - } - ] - }, - { - "id": 1935, - "data_source": "Mono3DRefer", - "scene_name": "001662", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1717.8, 522.5], [1344.17, 536.91], [876.56, 320.26], [1039.17, 317.84], [1717.8, 102.35], [1344.17, 99.45], [876.56, 143.13], [1039.17, 143.62]]}, {\"category\": \"car\", \"corners_3d\": [[106.99, 473.73], [-144.43, 476.89], [208.53, 330.24], [337.43, 329.39], [106.99, 236.58], [-144.43, 237.25], [208.53, 206.19], [337.43, 206.01]]}, {\"category\": \"car\", \"corners_3d\": [[439.59, 275.23], [352.24, 275.73], [401.91, 253.3], [470.1, 253.0], [439.59, 204.17], [352.24, 204.32], [401.91, 197.46], [470.1, 197.37]]}, {\"category\": \"car\", \"corners_3d\": [[488.43, 242.43], [425.22, 242.62], [452.97, 230.68], [505.32, 230.55], [488.43, 180.45], [425.22, 180.47], [452.97, 179.16], [505.32, 179.15]]}, {\"category\": \"truck\", \"corners_3d\": [[561.65, 204.14], [527.79, 204.06], [544.46, 199.88], [573.82, 199.95], [561.65, 152.51], [527.79, 152.57], [544.46, 155.29], [573.82, 155.24]]}, {\"category\": \"car\", \"corners_3d\": [[603.75, 204.87], [632.68, 204.82], [637.34, 207.72], [605.79, 207.78], [603.75, 177.21], [632.68, 177.2], [637.34, 177.59], [605.79, 177.6]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[596.04, 208.48], [603.81, 208.18], [608.89, 208.5], [601.1, 208.8], [596.04, 175.41], [603.81, 175.38], [608.89, 175.41], [601.1, 175.43]]}]\n```" - } - ] - }, - { - "id": 1936, - "data_source": "Mono3DRefer", - "scene_name": "001666", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[763.78, 309.59], [888.61, 308.89], [1125.68, 416.35], [903.53, 418.58], [763.78, 184.09], [888.61, 184.03], [1125.68, 192.86], [903.53, 193.04]]}, {\"category\": \"car\", \"corners_3d\": [[358.41, 218.44], [372.64, 216.08], [452.06, 216.23], [442.15, 218.61], [358.41, 177.88], [372.64, 177.62], [452.06, 177.64], [442.15, 177.9]]}, {\"category\": \"car\", \"corners_3d\": [[729.29, 198.71], [699.05, 198.76], [686.87, 196.26], [714.21, 196.22], [729.29, 167.0], [699.05, 166.99], [686.87, 167.56], [714.21, 167.57]]}, {\"category\": \"car\", \"corners_3d\": [[590.6, 204.13], [564.96, 204.16], [566.94, 201.39], [590.32, 201.37], [590.6, 179.21], [564.96, 179.21], [566.94, 178.65], [590.32, 178.65]]}]\n```" - } - ] - }, - { - "id": 1937, - "data_source": "Mono3DRefer", - "scene_name": "001675", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[288.8, 228.24], [245.89, 228.21], [288.12, 221.9], [326.16, 221.92], [288.8, 194.4], [245.89, 194.39], [288.12, 191.93], [326.16, 191.94]]}]\n```" - } - ] - }, - { - "id": 1938, - "data_source": "Mono3DRefer", - "scene_name": "001675", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[599.17, 191.1], [623.17, 191.1], [624.25, 194.48], [595.81, 194.47], [599.17, 165.1], [623.17, 165.09], [624.25, 163.66], [595.81, 163.66]]}]\n```" - } - ] - }, - { - "id": 1939, - "data_source": "Mono3DRefer", - "scene_name": "001675", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[599.17, 191.1], [623.17, 191.1], [624.25, 194.48], [595.81, 194.47], [599.17, 165.1], [623.17, 165.09], [624.25, 163.66], [595.81, 163.66]]}, {\"category\": \"car\", \"corners_3d\": [[288.8, 228.24], [245.89, 228.21], [288.12, 221.9], [326.16, 221.92], [288.8, 194.4], [245.89, 194.39], [288.12, 191.93], [326.16, 191.94]]}]\n```" - } - ] - }, - { - "id": 1940, - "data_source": "Mono3DRefer", - "scene_name": "001677", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[95.86, 294.79], [-2.41, 294.63], [152.53, 264.13], [226.31, 264.22], [95.86, 203.17], [-2.41, 203.13], [152.53, 195.54], [226.31, 195.57]]}, {\"category\": \"car\", \"corners_3d\": [[415.33, 220.37], [376.86, 220.32], [408.34, 214.26], [441.91, 214.3], [415.33, 183.05], [376.86, 183.03], [408.34, 181.74], [441.91, 181.74]]}]\n```" - } - ] - }, - { - "id": 1941, - "data_source": "Mono3DRefer", - "scene_name": "001677", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[876.66, 200.52], [887.54, 200.52], [904.68, 202.23], [893.14, 202.23], [876.66, 156.54], [887.54, 156.54], [904.68, 155.53], [893.14, 155.53]]}]\n```" - } - ] - }, - { - "id": 1942, - "data_source": "Mono3DRefer", - "scene_name": "001677", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[640.42, 185.25], [660.27, 185.21], [708.7, 189.83], [681.61, 189.9], [640.42, 156.97], [660.27, 157.02], [708.7, 151.11], [681.61, 151.02]]}]\n```" - } - ] - }, - { - "id": 1943, - "data_source": "Mono3DRefer", - "scene_name": "001677", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[876.66, 200.52], [887.54, 200.52], [904.68, 202.23], [893.14, 202.23], [876.66, 156.54], [887.54, 156.54], [904.68, 155.53], [893.14, 155.53]]}, {\"category\": \"car\", \"corners_3d\": [[95.86, 294.79], [-2.41, 294.63], [152.53, 264.13], [226.31, 264.22], [95.86, 203.17], [-2.41, 203.13], [152.53, 195.54], [226.31, 195.57]]}, {\"category\": \"car\", \"corners_3d\": [[415.33, 220.37], [376.86, 220.32], [408.34, 214.26], [441.91, 214.3], [415.33, 183.05], [376.86, 183.03], [408.34, 181.74], [441.91, 181.74]]}]\n```" - } - ] - }, - { - "id": 1944, - "data_source": "Mono3DRefer", - "scene_name": "001677", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[640.42, 185.25], [660.27, 185.21], [708.7, 189.83], [681.61, 189.9], [640.42, 156.97], [660.27, 157.02], [708.7, 151.11], [681.61, 151.02]]}, {\"category\": \"car\", \"corners_3d\": [[95.86, 294.79], [-2.41, 294.63], [152.53, 264.13], [226.31, 264.22], [95.86, 203.17], [-2.41, 203.13], [152.53, 195.54], [226.31, 195.57]]}, {\"category\": \"car\", \"corners_3d\": [[415.33, 220.37], [376.86, 220.32], [408.34, 214.26], [441.91, 214.3], [415.33, 183.05], [376.86, 183.03], [408.34, 181.74], [441.91, 181.74]]}]\n```" - } - ] - }, - { - "id": 1945, - "data_source": "Mono3DRefer", - "scene_name": "001677", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[876.66, 200.52], [887.54, 200.52], [904.68, 202.23], [893.14, 202.23], [876.66, 156.54], [887.54, 156.54], [904.68, 155.53], [893.14, 155.53]]}, {\"category\": \"tram\", \"corners_3d\": [[640.42, 185.25], [660.27, 185.21], [708.7, 189.83], [681.61, 189.9], [640.42, 156.97], [660.27, 157.02], [708.7, 151.11], [681.61, 151.02]]}]\n```" - } - ] - }, - { - "id": 1946, - "data_source": "Mono3DRefer", - "scene_name": "001677", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[876.66, 200.52], [887.54, 200.52], [904.68, 202.23], [893.14, 202.23], [876.66, 156.54], [887.54, 156.54], [904.68, 155.53], [893.14, 155.53]]}, {\"category\": \"tram\", \"corners_3d\": [[640.42, 185.25], [660.27, 185.21], [708.7, 189.83], [681.61, 189.9], [640.42, 156.97], [660.27, 157.02], [708.7, 151.11], [681.61, 151.02]]}, {\"category\": \"car\", \"corners_3d\": [[95.86, 294.79], [-2.41, 294.63], [152.53, 264.13], [226.31, 264.22], [95.86, 203.17], [-2.41, 203.13], [152.53, 195.54], [226.31, 195.57]]}, {\"category\": \"car\", \"corners_3d\": [[415.33, 220.37], [376.86, 220.32], [408.34, 214.26], [441.91, 214.3], [415.33, 183.05], [376.86, 183.03], [408.34, 181.74], [441.91, 181.74]]}]\n```" - } - ] - }, - { - "id": 1947, - "data_source": "Mono3DRefer", - "scene_name": "001685", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[815.71, 226.15], [870.89, 226.51], [911.77, 237.72], [844.82, 237.19], [815.71, 173.18], [870.89, 173.19], [911.77, 173.25], [844.82, 173.25]]}, {\"category\": \"car\", \"corners_3d\": [[900.37, 221.01], [953.11, 221.29], [1017.57, 232.24], [952.55, 231.82], [900.37, 174.92], [953.11, 174.93], [1017.57, 175.4], [952.55, 175.38]]}, {\"category\": \"car\", \"corners_3d\": [[140.83, 240.16], [109.16, 244.71], [-16.13, 244.71], [23.46, 240.16], [140.83, 188.65], [109.16, 189.72], [-16.13, 189.72], [23.46, 188.65]]}]\n```" - } - ] - }, - { - "id": 1948, - "data_source": "Mono3DRefer", - "scene_name": "001685", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[104.78, 262.5], [26.32, 272.27], [-119.6, 264.36], [-33.51, 256.02], [104.78, 182.89], [26.32, 183.99], [-119.6, 183.1], [-33.51, 182.17]]}, {\"category\": \"van\", \"corners_3d\": [[203.06, 227.99], [175.83, 231.62], [53.09, 231.52], [87.83, 227.9], [203.06, 180.31], [175.83, 180.8], [53.09, 180.79], [87.83, 180.3]]}, {\"category\": \"van\", \"corners_3d\": [[268.12, 216.79], [249.1, 219.05], [131.6, 218.74], [156.2, 216.51], [268.12, 165.59], [249.1, 165.22], [131.6, 165.27], [156.2, 165.64]]}, {\"category\": \"van\", \"corners_3d\": [[301.75, 212.79], [288.85, 214.54], [182.47, 214.66], [199.9, 212.9], [301.75, 167.9], [288.85, 167.68], [182.47, 167.67], [199.9, 167.89]]}]\n```" - } - ] - }, - { - "id": 1949, - "data_source": "Mono3DRefer", - "scene_name": "001685", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[104.78, 262.5], [26.32, 272.27], [-119.6, 264.36], [-33.51, 256.02], [104.78, 182.89], [26.32, 183.99], [-119.6, 183.1], [-33.51, 182.17]]}, {\"category\": \"car\", \"corners_3d\": [[815.71, 226.15], [870.89, 226.51], [911.77, 237.72], [844.82, 237.19], [815.71, 173.18], [870.89, 173.19], [911.77, 173.25], [844.82, 173.25]]}, {\"category\": \"car\", \"corners_3d\": [[900.37, 221.01], [953.11, 221.29], [1017.57, 232.24], [952.55, 231.82], [900.37, 174.92], [953.11, 174.93], [1017.57, 175.4], [952.55, 175.38]]}, {\"category\": \"van\", \"corners_3d\": [[203.06, 227.99], [175.83, 231.62], [53.09, 231.52], [87.83, 227.9], [203.06, 180.31], [175.83, 180.8], [53.09, 180.79], [87.83, 180.3]]}, {\"category\": \"car\", \"corners_3d\": [[140.83, 240.16], [109.16, 244.71], [-16.13, 244.71], [23.46, 240.16], [140.83, 188.65], [109.16, 189.72], [-16.13, 189.72], [23.46, 188.65]]}, {\"category\": \"van\", \"corners_3d\": [[268.12, 216.79], [249.1, 219.05], [131.6, 218.74], [156.2, 216.51], [268.12, 165.59], [249.1, 165.22], [131.6, 165.27], [156.2, 165.64]]}, {\"category\": \"van\", \"corners_3d\": [[301.75, 212.79], [288.85, 214.54], [182.47, 214.66], [199.9, 212.9], [301.75, 167.9], [288.85, 167.68], [182.47, 167.67], [199.9, 167.89]]}]\n```" - } - ] - }, - { - "id": 1950, - "data_source": "Mono3DRefer", - "scene_name": "001691", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[697.24, 203.62], [721.07, 203.8], [720.05, 206.23], [694.37, 206.03], [697.24, 181.37], [721.07, 181.42], [720.05, 182.09], [694.37, 182.04]]}, {\"category\": \"car\", \"corners_3d\": [[777.44, 287.62], [889.36, 287.81], [1039.55, 351.29], [865.62, 350.82], [777.44, 182.17], [889.36, 182.19], [1039.55, 187.34], [865.62, 187.31]]}, {\"category\": \"car\", \"corners_3d\": [[597.41, 230.71], [548.64, 230.08], [572.15, 222.47], [614.59, 222.94], [597.41, 184.84], [548.64, 184.71], [572.15, 183.13], [614.59, 183.23]]}, {\"category\": \"car\", \"corners_3d\": [[753.98, 254.85], [824.93, 255.49], [876.43, 282.61], [781.91, 281.48], [753.98, 159.86], [824.93, 159.76], [876.43, 155.47], [781.91, 155.64]]}]\n```" - } - ] - }, - { - "id": 1951, - "data_source": "Mono3DRefer", - "scene_name": "001692", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[868.17, 250.57], [888.4, 250.57], [935.41, 262.18], [911.84, 262.18], [868.17, 143.53], [888.4, 143.53], [935.41, 137.41], [911.84, 137.41]]}, {\"category\": \"cyclist\", \"corners_3d\": [[580.86, 217.32], [596.07, 217.81], [562.08, 220.68], [546.2, 220.1], [580.86, 164.27], [596.07, 164.05], [562.08, 162.79], [546.2, 163.04]]}]\n```" - } - ] - }, - { - "id": 1952, - "data_source": "Mono3DRefer", - "scene_name": "001692", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[848.8, 288.2], [901.93, 292.22], [888.65, 303.93], [830.7, 299.03], [848.8, 153.61], [901.93, 152.6], [888.65, 149.68], [830.7, 150.9]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1033.93, 249.04], [1078.59, 252.87], [1063.87, 254.08], [1019.3, 250.12], [1033.93, 131.55], [1078.59, 128.81], [1063.87, 127.94], [1019.3, 130.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[729.87, 212.97], [703.82, 212.85], [702.71, 211.62], [727.76, 211.73], [729.87, 152.62], [703.82, 152.72], [702.71, 153.78], [727.76, 153.68]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[730.93, 202.81], [743.65, 203.25], [734.79, 203.5], [722.11, 203.05], [730.93, 157.61], [743.65, 157.16], [734.79, 156.91], [722.11, 157.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[746.53, 199.94], [762.64, 200.13], [759.75, 200.44], [743.42, 200.24], [746.53, 153.81], [762.64, 153.54], [759.75, 153.12], [743.42, 153.4]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[711.14, 205.24], [702.61, 204.75], [729.62, 204.29], [738.53, 204.76], [711.14, 150.55], [702.61, 151.14], [729.62, 151.69], [738.53, 151.12]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[726.57, 205.75], [719.03, 205.14], [745.57, 204.88], [753.69, 205.48], [726.57, 147.13], [719.03, 147.93], [745.57, 148.28], [753.69, 147.49]]}]\n```" - } - ] - }, - { - "id": 1953, - "data_source": "Mono3DRefer", - "scene_name": "001692", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[868.17, 250.57], [888.4, 250.57], [935.41, 262.18], [911.84, 262.18], [868.17, 143.53], [888.4, 143.53], [935.41, 137.41], [911.84, 137.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[848.8, 288.2], [901.93, 292.22], [888.65, 303.93], [830.7, 299.03], [848.8, 153.61], [901.93, 152.6], [888.65, 149.68], [830.7, 150.9]]}, {\"category\": \"cyclist\", \"corners_3d\": [[580.86, 217.32], [596.07, 217.81], [562.08, 220.68], [546.2, 220.1], [580.86, 164.27], [596.07, 164.05], [562.08, 162.79], [546.2, 163.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1033.93, 249.04], [1078.59, 252.87], [1063.87, 254.08], [1019.3, 250.12], [1033.93, 131.55], [1078.59, 128.81], [1063.87, 127.94], [1019.3, 130.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[729.87, 212.97], [703.82, 212.85], [702.71, 211.62], [727.76, 211.73], [729.87, 152.62], [703.82, 152.72], [702.71, 153.78], [727.76, 153.68]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[730.93, 202.81], [743.65, 203.25], [734.79, 203.5], [722.11, 203.05], [730.93, 157.61], [743.65, 157.16], [734.79, 156.91], [722.11, 157.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[746.53, 199.94], [762.64, 200.13], [759.75, 200.44], [743.42, 200.24], [746.53, 153.81], [762.64, 153.54], [759.75, 153.12], [743.42, 153.4]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[711.14, 205.24], [702.61, 204.75], [729.62, 204.29], [738.53, 204.76], [711.14, 150.55], [702.61, 151.14], [729.62, 151.69], [738.53, 151.12]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[726.57, 205.75], [719.03, 205.14], [745.57, 204.88], [753.69, 205.48], [726.57, 147.13], [719.03, 147.93], [745.57, 148.28], [753.69, 147.49]]}]\n```" - } - ] - }, - { - "id": 1954, - "data_source": "Mono3DRefer", - "scene_name": "001698", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[559.19, 202.14], [536.19, 202.15], [541.21, 199.9], [562.45, 199.9], [559.19, 180.36], [536.19, 180.36], [541.21, 179.79], [562.45, 179.78]]}, {\"category\": \"car\", \"corners_3d\": [[515.67, 228.08], [469.67, 228.05], [490.19, 220.32], [529.75, 220.35], [515.67, 179.34], [469.67, 179.34], [490.19, 178.43], [529.75, 178.43]]}]\n```" - } - ] - }, - { - "id": 1955, - "data_source": "Mono3DRefer", - "scene_name": "001703", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[563.65, 216.61], [530.58, 216.85], [529.74, 212.54], [559.58, 212.34], [563.65, 182.73], [530.58, 182.78], [529.74, 181.81], [559.58, 181.77]]}, {\"category\": \"car\", \"corners_3d\": [[528.65, 203.06], [506.73, 203.3], [500.38, 201.15], [520.81, 200.94], [528.65, 184.16], [506.73, 184.25], [500.38, 183.45], [520.81, 183.37]]}]\n```" - } - ] - }, - { - "id": 1956, - "data_source": "Mono3DRefer", - "scene_name": "001704", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[582.9, 200.22], [598.38, 200.21], [598.15, 201.58], [581.9, 201.59], [582.9, 188.03], [598.38, 188.03], [598.15, 188.79], [581.9, 188.79]]}]\n```" - } - ] - }, - { - "id": 1957, - "data_source": "Mono3DRefer", - "scene_name": "001704", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[403.43, 244.64], [343.73, 244.83], [392.37, 230.45], [440.08, 230.34], [403.43, 160.1], [343.73, 160.07], [392.37, 162.62], [440.08, 162.64]]}]\n```" - } - ] - }, - { - "id": 1958, - "data_source": "Mono3DRefer", - "scene_name": "001704", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[403.43, 244.64], [343.73, 244.83], [392.37, 230.45], [440.08, 230.34], [403.43, 160.1], [343.73, 160.07], [392.37, 162.62], [440.08, 162.64]]}, {\"category\": \"car\", \"corners_3d\": [[582.9, 200.22], [598.38, 200.21], [598.15, 201.58], [581.9, 201.59], [582.9, 188.03], [598.38, 188.03], [598.15, 188.79], [581.9, 188.79]]}]\n```" - } - ] - }, - { - "id": 1959, - "data_source": "Mono3DRefer", - "scene_name": "001712", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[231.27, 235.23], [324.45, 235.32], [-1093.84, 537.77], [-1623.0, 534.81], [231.27, 123.42], [324.45, 123.35], [-1093.84, -116.34], [-1623.0, -113.99]]}, {\"category\": \"tram\", \"corners_3d\": [[372.5, 212.14], [431.9, 212.11], [326.26, 236.37], [229.99, 236.44], [372.5, 144.07], [431.9, 144.09], [326.26, 126.33], [229.99, 126.27]]}]\n```" - } - ] - }, - { - "id": 1960, - "data_source": "Mono3DRefer", - "scene_name": "001712", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[246.02, 251.59], [240.82, 247.9], [273.98, 246.2], [280.65, 249.73], [246.02, 164.14], [240.82, 164.55], [273.98, 164.74], [280.65, 164.35]]}]\n```" - } - ] - }, - { - "id": 1961, - "data_source": "Mono3DRefer", - "scene_name": "001712", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[565.64, 190.55], [581.5, 190.55], [580.06, 191.27], [563.56, 191.26], [565.64, 175.33], [581.5, 175.33], [580.06, 175.43], [563.56, 175.43]]}]\n```" - } - ] - }, - { - "id": 1962, - "data_source": "Mono3DRefer", - "scene_name": "001712", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[594.93, 189.31], [618.08, 189.3], [620.76, 192.22], [593.52, 192.23], [594.93, 164.23], [618.08, 164.23], [620.76, 162.7], [593.52, 162.7]]}]\n```" - } - ] - }, - { - "id": 1963, - "data_source": "Mono3DRefer", - "scene_name": "001712", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[231.27, 235.23], [324.45, 235.32], [-1093.84, 537.77], [-1623.0, 534.81], [231.27, 123.42], [324.45, 123.35], [-1093.84, -116.34], [-1623.0, -113.99]]}, {\"category\": \"tram\", \"corners_3d\": [[372.5, 212.14], [431.9, 212.11], [326.26, 236.37], [229.99, 236.44], [372.5, 144.07], [431.9, 144.09], [326.26, 126.33], [229.99, 126.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[246.02, 251.59], [240.82, 247.9], [273.98, 246.2], [280.65, 249.73], [246.02, 164.14], [240.82, 164.55], [273.98, 164.74], [280.65, 164.35]]}]\n```" - } - ] - }, - { - "id": 1964, - "data_source": "Mono3DRefer", - "scene_name": "001712", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[565.64, 190.55], [581.5, 190.55], [580.06, 191.27], [563.56, 191.26], [565.64, 175.33], [581.5, 175.33], [580.06, 175.43], [563.56, 175.43]]}, {\"category\": \"tram\", \"corners_3d\": [[231.27, 235.23], [324.45, 235.32], [-1093.84, 537.77], [-1623.0, 534.81], [231.27, 123.42], [324.45, 123.35], [-1093.84, -116.34], [-1623.0, -113.99]]}, {\"category\": \"tram\", \"corners_3d\": [[372.5, 212.14], [431.9, 212.11], [326.26, 236.37], [229.99, 236.44], [372.5, 144.07], [431.9, 144.09], [326.26, 126.33], [229.99, 126.27]]}]\n```" - } - ] - }, - { - "id": 1965, - "data_source": "Mono3DRefer", - "scene_name": "001712", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[594.93, 189.31], [618.08, 189.3], [620.76, 192.22], [593.52, 192.23], [594.93, 164.23], [618.08, 164.23], [620.76, 162.7], [593.52, 162.7]]}, {\"category\": \"tram\", \"corners_3d\": [[231.27, 235.23], [324.45, 235.32], [-1093.84, 537.77], [-1623.0, 534.81], [231.27, 123.42], [324.45, 123.35], [-1093.84, -116.34], [-1623.0, -113.99]]}, {\"category\": \"tram\", \"corners_3d\": [[372.5, 212.14], [431.9, 212.11], [326.26, 236.37], [229.99, 236.44], [372.5, 144.07], [431.9, 144.09], [326.26, 126.33], [229.99, 126.27]]}]\n```" - } - ] - }, - { - "id": 1966, - "data_source": "Mono3DRefer", - "scene_name": "001712", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[565.64, 190.55], [581.5, 190.55], [580.06, 191.27], [563.56, 191.26], [565.64, 175.33], [581.5, 175.33], [580.06, 175.43], [563.56, 175.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[246.02, 251.59], [240.82, 247.9], [273.98, 246.2], [280.65, 249.73], [246.02, 164.14], [240.82, 164.55], [273.98, 164.74], [280.65, 164.35]]}]\n```" - } - ] - }, - { - "id": 1967, - "data_source": "Mono3DRefer", - "scene_name": "001712", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[594.93, 189.31], [618.08, 189.3], [620.76, 192.22], [593.52, 192.23], [594.93, 164.23], [618.08, 164.23], [620.76, 162.7], [593.52, 162.7]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[246.02, 251.59], [240.82, 247.9], [273.98, 246.2], [280.65, 249.73], [246.02, 164.14], [240.82, 164.55], [273.98, 164.74], [280.65, 164.35]]}]\n```" - } - ] - }, - { - "id": 1968, - "data_source": "Mono3DRefer", - "scene_name": "001712", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[565.64, 190.55], [581.5, 190.55], [580.06, 191.27], [563.56, 191.26], [565.64, 175.33], [581.5, 175.33], [580.06, 175.43], [563.56, 175.43]]}, {\"category\": \"bus\", \"corners_3d\": [[594.93, 189.31], [618.08, 189.3], [620.76, 192.22], [593.52, 192.23], [594.93, 164.23], [618.08, 164.23], [620.76, 162.7], [593.52, 162.7]]}]\n```" - } - ] - }, - { - "id": 1969, - "data_source": "Mono3DRefer", - "scene_name": "001712", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[565.64, 190.55], [581.5, 190.55], [580.06, 191.27], [563.56, 191.26], [565.64, 175.33], [581.5, 175.33], [580.06, 175.43], [563.56, 175.43]]}, {\"category\": \"tram\", \"corners_3d\": [[231.27, 235.23], [324.45, 235.32], [-1093.84, 537.77], [-1623.0, 534.81], [231.27, 123.42], [324.45, 123.35], [-1093.84, -116.34], [-1623.0, -113.99]]}, {\"category\": \"tram\", \"corners_3d\": [[372.5, 212.14], [431.9, 212.11], [326.26, 236.37], [229.99, 236.44], [372.5, 144.07], [431.9, 144.09], [326.26, 126.33], [229.99, 126.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[246.02, 251.59], [240.82, 247.9], [273.98, 246.2], [280.65, 249.73], [246.02, 164.14], [240.82, 164.55], [273.98, 164.74], [280.65, 164.35]]}]\n```" - } - ] - }, - { - "id": 1970, - "data_source": "Mono3DRefer", - "scene_name": "001712", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, pedestrian, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[594.93, 189.31], [618.08, 189.3], [620.76, 192.22], [593.52, 192.23], [594.93, 164.23], [618.08, 164.23], [620.76, 162.7], [593.52, 162.7]]}, {\"category\": \"tram\", \"corners_3d\": [[231.27, 235.23], [324.45, 235.32], [-1093.84, 537.77], [-1623.0, 534.81], [231.27, 123.42], [324.45, 123.35], [-1093.84, -116.34], [-1623.0, -113.99]]}, {\"category\": \"tram\", \"corners_3d\": [[372.5, 212.14], [431.9, 212.11], [326.26, 236.37], [229.99, 236.44], [372.5, 144.07], [431.9, 144.09], [326.26, 126.33], [229.99, 126.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[246.02, 251.59], [240.82, 247.9], [273.98, 246.2], [280.65, 249.73], [246.02, 164.14], [240.82, 164.55], [273.98, 164.74], [280.65, 164.35]]}]\n```" - } - ] - }, - { - "id": 1971, - "data_source": "Mono3DRefer", - "scene_name": "001712", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[565.64, 190.55], [581.5, 190.55], [580.06, 191.27], [563.56, 191.26], [565.64, 175.33], [581.5, 175.33], [580.06, 175.43], [563.56, 175.43]]}, {\"category\": \"bus\", \"corners_3d\": [[594.93, 189.31], [618.08, 189.3], [620.76, 192.22], [593.52, 192.23], [594.93, 164.23], [618.08, 164.23], [620.76, 162.7], [593.52, 162.7]]}, {\"category\": \"tram\", \"corners_3d\": [[231.27, 235.23], [324.45, 235.32], [-1093.84, 537.77], [-1623.0, 534.81], [231.27, 123.42], [324.45, 123.35], [-1093.84, -116.34], [-1623.0, -113.99]]}, {\"category\": \"tram\", \"corners_3d\": [[372.5, 212.14], [431.9, 212.11], [326.26, 236.37], [229.99, 236.44], [372.5, 144.07], [431.9, 144.09], [326.26, 126.33], [229.99, 126.27]]}]\n```" - } - ] - }, - { - "id": 1972, - "data_source": "Mono3DRefer", - "scene_name": "001712", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[565.64, 190.55], [581.5, 190.55], [580.06, 191.27], [563.56, 191.26], [565.64, 175.33], [581.5, 175.33], [580.06, 175.43], [563.56, 175.43]]}, {\"category\": \"bus\", \"corners_3d\": [[594.93, 189.31], [618.08, 189.3], [620.76, 192.22], [593.52, 192.23], [594.93, 164.23], [618.08, 164.23], [620.76, 162.7], [593.52, 162.7]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[246.02, 251.59], [240.82, 247.9], [273.98, 246.2], [280.65, 249.73], [246.02, 164.14], [240.82, 164.55], [273.98, 164.74], [280.65, 164.35]]}]\n```" - } - ] - }, - { - "id": 1973, - "data_source": "Mono3DRefer", - "scene_name": "001712", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, pedestrian, car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[565.64, 190.55], [581.5, 190.55], [580.06, 191.27], [563.56, 191.26], [565.64, 175.33], [581.5, 175.33], [580.06, 175.43], [563.56, 175.43]]}, {\"category\": \"bus\", \"corners_3d\": [[594.93, 189.31], [618.08, 189.3], [620.76, 192.22], [593.52, 192.23], [594.93, 164.23], [618.08, 164.23], [620.76, 162.7], [593.52, 162.7]]}, {\"category\": \"tram\", \"corners_3d\": [[231.27, 235.23], [324.45, 235.32], [-1093.84, 537.77], [-1623.0, 534.81], [231.27, 123.42], [324.45, 123.35], [-1093.84, -116.34], [-1623.0, -113.99]]}, {\"category\": \"tram\", \"corners_3d\": [[372.5, 212.14], [431.9, 212.11], [326.26, 236.37], [229.99, 236.44], [372.5, 144.07], [431.9, 144.09], [326.26, 126.33], [229.99, 126.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[246.02, 251.59], [240.82, 247.9], [273.98, 246.2], [280.65, 249.73], [246.02, 164.14], [240.82, 164.55], [273.98, 164.74], [280.65, 164.35]]}]\n```" - } - ] - }, - { - "id": 1974, - "data_source": "Mono3DRefer", - "scene_name": "001715", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[567.72, 202.03], [541.43, 202.03], [546.88, 199.68], [571.04, 199.68], [567.72, 180.03], [541.43, 180.03], [546.88, 179.45], [571.04, 179.45]]}]\n```" - } - ] - }, - { - "id": 1975, - "data_source": "Mono3DRefer", - "scene_name": "001727", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[755.32, 208.82], [767.89, 208.82], [776.52, 210.79], [763.27, 210.79], [755.32, 165.7], [767.89, 165.7], [776.52, 165.31], [763.27, 165.31]]}]\n```" - } - ] - }, - { - "id": 1976, - "data_source": "Mono3DRefer", - "scene_name": "001727", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[252.25, 236.2], [200.91, 236.2], [255.78, 227.69], [300.22, 227.68], [252.25, 190.33], [200.91, 190.33], [255.78, 187.98], [300.22, 187.98]]}]\n```" - } - ] - }, - { - "id": 1977, - "data_source": "Mono3DRefer", - "scene_name": "001727", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[395.54, 220.95], [325.76, 221.0], [439.13, 201.23], [480.19, 201.21], [395.54, 138.23], [325.76, 138.19], [439.13, 152.43], [480.19, 152.44]]}]\n```" - } - ] - }, - { - "id": 1978, - "data_source": "Mono3DRefer", - "scene_name": "001727", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[252.25, 236.2], [200.91, 236.2], [255.78, 227.69], [300.22, 227.68], [252.25, 190.33], [200.91, 190.33], [255.78, 187.98], [300.22, 187.98]]}, {\"category\": \"cyclist\", \"corners_3d\": [[755.32, 208.82], [767.89, 208.82], [776.52, 210.79], [763.27, 210.79], [755.32, 165.7], [767.89, 165.7], [776.52, 165.31], [763.27, 165.31]]}]\n```" - } - ] - }, - { - "id": 1979, - "data_source": "Mono3DRefer", - "scene_name": "001727", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[395.54, 220.95], [325.76, 221.0], [439.13, 201.23], [480.19, 201.21], [395.54, 138.23], [325.76, 138.19], [439.13, 152.43], [480.19, 152.44]]}, {\"category\": \"cyclist\", \"corners_3d\": [[755.32, 208.82], [767.89, 208.82], [776.52, 210.79], [763.27, 210.79], [755.32, 165.7], [767.89, 165.7], [776.52, 165.31], [763.27, 165.31]]}]\n```" - } - ] - }, - { - "id": 1980, - "data_source": "Mono3DRefer", - "scene_name": "001727", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[252.25, 236.2], [200.91, 236.2], [255.78, 227.69], [300.22, 227.68], [252.25, 190.33], [200.91, 190.33], [255.78, 187.98], [300.22, 187.98]]}, {\"category\": \"tram\", \"corners_3d\": [[395.54, 220.95], [325.76, 221.0], [439.13, 201.23], [480.19, 201.21], [395.54, 138.23], [325.76, 138.19], [439.13, 152.43], [480.19, 152.44]]}]\n```" - } - ] - }, - { - "id": 1981, - "data_source": "Mono3DRefer", - "scene_name": "001727", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[252.25, 236.2], [200.91, 236.2], [255.78, 227.69], [300.22, 227.68], [252.25, 190.33], [200.91, 190.33], [255.78, 187.98], [300.22, 187.98]]}, {\"category\": \"tram\", \"corners_3d\": [[395.54, 220.95], [325.76, 221.0], [439.13, 201.23], [480.19, 201.21], [395.54, 138.23], [325.76, 138.19], [439.13, 152.43], [480.19, 152.44]]}, {\"category\": \"cyclist\", \"corners_3d\": [[755.32, 208.82], [767.89, 208.82], [776.52, 210.79], [763.27, 210.79], [755.32, 165.7], [767.89, 165.7], [776.52, 165.31], [763.27, 165.31]]}]\n```" - } - ] - }, - { - "id": 1982, - "data_source": "Mono3DRefer", - "scene_name": "001728", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[152.2, 259.21], [91.36, 259.5], [172.63, 245.09], [223.16, 244.89], [152.2, 205.97], [91.36, 206.08], [172.63, 200.55], [223.16, 200.48]]}]\n```" - } - ] - }, - { - "id": 1983, - "data_source": "Mono3DRefer", - "scene_name": "001728", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[388.93, 236.1], [449.96, 236.05], [406.09, 254.18], [327.48, 254.26], [388.93, 158.79], [449.96, 158.8], [406.09, 154.76], [327.48, 154.75]]}]\n```" - } - ] - }, - { - "id": 1984, - "data_source": "Mono3DRefer", - "scene_name": "001728", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[388.93, 236.1], [449.96, 236.05], [406.09, 254.18], [327.48, 254.26], [388.93, 158.79], [449.96, 158.8], [406.09, 154.76], [327.48, 154.75]]}, {\"category\": \"car\", \"corners_3d\": [[152.2, 259.21], [91.36, 259.5], [172.63, 245.09], [223.16, 244.89], [152.2, 205.97], [91.36, 206.08], [172.63, 200.55], [223.16, 200.48]]}]\n```" - } - ] - }, - { - "id": 1985, - "data_source": "Mono3DRefer", - "scene_name": "001729", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1028.56, 412.11], [1299.78, 425.44], [3347.7, 1327.88], [1888.39, 1093.35], [1028.56, 209.06], [1299.78, 211.08], [3347.7, 347.65], [1888.39, 312.15]]}, {\"category\": \"car\", \"corners_3d\": [[786.55, 223.21], [829.57, 223.65], [849.98, 233.04], [798.88, 232.43], [786.55, 187.05], [829.57, 187.18], [849.98, 189.82], [798.88, 189.65]]}]\n```" - } - ] - }, - { - "id": 1986, - "data_source": "Mono3DRefer", - "scene_name": "001733", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[634.81, 208.34], [664.93, 208.28], [673.27, 211.81], [640.16, 211.88], [634.81, 179.18], [664.93, 179.17], [673.27, 179.8], [640.16, 179.82]]}, {\"category\": \"car\", \"corners_3d\": [[704.22, 259.91], [792.42, 259.49], [865.57, 289.5], [747.08, 290.27], [704.22, 182.22], [792.42, 182.18], [865.57, 185.4], [747.08, 185.49]]}, {\"category\": \"car\", \"corners_3d\": [[413.05, 244.32], [479.31, 243.94], [461.04, 258.78], [380.74, 259.34], [413.05, 193.89], [479.31, 193.77], [461.04, 198.14], [380.74, 198.31]]}, {\"category\": \"car\", \"corners_3d\": [[657.65, 226.78], [704.16, 226.68], [723.06, 235.47], [668.98, 235.61], [657.65, 182.49], [704.16, 182.47], [723.06, 184.04], [668.98, 184.06]]}, {\"category\": \"car\", \"corners_3d\": [[471.25, 224.24], [515.09, 224.24], [499.56, 232.64], [448.56, 232.63], [471.25, 182.34], [515.09, 182.34], [499.56, 183.9], [448.56, 183.89]]}, {\"category\": \"car\", \"corners_3d\": [[528.35, 215.36], [497.97, 215.4], [506.47, 211.66], [534.17, 211.63], [528.35, 185.8], [497.97, 185.81], [506.47, 184.67], [534.17, 184.66]]}]\n```" - } - ] - }, - { - "id": 1987, - "data_source": "Mono3DRefer", - "scene_name": "001733", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[576.33, 199.5], [569.9, 199.49], [571.22, 198.74], [577.47, 198.74], [576.33, 180.61], [569.9, 180.61], [571.22, 180.39], [577.47, 180.39]]}, {\"category\": \"cyclist\", \"corners_3d\": [[584.43, 198.19], [577.71, 198.19], [578.5, 197.55], [585.05, 197.55], [584.43, 180.28], [577.71, 180.28], [578.5, 180.09], [585.05, 180.09]]}]\n```" - } - ] - }, - { - "id": 1988, - "data_source": "Mono3DRefer", - "scene_name": "001733", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[634.81, 208.34], [664.93, 208.28], [673.27, 211.81], [640.16, 211.88], [634.81, 179.18], [664.93, 179.17], [673.27, 179.8], [640.16, 179.82]]}, {\"category\": \"car\", \"corners_3d\": [[704.22, 259.91], [792.42, 259.49], [865.57, 289.5], [747.08, 290.27], [704.22, 182.22], [792.42, 182.18], [865.57, 185.4], [747.08, 185.49]]}, {\"category\": \"car\", \"corners_3d\": [[413.05, 244.32], [479.31, 243.94], [461.04, 258.78], [380.74, 259.34], [413.05, 193.89], [479.31, 193.77], [461.04, 198.14], [380.74, 198.31]]}, {\"category\": \"car\", \"corners_3d\": [[657.65, 226.78], [704.16, 226.68], [723.06, 235.47], [668.98, 235.61], [657.65, 182.49], [704.16, 182.47], [723.06, 184.04], [668.98, 184.06]]}, {\"category\": \"car\", \"corners_3d\": [[471.25, 224.24], [515.09, 224.24], [499.56, 232.64], [448.56, 232.63], [471.25, 182.34], [515.09, 182.34], [499.56, 183.9], [448.56, 183.89]]}, {\"category\": \"car\", \"corners_3d\": [[528.35, 215.36], [497.97, 215.4], [506.47, 211.66], [534.17, 211.63], [528.35, 185.8], [497.97, 185.81], [506.47, 184.67], [534.17, 184.66]]}, {\"category\": \"cyclist\", \"corners_3d\": [[576.33, 199.5], [569.9, 199.49], [571.22, 198.74], [577.47, 198.74], [576.33, 180.61], [569.9, 180.61], [571.22, 180.39], [577.47, 180.39]]}, {\"category\": \"cyclist\", \"corners_3d\": [[584.43, 198.19], [577.71, 198.19], [578.5, 197.55], [585.05, 197.55], [584.43, 180.28], [577.71, 180.28], [578.5, 180.09], [585.05, 180.09]]}]\n```" - } - ] - }, - { - "id": 1989, - "data_source": "Mono3DRefer", - "scene_name": "001739", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[569.47, 222.81], [538.12, 223.05], [537.29, 218.35], [565.72, 218.15], [569.47, 190.65], [538.12, 190.73], [537.29, 189.06], [565.72, 188.99]]}, {\"category\": \"car\", \"corners_3d\": [[530.72, 207.32], [511.11, 207.55], [506.08, 205.33], [524.47, 205.13], [530.72, 190.43], [511.11, 190.54], [506.08, 189.41], [524.47, 189.31]]}]\n```" - } - ] - }, - { - "id": 1990, - "data_source": "Mono3DRefer", - "scene_name": "001750", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[57.66, 259.37], [-11.6, 259.46], [90.83, 245.0], [148.45, 244.94], [57.66, 196.28], [-11.6, 196.31], [90.83, 192.39], [148.45, 192.37]]}, {\"category\": \"car\", \"corners_3d\": [[-6.89, 251.05], [-77.25, 251.14], [51.15, 236.33], [108.11, 236.28], [-6.89, 195.02], [-77.25, 195.04], [51.15, 190.85], [108.11, 190.83]]}, {\"category\": \"car\", \"corners_3d\": [[318.57, 219.78], [278.59, 219.81], [313.12, 214.79], [348.81, 214.77], [318.57, 184.02], [278.59, 184.03], [313.12, 182.83], [348.81, 182.83]]}, {\"category\": \"car\", \"corners_3d\": [[277.94, 217.08], [241.18, 217.06], [278.73, 212.63], [311.83, 212.65], [277.94, 182.65], [241.18, 182.65], [278.73, 181.67], [311.83, 181.67]]}, {\"category\": \"car\", \"corners_3d\": [[434.46, 201.22], [410.57, 201.21], [424.63, 199.33], [446.95, 199.35], [434.46, 180.37], [410.57, 180.37], [424.63, 179.87], [446.95, 179.88]]}]\n```" - } - ] - }, - { - "id": 1991, - "data_source": "Mono3DRefer", - "scene_name": "001750", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[718.18, 205.7], [705.45, 205.71], [703.8, 205.22], [716.34, 205.21], [718.18, 176.08], [705.45, 176.08], [703.8, 176.04], [716.34, 176.03]]}]\n```" - } - ] - }, - { - "id": 1992, - "data_source": "Mono3DRefer", - "scene_name": "001750", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[12.85, 238.05], [-43.41, 238.11], [72.01, 226.43], [118.14, 226.4], [12.85, 180.55], [-43.41, 180.56], [72.01, 179.18], [118.14, 179.17]]}]\n```" - } - ] - }, - { - "id": 1993, - "data_source": "Mono3DRefer", - "scene_name": "001750", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[718.18, 205.7], [705.45, 205.71], [703.8, 205.22], [716.34, 205.21], [718.18, 176.08], [705.45, 176.08], [703.8, 176.04], [716.34, 176.03]]}, {\"category\": \"car\", \"corners_3d\": [[57.66, 259.37], [-11.6, 259.46], [90.83, 245.0], [148.45, 244.94], [57.66, 196.28], [-11.6, 196.31], [90.83, 192.39], [148.45, 192.37]]}, {\"category\": \"car\", \"corners_3d\": [[-6.89, 251.05], [-77.25, 251.14], [51.15, 236.33], [108.11, 236.28], [-6.89, 195.02], [-77.25, 195.04], [51.15, 190.85], [108.11, 190.83]]}, {\"category\": \"car\", \"corners_3d\": [[318.57, 219.78], [278.59, 219.81], [313.12, 214.79], [348.81, 214.77], [318.57, 184.02], [278.59, 184.03], [313.12, 182.83], [348.81, 182.83]]}, {\"category\": \"car\", \"corners_3d\": [[277.94, 217.08], [241.18, 217.06], [278.73, 212.63], [311.83, 212.65], [277.94, 182.65], [241.18, 182.65], [278.73, 181.67], [311.83, 181.67]]}, {\"category\": \"car\", \"corners_3d\": [[434.46, 201.22], [410.57, 201.21], [424.63, 199.33], [446.95, 199.35], [434.46, 180.37], [410.57, 180.37], [424.63, 179.87], [446.95, 179.88]]}]\n```" - } - ] - }, - { - "id": 1994, - "data_source": "Mono3DRefer", - "scene_name": "001750", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[12.85, 238.05], [-43.41, 238.11], [72.01, 226.43], [118.14, 226.4], [12.85, 180.55], [-43.41, 180.56], [72.01, 179.18], [118.14, 179.17]]}, {\"category\": \"car\", \"corners_3d\": [[57.66, 259.37], [-11.6, 259.46], [90.83, 245.0], [148.45, 244.94], [57.66, 196.28], [-11.6, 196.31], [90.83, 192.39], [148.45, 192.37]]}, {\"category\": \"car\", \"corners_3d\": [[-6.89, 251.05], [-77.25, 251.14], [51.15, 236.33], [108.11, 236.28], [-6.89, 195.02], [-77.25, 195.04], [51.15, 190.85], [108.11, 190.83]]}, {\"category\": \"car\", \"corners_3d\": [[318.57, 219.78], [278.59, 219.81], [313.12, 214.79], [348.81, 214.77], [318.57, 184.02], [278.59, 184.03], [313.12, 182.83], [348.81, 182.83]]}, {\"category\": \"car\", \"corners_3d\": [[277.94, 217.08], [241.18, 217.06], [278.73, 212.63], [311.83, 212.65], [277.94, 182.65], [241.18, 182.65], [278.73, 181.67], [311.83, 181.67]]}, {\"category\": \"car\", \"corners_3d\": [[434.46, 201.22], [410.57, 201.21], [424.63, 199.33], [446.95, 199.35], [434.46, 180.37], [410.57, 180.37], [424.63, 179.87], [446.95, 179.88]]}]\n```" - } - ] - }, - { - "id": 1995, - "data_source": "Mono3DRefer", - "scene_name": "001750", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[718.18, 205.7], [705.45, 205.71], [703.8, 205.22], [716.34, 205.21], [718.18, 176.08], [705.45, 176.08], [703.8, 176.04], [716.34, 176.03]]}, {\"category\": \"van\", \"corners_3d\": [[12.85, 238.05], [-43.41, 238.11], [72.01, 226.43], [118.14, 226.4], [12.85, 180.55], [-43.41, 180.56], [72.01, 179.18], [118.14, 179.17]]}]\n```" - } - ] - }, - { - "id": 1996, - "data_source": "Mono3DRefer", - "scene_name": "001750", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[718.18, 205.7], [705.45, 205.71], [703.8, 205.22], [716.34, 205.21], [718.18, 176.08], [705.45, 176.08], [703.8, 176.04], [716.34, 176.03]]}, {\"category\": \"van\", \"corners_3d\": [[12.85, 238.05], [-43.41, 238.11], [72.01, 226.43], [118.14, 226.4], [12.85, 180.55], [-43.41, 180.56], [72.01, 179.18], [118.14, 179.17]]}, {\"category\": \"car\", \"corners_3d\": [[57.66, 259.37], [-11.6, 259.46], [90.83, 245.0], [148.45, 244.94], [57.66, 196.28], [-11.6, 196.31], [90.83, 192.39], [148.45, 192.37]]}, {\"category\": \"car\", \"corners_3d\": [[-6.89, 251.05], [-77.25, 251.14], [51.15, 236.33], [108.11, 236.28], [-6.89, 195.02], [-77.25, 195.04], [51.15, 190.85], [108.11, 190.83]]}, {\"category\": \"car\", \"corners_3d\": [[318.57, 219.78], [278.59, 219.81], [313.12, 214.79], [348.81, 214.77], [318.57, 184.02], [278.59, 184.03], [313.12, 182.83], [348.81, 182.83]]}, {\"category\": \"car\", \"corners_3d\": [[277.94, 217.08], [241.18, 217.06], [278.73, 212.63], [311.83, 212.65], [277.94, 182.65], [241.18, 182.65], [278.73, 181.67], [311.83, 181.67]]}, {\"category\": \"car\", \"corners_3d\": [[434.46, 201.22], [410.57, 201.21], [424.63, 199.33], [446.95, 199.35], [434.46, 180.37], [410.57, 180.37], [424.63, 179.87], [446.95, 179.88]]}]\n```" - } - ] - }, - { - "id": 1997, - "data_source": "Mono3DRefer", - "scene_name": "001752", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[448.47, 201.3], [427.19, 201.27], [439.89, 199.55], [459.9, 199.58], [448.47, 184.41], [427.19, 184.4], [439.89, 183.7], [459.9, 183.72]]}]\n```" - } - ] - }, - { - "id": 1998, - "data_source": "Mono3DRefer", - "scene_name": "001752", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[594.56, 189.58], [618.35, 189.59], [618.53, 192.65], [590.39, 192.64], [594.56, 163.81], [618.35, 163.8], [618.53, 162.15], [590.39, 162.15]]}]\n```" - } - ] - }, - { - "id": 1999, - "data_source": "Mono3DRefer", - "scene_name": "001752", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[594.56, 189.58], [618.35, 189.59], [618.53, 192.65], [590.39, 192.64], [594.56, 163.81], [618.35, 163.8], [618.53, 162.15], [590.39, 162.15]]}, {\"category\": \"car\", \"corners_3d\": [[448.47, 201.3], [427.19, 201.27], [439.89, 199.55], [459.9, 199.58], [448.47, 184.41], [427.19, 184.4], [439.89, 183.7], [459.9, 183.72]]}]\n```" - } - ] - }, - { - "id": 2000, - "data_source": "Mono3DRefer", - "scene_name": "001758", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[139.41, 328.07], [169.55, 319.44], [259.1, 320.76], [234.28, 329.56], [139.41, 155.87], [169.55, 157.32], [259.1, 157.09], [234.28, 155.63]]}]\n```" - } - ] - }, - { - "id": 2001, - "data_source": "Mono3DRefer", - "scene_name": "001759", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[683.33, 196.91], [721.49, 196.85], [750.73, 201.6], [705.07, 201.69], [683.33, 143.52], [721.49, 143.6], [750.73, 137.8], [705.07, 137.69]]}]\n```" - } - ] - }, - { - "id": 2002, - "data_source": "Mono3DRefer", - "scene_name": "001759", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[636.04, 198.21], [660.74, 198.19], [667.21, 200.46], [640.31, 200.49], [636.04, 175.29], [660.74, 175.29], [667.21, 175.51], [640.31, 175.51]]}, {\"category\": \"car\", \"corners_3d\": [[685.1, 223.95], [733.27, 223.78], [765.38, 234.05], [707.58, 234.29], [685.1, 171.92], [733.27, 171.92], [765.38, 171.73], [707.58, 171.73]]}, {\"category\": \"car\", \"corners_3d\": [[756.71, 258.53], [838.77, 258.54], [908.59, 284.71], [801.47, 284.69], [756.71, 184.85], [838.77, 184.85], [908.59, 188.52], [801.47, 188.52]]}, {\"category\": \"car\", \"corners_3d\": [[946.75, 254.93], [1026.92, 254.66], [1209.4, 288.71], [1096.56, 289.25], [946.75, 181.45], [1026.92, 181.42], [1209.4, 184.99], [1096.56, 185.05]]}, {\"category\": \"car\", \"corners_3d\": [[821.09, 202.42], [836.11, 201.85], [891.22, 203.97], [876.24, 204.63], [821.09, 176.21], [836.11, 176.15], [891.22, 176.39], [876.24, 176.46]]}]\n```" - } - ] - }, - { - "id": 2003, - "data_source": "Mono3DRefer", - "scene_name": "001759", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[636.04, 198.21], [660.74, 198.19], [667.21, 200.46], [640.31, 200.49], [636.04, 175.29], [660.74, 175.29], [667.21, 175.51], [640.31, 175.51]]}, {\"category\": \"car\", \"corners_3d\": [[685.1, 223.95], [733.27, 223.78], [765.38, 234.05], [707.58, 234.29], [685.1, 171.92], [733.27, 171.92], [765.38, 171.73], [707.58, 171.73]]}, {\"category\": \"car\", \"corners_3d\": [[756.71, 258.53], [838.77, 258.54], [908.59, 284.71], [801.47, 284.69], [756.71, 184.85], [838.77, 184.85], [908.59, 188.52], [801.47, 188.52]]}, {\"category\": \"truck\", \"corners_3d\": [[683.33, 196.91], [721.49, 196.85], [750.73, 201.6], [705.07, 201.69], [683.33, 143.52], [721.49, 143.6], [750.73, 137.8], [705.07, 137.69]]}, {\"category\": \"car\", \"corners_3d\": [[946.75, 254.93], [1026.92, 254.66], [1209.4, 288.71], [1096.56, 289.25], [946.75, 181.45], [1026.92, 181.42], [1209.4, 184.99], [1096.56, 185.05]]}, {\"category\": \"car\", \"corners_3d\": [[821.09, 202.42], [836.11, 201.85], [891.22, 203.97], [876.24, 204.63], [821.09, 176.21], [836.11, 176.15], [891.22, 176.39], [876.24, 176.46]]}]\n```" - } - ] - }, - { - "id": 2004, - "data_source": "Mono3DRefer", - "scene_name": "001762", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[738.47, 206.01], [749.66, 206.03], [755.34, 207.62], [743.61, 207.61], [738.47, 167.83], [749.66, 167.83], [755.34, 167.58], [743.61, 167.59]]}]\n```" - } - ] - }, - { - "id": 2005, - "data_source": "Mono3DRefer", - "scene_name": "001762", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[451.31, 208.95], [400.08, 208.98], [468.39, 196.75], [502.25, 196.73], [451.31, 148.15], [400.08, 148.13], [468.39, 156.51], [502.25, 156.51]]}]\n```" - } - ] - }, - { - "id": 2006, - "data_source": "Mono3DRefer", - "scene_name": "001762", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[136.19, 244.03], [81.55, 243.98], [168.77, 232.37], [214.53, 232.41], [136.19, 193.55], [81.55, 193.54], [168.77, 190.16], [214.53, 190.17]]}, {\"category\": \"car\", \"corners_3d\": [[354.61, 217.33], [318.07, 217.33], [347.04, 212.91], [379.94, 212.9], [354.61, 184.68], [318.07, 184.68], [347.04, 183.51], [379.94, 183.5]]}]\n```" - } - ] - }, - { - "id": 2007, - "data_source": "Mono3DRefer", - "scene_name": "001762", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[451.31, 208.95], [400.08, 208.98], [468.39, 196.75], [502.25, 196.73], [451.31, 148.15], [400.08, 148.13], [468.39, 156.51], [502.25, 156.51]]}, {\"category\": \"cyclist\", \"corners_3d\": [[738.47, 206.01], [749.66, 206.03], [755.34, 207.62], [743.61, 207.61], [738.47, 167.83], [749.66, 167.83], [755.34, 167.58], [743.61, 167.59]]}]\n```" - } - ] - }, - { - "id": 2008, - "data_source": "Mono3DRefer", - "scene_name": "001762", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[738.47, 206.01], [749.66, 206.03], [755.34, 207.62], [743.61, 207.61], [738.47, 167.83], [749.66, 167.83], [755.34, 167.58], [743.61, 167.59]]}, {\"category\": \"car\", \"corners_3d\": [[136.19, 244.03], [81.55, 243.98], [168.77, 232.37], [214.53, 232.41], [136.19, 193.55], [81.55, 193.54], [168.77, 190.16], [214.53, 190.17]]}, {\"category\": \"car\", \"corners_3d\": [[354.61, 217.33], [318.07, 217.33], [347.04, 212.91], [379.94, 212.9], [354.61, 184.68], [318.07, 184.68], [347.04, 183.51], [379.94, 183.5]]}]\n```" - } - ] - }, - { - "id": 2009, - "data_source": "Mono3DRefer", - "scene_name": "001762", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[451.31, 208.95], [400.08, 208.98], [468.39, 196.75], [502.25, 196.73], [451.31, 148.15], [400.08, 148.13], [468.39, 156.51], [502.25, 156.51]]}, {\"category\": \"car\", \"corners_3d\": [[136.19, 244.03], [81.55, 243.98], [168.77, 232.37], [214.53, 232.41], [136.19, 193.55], [81.55, 193.54], [168.77, 190.16], [214.53, 190.17]]}, {\"category\": \"car\", \"corners_3d\": [[354.61, 217.33], [318.07, 217.33], [347.04, 212.91], [379.94, 212.9], [354.61, 184.68], [318.07, 184.68], [347.04, 183.51], [379.94, 183.5]]}]\n```" - } - ] - }, - { - "id": 2010, - "data_source": "Mono3DRefer", - "scene_name": "001762", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[451.31, 208.95], [400.08, 208.98], [468.39, 196.75], [502.25, 196.73], [451.31, 148.15], [400.08, 148.13], [468.39, 156.51], [502.25, 156.51]]}, {\"category\": \"cyclist\", \"corners_3d\": [[738.47, 206.01], [749.66, 206.03], [755.34, 207.62], [743.61, 207.61], [738.47, 167.83], [749.66, 167.83], [755.34, 167.58], [743.61, 167.59]]}, {\"category\": \"car\", \"corners_3d\": [[136.19, 244.03], [81.55, 243.98], [168.77, 232.37], [214.53, 232.41], [136.19, 193.55], [81.55, 193.54], [168.77, 190.16], [214.53, 190.17]]}, {\"category\": \"car\", \"corners_3d\": [[354.61, 217.33], [318.07, 217.33], [347.04, 212.91], [379.94, 212.9], [354.61, 184.68], [318.07, 184.68], [347.04, 183.51], [379.94, 183.5]]}]\n```" - } - ] - }, - { - "id": 2011, - "data_source": "Mono3DRefer", - "scene_name": "001772", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-761.41, 1436.24], [-2107.72, 1483.4], [122.33, 401.87], [350.34, 400.39], [-761.41, 284.09], [-2107.72, 288.24], [122.33, 193.02], [350.34, 192.89]]}, {\"category\": \"car\", \"corners_3d\": [[718.38, 268.31], [804.85, 267.85], [914.67, 314.51], [786.21, 315.52], [718.38, 176.84], [804.85, 176.82], [914.67, 178.76], [786.21, 178.8]]}, {\"category\": \"car\", \"corners_3d\": [[673.47, 240.36], [734.72, 240.02], [781.07, 258.56], [703.11, 259.12], [673.47, 178.76], [734.72, 178.73], [781.07, 180.36], [703.11, 180.4]]}, {\"category\": \"car\", \"corners_3d\": [[697.82, 194.01], [673.28, 194.06], [664.11, 192.37], [686.71, 192.32], [697.82, 168.24], [673.28, 168.23], [664.11, 168.6], [686.71, 168.61]]}]\n```" - } - ] - }, - { - "id": 2012, - "data_source": "Mono3DRefer", - "scene_name": "001772", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[900.52, 400.29], [1135.6, 396.57], [4916.89, 1902.94], [3417.51, 2153.69], [900.52, 122.87], [1135.6, 123.69], [4916.89, -207.38], [3417.51, -262.49]]}]\n```" - } - ] - }, - { - "id": 2013, - "data_source": "Mono3DRefer", - "scene_name": "001772", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-761.41, 1436.24], [-2107.72, 1483.4], [122.33, 401.87], [350.34, 400.39], [-761.41, 284.09], [-2107.72, 288.24], [122.33, 193.02], [350.34, 192.89]]}, {\"category\": \"van\", \"corners_3d\": [[900.52, 400.29], [1135.6, 396.57], [4916.89, 1902.94], [3417.51, 2153.69], [900.52, 122.87], [1135.6, 123.69], [4916.89, -207.38], [3417.51, -262.49]]}, {\"category\": \"car\", \"corners_3d\": [[718.38, 268.31], [804.85, 267.85], [914.67, 314.51], [786.21, 315.52], [718.38, 176.84], [804.85, 176.82], [914.67, 178.76], [786.21, 178.8]]}, {\"category\": \"car\", \"corners_3d\": [[673.47, 240.36], [734.72, 240.02], [781.07, 258.56], [703.11, 259.12], [673.47, 178.76], [734.72, 178.73], [781.07, 180.36], [703.11, 180.4]]}, {\"category\": \"car\", \"corners_3d\": [[697.82, 194.01], [673.28, 194.06], [664.11, 192.37], [686.71, 192.32], [697.82, 168.24], [673.28, 168.23], [664.11, 168.6], [686.71, 168.61]]}]\n```" - } - ] - }, - { - "id": 2014, - "data_source": "Mono3DRefer", - "scene_name": "001778", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[551.94, 233.64], [487.8, 233.75], [508.7, 221.83], [560.27, 221.76], [551.94, 176.37], [487.8, 176.38], [508.7, 175.69], [560.27, 175.68]]}, {\"category\": \"car\", \"corners_3d\": [[578.3, 207.49], [534.76, 207.24], [555.58, 202.8], [593.6, 202.99], [578.3, 166.01], [534.76, 166.06], [555.58, 166.94], [593.6, 166.9]]}, {\"category\": \"car\", \"corners_3d\": [[642.41, 192.36], [657.67, 192.93], [622.25, 193.63], [607.53, 193.01], [642.41, 167.25], [657.67, 167.09], [622.25, 166.89], [607.53, 167.06]]}, {\"category\": \"car\", \"corners_3d\": [[-262.35, 337.02], [-97.18, 304.25], [278.38, 300.27], [209.56, 330.85], [-262.35, 179.25], [-97.18, 177.98], [278.38, 177.82], [209.56, 179.01]]}]\n```" - } - ] - }, - { - "id": 2015, - "data_source": "Mono3DRefer", - "scene_name": "001778", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[475.8, 286.21], [341.16, 286.23], [419.26, 253.17], [514.63, 253.16], [475.8, 150.88], [341.16, 150.88], [419.26, 157.29], [514.63, 157.29]]}, {\"category\": \"van\", \"corners_3d\": [[718.21, 262.73], [806.51, 262.74], [883.73, 298.09], [760.71, 298.07], [718.21, 166.88], [806.51, 166.88], [883.73, 164.53], [760.71, 164.53]]}, {\"category\": \"van\", \"corners_3d\": [[566.29, 218.48], [515.98, 218.45], [531.16, 211.54], [573.86, 211.56], [566.29, 163.66], [515.98, 163.66], [531.16, 165.06], [573.86, 165.05]]}]\n```" - } - ] - }, - { - "id": 2016, - "data_source": "Mono3DRefer", - "scene_name": "001778", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[475.8, 286.21], [341.16, 286.23], [419.26, 253.17], [514.63, 253.16], [475.8, 150.88], [341.16, 150.88], [419.26, 157.29], [514.63, 157.29]]}, {\"category\": \"van\", \"corners_3d\": [[718.21, 262.73], [806.51, 262.74], [883.73, 298.09], [760.71, 298.07], [718.21, 166.88], [806.51, 166.88], [883.73, 164.53], [760.71, 164.53]]}, {\"category\": \"car\", \"corners_3d\": [[551.94, 233.64], [487.8, 233.75], [508.7, 221.83], [560.27, 221.76], [551.94, 176.37], [487.8, 176.38], [508.7, 175.69], [560.27, 175.68]]}, {\"category\": \"van\", \"corners_3d\": [[566.29, 218.48], [515.98, 218.45], [531.16, 211.54], [573.86, 211.56], [566.29, 163.66], [515.98, 163.66], [531.16, 165.06], [573.86, 165.05]]}, {\"category\": \"car\", \"corners_3d\": [[578.3, 207.49], [534.76, 207.24], [555.58, 202.8], [593.6, 202.99], [578.3, 166.01], [534.76, 166.06], [555.58, 166.94], [593.6, 166.9]]}, {\"category\": \"car\", \"corners_3d\": [[642.41, 192.36], [657.67, 192.93], [622.25, 193.63], [607.53, 193.01], [642.41, 167.25], [657.67, 167.09], [622.25, 166.89], [607.53, 167.06]]}, {\"category\": \"car\", \"corners_3d\": [[-262.35, 337.02], [-97.18, 304.25], [278.38, 300.27], [209.56, 330.85], [-262.35, 179.25], [-97.18, 177.98], [278.38, 177.82], [209.56, 179.01]]}]\n```" - } - ] - }, - { - "id": 2017, - "data_source": "Mono3DRefer", - "scene_name": "001779", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[362.32, 226.59], [382.85, 225.87], [390.58, 228.2], [369.25, 228.98], [362.32, 170.45], [382.85, 170.49], [390.58, 170.38], [369.25, 170.35]]}]\n```" - } - ] - }, - { - "id": 2018, - "data_source": "Mono3DRefer", - "scene_name": "001779", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-388.09, 375.46], [-171.03, 332.06], [220.5, 333.97], [108.49, 378.56], [-388.09, 202.27], [-171.03, 195.97], [220.5, 196.24], [108.49, 202.72]]}, {\"category\": \"car\", \"corners_3d\": [[48.98, 284.61], [108.61, 270.05], [366.39, 265.55], [346.49, 278.7], [48.98, 194.4], [108.61, 191.6], [366.39, 190.73], [346.49, 193.26]]}, {\"category\": \"car\", \"corners_3d\": [[1735.88, 438.72], [2487.91, 608.61], [1648.39, 652.1], [1198.77, 454.3], [1735.88, 163.87], [2487.91, 158.12], [1648.39, 156.65], [1198.77, 163.34]]}, {\"category\": \"car\", \"corners_3d\": [[1006.18, 243.11], [1083.97, 243.19], [1192.75, 259.59], [1096.7, 259.46], [1006.18, 160.5], [1083.97, 160.48], [1192.75, 157.6], [1096.7, 157.62]]}, {\"category\": \"car\", \"corners_3d\": [[1116.32, 238.95], [1209.25, 238.79], [1410.87, 260.45], [1287.97, 260.75], [1116.32, 155.52], [1209.25, 155.56], [1410.87, 149.88], [1287.97, 149.8]]}, {\"category\": \"car\", \"corners_3d\": [[341.25, 236.85], [314.64, 241.74], [157.99, 240.01], [194.85, 235.35], [341.25, 184.91], [314.64, 185.83], [157.99, 185.51], [194.85, 184.63]]}]\n```" - } - ] - }, - { - "id": 2019, - "data_source": "Mono3DRefer", - "scene_name": "001779", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-388.09, 375.46], [-171.03, 332.06], [220.5, 333.97], [108.49, 378.56], [-388.09, 202.27], [-171.03, 195.97], [220.5, 196.24], [108.49, 202.72]]}, {\"category\": \"car\", \"corners_3d\": [[48.98, 284.61], [108.61, 270.05], [366.39, 265.55], [346.49, 278.7], [48.98, 194.4], [108.61, 191.6], [366.39, 190.73], [346.49, 193.26]]}, {\"category\": \"car\", \"corners_3d\": [[1735.88, 438.72], [2487.91, 608.61], [1648.39, 652.1], [1198.77, 454.3], [1735.88, 163.87], [2487.91, 158.12], [1648.39, 156.65], [1198.77, 163.34]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[362.32, 226.59], [382.85, 225.87], [390.58, 228.2], [369.25, 228.98], [362.32, 170.45], [382.85, 170.49], [390.58, 170.38], [369.25, 170.35]]}, {\"category\": \"car\", \"corners_3d\": [[1006.18, 243.11], [1083.97, 243.19], [1192.75, 259.59], [1096.7, 259.46], [1006.18, 160.5], [1083.97, 160.48], [1192.75, 157.6], [1096.7, 157.62]]}, {\"category\": \"car\", \"corners_3d\": [[1116.32, 238.95], [1209.25, 238.79], [1410.87, 260.45], [1287.97, 260.75], [1116.32, 155.52], [1209.25, 155.56], [1410.87, 149.88], [1287.97, 149.8]]}, {\"category\": \"car\", \"corners_3d\": [[341.25, 236.85], [314.64, 241.74], [157.99, 240.01], [194.85, 235.35], [341.25, 184.91], [314.64, 185.83], [157.99, 185.51], [194.85, 184.63]]}]\n```" - } - ] - }, - { - "id": 2020, - "data_source": "Mono3DRefer", - "scene_name": "001783", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[634.91, 208.79], [605.8, 208.78], [606.78, 205.38], [633.14, 205.39], [634.91, 181.83], [605.8, 181.83], [606.78, 180.98], [633.14, 180.98]]}]\n```" - } - ] - }, - { - "id": 2021, - "data_source": "Mono3DRefer", - "scene_name": "001783", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (motorcyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"motorcyclist\", \"corners_3d\": [[-153.02, 761.49], [-355.02, 743.16], [72.35, 507.98], [197.47, 514.23], [-153.02, 199.5], [-355.02, 198.67], [72.35, 188.02], [197.47, 188.3]]}]\n```" - } - ] - }, - { - "id": 2022, - "data_source": "Mono3DRefer", - "scene_name": "001783", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, motorcyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"motorcyclist\", \"corners_3d\": [[-153.02, 761.49], [-355.02, 743.16], [72.35, 507.98], [197.47, 514.23], [-153.02, 199.5], [-355.02, 198.67], [72.35, 188.02], [197.47, 188.3]]}, {\"category\": \"car\", \"corners_3d\": [[634.91, 208.79], [605.8, 208.78], [606.78, 205.38], [633.14, 205.39], [634.91, 181.83], [605.8, 181.83], [606.78, 180.98], [633.14, 180.98]]}]\n```" - } - ] - }, - { - "id": 2023, - "data_source": "Mono3DRefer", - "scene_name": "001784", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[357.55, 353.19], [189.55, 352.8], [337.19, 290.53], [447.18, 290.7], [357.55, 203.05], [189.55, 202.99], [337.19, 192.56], [447.18, 192.59]]}, {\"category\": \"car\", \"corners_3d\": [[505.71, 253.66], [430.88, 253.66], [464.83, 238.26], [525.4, 238.25], [505.71, 183.92], [430.88, 183.92], [464.83, 181.81], [525.4, 181.81]]}]\n```" - } - ] - }, - { - "id": 2024, - "data_source": "Mono3DRefer", - "scene_name": "001807", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[407.68, 209.81], [380.1, 209.77], [398.61, 207.04], [424.17, 207.08], [407.68, 185.71], [380.1, 185.7], [398.61, 184.75], [424.17, 184.76]]}]\n```" - } - ] - }, - { - "id": 2025, - "data_source": "Mono3DRefer", - "scene_name": "001807", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[655.36, 208.33], [711.92, 208.33], [771.13, 228.97], [681.66, 228.96], [655.36, 147.04], [711.92, 147.04], [771.13, 132.02], [681.66, 132.03]]}]\n```" - } - ] - }, - { - "id": 2026, - "data_source": "Mono3DRefer", - "scene_name": "001807", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[655.36, 208.33], [711.92, 208.33], [771.13, 228.97], [681.66, 228.96], [655.36, 147.04], [711.92, 147.04], [771.13, 132.02], [681.66, 132.03]]}, {\"category\": \"car\", \"corners_3d\": [[407.68, 209.81], [380.1, 209.77], [398.61, 207.04], [424.17, 207.08], [407.68, 185.71], [380.1, 185.7], [398.61, 184.75], [424.17, 184.76]]}]\n```" - } - ] - }, - { - "id": 2027, - "data_source": "Mono3DRefer", - "scene_name": "001809", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[544.95, 209.19], [513.42, 209.24], [520.39, 205.81], [548.94, 205.77], [544.95, 178.14], [513.42, 178.15], [520.39, 177.65], [548.94, 177.65]]}]\n```" - } - ] - }, - { - "id": 2028, - "data_source": "Mono3DRefer", - "scene_name": "001811", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[807.33, 237.26], [845.59, 242.33], [712.18, 248.24], [681.43, 242.3], [807.33, 174.77], [845.59, 174.92], [712.18, 175.1], [681.43, 174.92]]}, {\"category\": \"car\", \"corners_3d\": [[1386.18, 406.42], [1783.85, 488.97], [1529.62, 740.58], [1037.05, 520.14], [1386.18, 214.37], [1783.85, 229.04], [1529.62, 273.76], [1037.05, 234.58]]}, {\"category\": \"car\", \"corners_3d\": [[484.38, 354.91], [338.05, 354.94], [430.28, 292.95], [526.79, 292.94], [484.38, 213.74], [338.05, 213.74], [430.28, 199.82], [526.79, 199.82]]}]\n```" - } - ] - }, - { - "id": 2029, - "data_source": "Mono3DRefer", - "scene_name": "001811", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[581.54, 282.72], [555.03, 284.88], [514.83, 271.65], [538.94, 269.96], [581.54, 172.21], [555.03, 172.2], [514.83, 172.28], [538.94, 172.29]]}]\n```" - } - ] - }, - { - "id": 2030, - "data_source": "Mono3DRefer", - "scene_name": "001811", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[581.54, 282.72], [555.03, 284.88], [514.83, 271.65], [538.94, 269.96], [581.54, 172.21], [555.03, 172.2], [514.83, 172.28], [538.94, 172.29]]}, {\"category\": \"car\", \"corners_3d\": [[807.33, 237.26], [845.59, 242.33], [712.18, 248.24], [681.43, 242.3], [807.33, 174.77], [845.59, 174.92], [712.18, 175.1], [681.43, 174.92]]}, {\"category\": \"car\", \"corners_3d\": [[1386.18, 406.42], [1783.85, 488.97], [1529.62, 740.58], [1037.05, 520.14], [1386.18, 214.37], [1783.85, 229.04], [1529.62, 273.76], [1037.05, 234.58]]}, {\"category\": \"car\", \"corners_3d\": [[484.38, 354.91], [338.05, 354.94], [430.28, 292.95], [526.79, 292.94], [484.38, 213.74], [338.05, 213.74], [430.28, 199.82], [526.79, 199.82]]}]\n```" - } - ] - }, - { - "id": 2031, - "data_source": "Mono3DRefer", - "scene_name": "001819", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[513.16, 302.82], [516.98, 296.02], [605.04, 295.56], [606.41, 302.31], [513.16, 158.7], [516.98, 159.91], [605.04, 159.99], [606.41, 158.79]]}]\n```" - } - ] - }, - { - "id": 2032, - "data_source": "Mono3DRefer", - "scene_name": "001824", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[242.49, 229.78], [199.88, 229.75], [254.75, 222.25], [291.76, 222.28], [242.49, 190.48], [199.88, 190.47], [254.75, 188.15], [291.76, 188.16]]}, {\"category\": \"car\", \"corners_3d\": [[394.7, 210.3], [364.18, 210.29], [385.55, 207.11], [413.49, 207.13], [394.7, 182.94], [364.18, 182.93], [385.55, 182.08], [413.49, 182.08]]}]\n```" - } - ] - }, - { - "id": 2033, - "data_source": "Mono3DRefer", - "scene_name": "001824", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[729.03, 204.47], [739.45, 204.49], [743.61, 205.9], [732.72, 205.88], [729.03, 169.0], [739.45, 168.99], [743.61, 168.82], [732.72, 168.82]]}]\n```" - } - ] - }, - { - "id": 2034, - "data_source": "Mono3DRefer", - "scene_name": "001824", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[474.61, 204.01], [431.1, 204.03], [482.85, 194.57], [513.15, 194.56], [474.61, 152.34], [431.1, 152.33], [482.85, 158.55], [513.15, 158.56]]}]\n```" - } - ] - }, - { - "id": 2035, - "data_source": "Mono3DRefer", - "scene_name": "001824", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[729.03, 204.47], [739.45, 204.49], [743.61, 205.9], [732.72, 205.88], [729.03, 169.0], [739.45, 168.99], [743.61, 168.82], [732.72, 168.82]]}, {\"category\": \"car\", \"corners_3d\": [[242.49, 229.78], [199.88, 229.75], [254.75, 222.25], [291.76, 222.28], [242.49, 190.48], [199.88, 190.47], [254.75, 188.15], [291.76, 188.16]]}, {\"category\": \"car\", \"corners_3d\": [[394.7, 210.3], [364.18, 210.29], [385.55, 207.11], [413.49, 207.13], [394.7, 182.94], [364.18, 182.93], [385.55, 182.08], [413.49, 182.08]]}]\n```" - } - ] - }, - { - "id": 2036, - "data_source": "Mono3DRefer", - "scene_name": "001824", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[474.61, 204.01], [431.1, 204.03], [482.85, 194.57], [513.15, 194.56], [474.61, 152.34], [431.1, 152.33], [482.85, 158.55], [513.15, 158.56]]}, {\"category\": \"car\", \"corners_3d\": [[242.49, 229.78], [199.88, 229.75], [254.75, 222.25], [291.76, 222.28], [242.49, 190.48], [199.88, 190.47], [254.75, 188.15], [291.76, 188.16]]}, {\"category\": \"car\", \"corners_3d\": [[394.7, 210.3], [364.18, 210.29], [385.55, 207.11], [413.49, 207.13], [394.7, 182.94], [364.18, 182.93], [385.55, 182.08], [413.49, 182.08]]}]\n```" - } - ] - }, - { - "id": 2037, - "data_source": "Mono3DRefer", - "scene_name": "001824", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[474.61, 204.01], [431.1, 204.03], [482.85, 194.57], [513.15, 194.56], [474.61, 152.34], [431.1, 152.33], [482.85, 158.55], [513.15, 158.56]]}, {\"category\": \"cyclist\", \"corners_3d\": [[729.03, 204.47], [739.45, 204.49], [743.61, 205.9], [732.72, 205.88], [729.03, 169.0], [739.45, 168.99], [743.61, 168.82], [732.72, 168.82]]}]\n```" - } - ] - }, - { - "id": 2038, - "data_source": "Mono3DRefer", - "scene_name": "001824", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[474.61, 204.01], [431.1, 204.03], [482.85, 194.57], [513.15, 194.56], [474.61, 152.34], [431.1, 152.33], [482.85, 158.55], [513.15, 158.56]]}, {\"category\": \"cyclist\", \"corners_3d\": [[729.03, 204.47], [739.45, 204.49], [743.61, 205.9], [732.72, 205.88], [729.03, 169.0], [739.45, 168.99], [743.61, 168.82], [732.72, 168.82]]}, {\"category\": \"car\", \"corners_3d\": [[242.49, 229.78], [199.88, 229.75], [254.75, 222.25], [291.76, 222.28], [242.49, 190.48], [199.88, 190.47], [254.75, 188.15], [291.76, 188.16]]}, {\"category\": \"car\", \"corners_3d\": [[394.7, 210.3], [364.18, 210.29], [385.55, 207.11], [413.49, 207.13], [394.7, 182.94], [364.18, 182.93], [385.55, 182.08], [413.49, 182.08]]}]\n```" - } - ] - }, - { - "id": 2039, - "data_source": "Mono3DRefer", - "scene_name": "001829", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1024.47, 203.35], [984.56, 202.54], [993.56, 200.05], [1030.26, 200.73], [1024.47, 173.46], [984.56, 173.44], [993.56, 173.4], [1030.26, 173.41]]}, {\"category\": \"car\", \"corners_3d\": [[853.01, 234.55], [792.45, 231.77], [835.62, 222.97], [888.49, 224.97], [853.01, 182.08], [792.45, 181.66], [835.62, 180.35], [888.49, 180.65]]}, {\"category\": \"car\", \"corners_3d\": [[910.32, 219.73], [870.46, 218.37], [893.7, 213.59], [929.87, 214.68], [910.32, 182.22], [870.46, 181.95], [893.7, 181.0], [929.87, 181.21]]}, {\"category\": \"car\", \"corners_3d\": [[969.21, 204.98], [934.9, 204.33], [939.18, 202.35], [971.38, 202.92], [969.21, 177.6], [934.9, 177.5], [939.18, 177.21], [971.38, 177.3]]}, {\"category\": \"car\", \"corners_3d\": [[778.37, 254.23], [710.81, 249.99], [779.3, 235.33], [836.48, 238.09], [778.37, 182.1], [710.81, 181.62], [779.3, 179.95], [836.48, 180.26]]}]\n```" - } - ] - }, - { - "id": 2040, - "data_source": "Mono3DRefer", - "scene_name": "001829", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[992.66, 198.61], [954.31, 198.02], [960.07, 195.52], [994.66, 196.0], [992.66, 157.0], [954.31, 157.36], [960.07, 158.9], [994.66, 158.6]]}]\n```" - } - ] - }, - { - "id": 2041, - "data_source": "Mono3DRefer", - "scene_name": "001829", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1024.47, 203.35], [984.56, 202.54], [993.56, 200.05], [1030.26, 200.73], [1024.47, 173.46], [984.56, 173.44], [993.56, 173.4], [1030.26, 173.41]]}, {\"category\": \"car\", \"corners_3d\": [[853.01, 234.55], [792.45, 231.77], [835.62, 222.97], [888.49, 224.97], [853.01, 182.08], [792.45, 181.66], [835.62, 180.35], [888.49, 180.65]]}, {\"category\": \"car\", \"corners_3d\": [[910.32, 219.73], [870.46, 218.37], [893.7, 213.59], [929.87, 214.68], [910.32, 182.22], [870.46, 181.95], [893.7, 181.0], [929.87, 181.21]]}, {\"category\": \"van\", \"corners_3d\": [[992.66, 198.61], [954.31, 198.02], [960.07, 195.52], [994.66, 196.0], [992.66, 157.0], [954.31, 157.36], [960.07, 158.9], [994.66, 158.6]]}, {\"category\": \"car\", \"corners_3d\": [[969.21, 204.98], [934.9, 204.33], [939.18, 202.35], [971.38, 202.92], [969.21, 177.6], [934.9, 177.5], [939.18, 177.21], [971.38, 177.3]]}, {\"category\": \"car\", \"corners_3d\": [[778.37, 254.23], [710.81, 249.99], [779.3, 235.33], [836.48, 238.09], [778.37, 182.1], [710.81, 181.62], [779.3, 179.95], [836.48, 180.26]]}]\n```" - } - ] - }, - { - "id": 2042, - "data_source": "Mono3DRefer", - "scene_name": "001832", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[445.4, 331.41], [303.41, 330.17], [417.4, 276.55], [511.34, 277.09], [445.4, 184.08], [303.41, 183.99], [417.4, 180.2], [511.34, 180.23]]}, {\"category\": \"car\", \"corners_3d\": [[-522.56, 1520.87], [-2521.24, 1846.8], [-6.45, 470.13], [312.63, 457.89], [-522.56, 204.48], [-2521.24, 212.13], [-6.45, 179.83], [312.63, 179.54]]}]\n```" - } - ] - }, - { - "id": 2043, - "data_source": "Mono3DRefer", - "scene_name": "001842", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[616.57, 198.74], [638.85, 198.84], [634.5, 200.77], [610.58, 200.66], [616.57, 177.64], [638.85, 177.66], [634.5, 178.02], [610.58, 178.0]]}, {\"category\": \"car\", \"corners_3d\": [[481.49, 216.03], [451.05, 215.7], [476.89, 211.78], [504.7, 212.04], [481.49, 184.55], [451.05, 184.46], [476.89, 183.4], [504.7, 183.47]]}, {\"category\": \"car\", \"corners_3d\": [[506.6, 210.14], [479.18, 209.92], [500.49, 206.62], [525.57, 206.8], [506.6, 175.54], [479.18, 175.52], [500.49, 175.28], [525.57, 175.3]]}, {\"category\": \"car\", \"corners_3d\": [[657.72, 198.48], [682.86, 198.53], [685.4, 200.76], [658.08, 200.7], [657.72, 174.8], [682.86, 174.8], [685.4, 174.97], [658.08, 174.97]]}, {\"category\": \"car\", \"corners_3d\": [[958.1, 191.98], [947.04, 191.21], [1018.02, 191.38], [1032.2, 192.17], [958.1, 163.99], [947.04, 164.34], [1018.02, 164.26], [1032.2, 163.9]]}]\n```" - } - ] - }, - { - "id": 2044, - "data_source": "Mono3DRefer", - "scene_name": "001856", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[760.3, 295.29], [908.76, 294.81], [1680.69, 595.52], [1171.81, 601.41], [760.3, 88.78], [908.76, 89.11], [1680.69, -117.37], [1171.81, -121.42]]}, {\"category\": \"van\", \"corners_3d\": [[690.57, 180.86], [708.09, 180.96], [689.8, 181.49], [671.37, 181.38], [690.57, 156.73], [708.09, 156.54], [689.8, 155.47], [671.37, 155.68]]}]\n```" - } - ] - }, - { - "id": 2045, - "data_source": "Mono3DRefer", - "scene_name": "001856", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[384.67, 272.95], [489.22, 272.81], [429.18, 325.59], [269.2, 325.9], [384.67, 191.76], [489.22, 191.73], [429.18, 201.7], [269.2, 201.76]]}, {\"category\": \"car\", \"corners_3d\": [[655.29, 212.59], [698.41, 212.5], [721.83, 220.42], [670.14, 220.56], [655.29, 171.18], [698.41, 171.18], [721.83, 170.85], [670.14, 170.84]]}, {\"category\": \"car\", \"corners_3d\": [[477.39, 228.09], [530.55, 228.1], [510.85, 241.77], [444.52, 241.77], [477.39, 178.75], [530.55, 178.75], [510.85, 180.21], [444.52, 180.21]]}, {\"category\": \"car\", \"corners_3d\": [[706.38, 258.85], [792.26, 258.55], [876.45, 293.95], [755.32, 294.55], [706.38, 172.29], [792.26, 172.29], [876.45, 172.06], [755.32, 172.06]]}, {\"category\": \"car\", \"corners_3d\": [[633.63, 195.48], [661.61, 195.55], [661.03, 197.6], [630.54, 197.52], [633.63, 172.0], [661.61, 171.99], [661.03, 171.92], [630.54, 171.92]]}]\n```" - } - ] - }, - { - "id": 2046, - "data_source": "Mono3DRefer", - "scene_name": "001856", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[384.67, 272.95], [489.22, 272.81], [429.18, 325.59], [269.2, 325.9], [384.67, 191.76], [489.22, 191.73], [429.18, 201.7], [269.2, 201.76]]}, {\"category\": \"van\", \"corners_3d\": [[760.3, 295.29], [908.76, 294.81], [1680.69, 595.52], [1171.81, 601.41], [760.3, 88.78], [908.76, 89.11], [1680.69, -117.37], [1171.81, -121.42]]}, {\"category\": \"car\", \"corners_3d\": [[655.29, 212.59], [698.41, 212.5], [721.83, 220.42], [670.14, 220.56], [655.29, 171.18], [698.41, 171.18], [721.83, 170.85], [670.14, 170.84]]}, {\"category\": \"car\", \"corners_3d\": [[477.39, 228.09], [530.55, 228.1], [510.85, 241.77], [444.52, 241.77], [477.39, 178.75], [530.55, 178.75], [510.85, 180.21], [444.52, 180.21]]}, {\"category\": \"car\", \"corners_3d\": [[706.38, 258.85], [792.26, 258.55], [876.45, 293.95], [755.32, 294.55], [706.38, 172.29], [792.26, 172.29], [876.45, 172.06], [755.32, 172.06]]}, {\"category\": \"car\", \"corners_3d\": [[633.63, 195.48], [661.61, 195.55], [661.03, 197.6], [630.54, 197.52], [633.63, 172.0], [661.61, 171.99], [661.03, 171.92], [630.54, 171.92]]}, {\"category\": \"van\", \"corners_3d\": [[690.57, 180.86], [708.09, 180.96], [689.8, 181.49], [671.37, 181.38], [690.57, 156.73], [708.09, 156.54], [689.8, 155.47], [671.37, 155.68]]}]\n```" - } - ] - }, - { - "id": 2047, - "data_source": "Mono3DRefer", - "scene_name": "001860", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-280.04, 320.71], [-125.63, 295.08], [122.14, 295.15], [19.62, 320.81], [-280.04, 176.24], [-125.63, 175.66], [122.14, 175.66], [19.62, 176.25]]}, {\"category\": \"car\", \"corners_3d\": [[-44.76, 290.71], [36.33, 275.58], [255.11, 274.62], [206.78, 289.45], [-44.76, 187.57], [36.33, 185.68], [255.11, 185.56], [206.78, 187.41]]}, {\"category\": \"car\", \"corners_3d\": [[147.41, 259.6], [147.24, 252.0], [327.73, 242.98], [343.08, 248.89], [147.41, 182.96], [147.24, 182.07], [327.73, 181.02], [343.08, 181.71]]}, {\"category\": \"car\", \"corners_3d\": [[202.07, 252.48], [194.11, 246.68], [346.32, 238.78], [363.96, 243.37], [202.07, 181.09], [194.11, 180.49], [346.32, 179.67], [363.96, 180.15]]}, {\"category\": \"car\", \"corners_3d\": [[316.55, 224.25], [312.82, 221.34], [385.75, 219.32], [393.5, 221.98], [316.55, 180.23], [312.82, 179.81], [385.75, 179.52], [393.5, 179.91]]}, {\"category\": \"car\", \"corners_3d\": [[339.48, 221.16], [339.31, 218.32], [439.67, 216.22], [445.81, 218.81], [339.48, 180.06], [339.31, 179.64], [439.67, 179.33], [445.81, 179.71]]}, {\"category\": \"car\", \"corners_3d\": [[485.97, 208.71], [456.72, 208.68], [469.18, 206.0], [496.25, 206.02], [485.97, 179.14], [456.72, 179.14], [469.18, 178.67], [496.25, 178.67]]}]\n```" - } - ] - }, - { - "id": 2048, - "data_source": "Mono3DRefer", - "scene_name": "001864", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[732.56, 297.25], [726.53, 291.39], [801.55, 291.53], [811.56, 297.4], [732.56, 155.2], [726.53, 156.47], [801.55, 156.44], [811.56, 155.17]]}]\n```" - } - ] - }, - { - "id": 2049, - "data_source": "Mono3DRefer", - "scene_name": "001866", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[427.79, 213.85], [433.64, 212.19], [505.71, 211.99], [502.91, 213.64], [427.79, 188.09], [433.64, 187.48], [505.71, 187.4], [502.91, 188.01]]}]\n```" - } - ] - }, - { - "id": 2050, - "data_source": "Mono3DRefer", - "scene_name": "001867", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-437.51, 842.59], [-1126.13, 836.62], [92.94, 372.2], [301.73, 372.73], [-437.51, 207.05], [-1126.13, 206.75], [92.94, 183.03], [301.73, 183.06]]}, {\"category\": \"car\", \"corners_3d\": [[869.58, 380.76], [1084.24, 381.43], [2142.53, 854.2], [1435.41, 847.15], [869.58, 205.3], [1084.24, 205.4], [2142.53, 279.19], [1435.41, 278.09]]}, {\"category\": \"car\", \"corners_3d\": [[824.13, 271.47], [728.67, 271.61], [695.79, 246.02], [766.57, 245.94], [824.13, 186.06], [728.67, 186.08], [695.79, 182.66], [766.57, 182.65]]}, {\"category\": \"car\", \"corners_3d\": [[474.22, 264.46], [382.01, 264.7], [428.73, 244.51], [500.61, 244.36], [474.22, 185.12], [382.01, 185.16], [428.73, 182.45], [500.61, 182.43]]}, {\"category\": \"car\", \"corners_3d\": [[672.49, 224.81], [724.38, 224.78], [748.68, 235.17], [686.42, 235.22], [672.49, 179.42], [724.38, 179.42], [748.68, 180.73], [686.42, 180.74]]}, {\"category\": \"car\", \"corners_3d\": [[719.65, 221.59], [669.37, 221.63], [659.8, 214.73], [702.97, 214.7], [719.65, 176.93], [669.37, 176.93], [659.8, 176.36], [702.97, 176.35]]}, {\"category\": \"car\", \"corners_3d\": [[786.32, 196.44], [777.2, 195.44], [852.31, 195.31], [864.71, 196.31], [786.32, 165.83], [777.2, 166.13], [852.31, 166.16], [864.71, 165.87]]}]\n```" - } - ] - }, - { - "id": 2051, - "data_source": "Mono3DRefer", - "scene_name": "001871", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[554.26, 212.2], [519.9, 212.2], [529.06, 208.15], [559.89, 208.15], [554.26, 183.43], [519.9, 183.43], [529.06, 182.35], [559.89, 182.35]]}]\n```" - } - ] - }, - { - "id": 2052, - "data_source": "Mono3DRefer", - "scene_name": "001875", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[127.85, 297.1], [12.62, 296.92], [175.71, 263.4], [259.95, 263.49], [127.85, 195.92], [12.62, 195.88], [175.71, 189.66], [259.95, 189.68]]}, {\"category\": \"car\", \"corners_3d\": [[520.08, 197.75], [497.81, 197.77], [504.55, 196.03], [525.27, 196.02], [520.08, 177.73], [497.81, 177.73], [504.55, 177.39], [525.27, 177.39]]}, {\"category\": \"car\", \"corners_3d\": [[585.84, 189.01], [601.51, 189.0], [602.17, 189.82], [585.71, 189.83], [585.84, 176.67], [601.51, 176.67], [602.17, 176.86], [585.71, 176.86]]}]\n```" - } - ] - }, - { - "id": 2053, - "data_source": "Mono3DRefer", - "scene_name": "001885", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[248.64, 214.85], [221.4, 214.27], [287.2, 209.81], [312.29, 210.27], [248.64, 181.91], [221.4, 181.78], [287.2, 180.82], [312.29, 180.92]]}, {\"category\": \"car\", \"corners_3d\": [[405.42, 201.81], [390.51, 201.13], [440.13, 199.45], [455.25, 200.05], [405.42, 173.48], [390.51, 173.47], [440.13, 173.43], [455.25, 173.45]]}, {\"category\": \"car\", \"corners_3d\": [[347.38, 200.74], [338.86, 200.01], [394.97, 198.69], [404.52, 199.36], [347.38, 178.0], [338.86, 177.86], [394.97, 177.62], [404.52, 177.74]]}, {\"category\": \"car\", \"corners_3d\": [[502.49, 192.06], [501.6, 191.54], [565.24, 191.22], [567.86, 191.73], [502.49, 170.37], [501.6, 170.43], [565.24, 170.47], [567.86, 170.41]]}]\n```" - } - ] - }, - { - "id": 2054, - "data_source": "Mono3DRefer", - "scene_name": "001887", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[526.66, 226.66], [483.13, 226.67], [499.85, 219.52], [537.6, 219.51], [526.66, 185.78], [483.13, 185.78], [499.85, 184.07], [537.6, 184.07]]}]\n```" - } - ] - }, - { - "id": 2055, - "data_source": "Mono3DRefer", - "scene_name": "001887", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[301.3, 351.31], [139.59, 350.54], [307.84, 288.75], [413.63, 289.08], [301.3, 181.65], [139.59, 181.61], [307.84, 178.56], [413.63, 178.58]]}, {\"category\": \"van\", \"corners_3d\": [[544.36, 212.41], [505.32, 212.39], [518.26, 207.77], [552.74, 207.79], [544.36, 169.78], [505.32, 169.79], [518.26, 170.14], [552.74, 170.14]]}]\n```" - } - ] - }, - { - "id": 2056, - "data_source": "Mono3DRefer", - "scene_name": "001887", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[301.3, 351.31], [139.59, 350.54], [307.84, 288.75], [413.63, 289.08], [301.3, 181.65], [139.59, 181.61], [307.84, 178.56], [413.63, 178.58]]}, {\"category\": \"car\", \"corners_3d\": [[526.66, 226.66], [483.13, 226.67], [499.85, 219.52], [537.6, 219.51], [526.66, 185.78], [483.13, 185.78], [499.85, 184.07], [537.6, 184.07]]}, {\"category\": \"van\", \"corners_3d\": [[544.36, 212.41], [505.32, 212.39], [518.26, 207.77], [552.74, 207.79], [544.36, 169.78], [505.32, 169.79], [518.26, 170.14], [552.74, 170.14]]}]\n```" - } - ] - }, - { - "id": 2057, - "data_source": "Mono3DRefer", - "scene_name": "001889", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[651.76, 188.98], [669.7, 188.96], [675.46, 190.01], [656.35, 190.03], [651.76, 168.47], [669.7, 168.47], [675.46, 168.19], [656.35, 168.18]]}]\n```" - } - ] - }, - { - "id": 2058, - "data_source": "Mono3DRefer", - "scene_name": "001889", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[724.6, 197.08], [741.69, 196.83], [771.65, 198.5], [753.69, 198.79], [724.6, 175.34], [741.69, 175.32], [771.65, 175.49], [753.69, 175.52]]}]\n```" - } - ] - }, - { - "id": 2059, - "data_source": "Mono3DRefer", - "scene_name": "001889", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[651.76, 188.98], [669.7, 188.96], [675.46, 190.01], [656.35, 190.03], [651.76, 168.47], [669.7, 168.47], [675.46, 168.19], [656.35, 168.18]]}, {\"category\": \"car\", \"corners_3d\": [[724.6, 197.08], [741.69, 196.83], [771.65, 198.5], [753.69, 198.79], [724.6, 175.34], [741.69, 175.32], [771.65, 175.49], [753.69, 175.52]]}]\n```" - } - ] - }, - { - "id": 2060, - "data_source": "Mono3DRefer", - "scene_name": "001892", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[425.63, 197.54], [413.7, 197.08], [458.76, 195.76], [470.84, 196.17], [425.63, 177.49], [413.7, 177.4], [458.76, 177.15], [470.84, 177.23]]}, {\"category\": \"car\", \"corners_3d\": [[488.15, 190.96], [486.62, 190.52], [528.68, 190.27], [531.2, 190.69], [488.15, 174.51], [486.62, 174.47], [528.68, 174.44], [531.2, 174.48]]}]\n```" - } - ] - }, - { - "id": 2061, - "data_source": "Mono3DRefer", - "scene_name": "001897", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[390.53, 276.92], [350.87, 277.59], [369.89, 265.95], [405.06, 265.42], [390.53, 178.22], [350.87, 178.25], [369.89, 177.65], [405.06, 177.62]]}]\n```" - } - ] - }, - { - "id": 2062, - "data_source": "Mono3DRefer", - "scene_name": "001897", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[649.93, 253.01], [722.52, 251.45], [795.29, 274.23], [703.0, 276.84], [649.93, 177.82], [722.52, 177.72], [795.29, 179.13], [703.0, 179.29]]}, {\"category\": \"car\", \"corners_3d\": [[550.8, 215.05], [587.95, 214.46], [610.3, 219.7], [568.75, 220.44], [550.8, 180.54], [587.95, 180.43], [610.3, 181.39], [568.75, 181.52]]}, {\"category\": \"car\", \"corners_3d\": [[1267.39, 405.37], [1146.49, 351.09], [1641.08, 371.8], [1955.29, 441.91], [1267.39, 182.98], [1146.49, 180.61], [1641.08, 181.52], [1955.29, 184.57]]}, {\"category\": \"car\", \"corners_3d\": [[510.49, 201.87], [496.51, 201.17], [547.35, 199.54], [561.7, 200.16], [510.49, 174.26], [496.51, 174.23], [547.35, 174.15], [561.7, 174.18]]}, {\"category\": \"car\", \"corners_3d\": [[652.15, 199.48], [674.45, 200.27], [627.47, 202.21], [605.08, 201.31], [652.15, 172.33], [674.45, 172.31], [627.47, 172.27], [605.08, 172.29]]}, {\"category\": \"car\", \"corners_3d\": [[699.55, 196.37], [720.58, 196.92], [680.74, 198.71], [659.17, 198.07], [699.55, 172.04], [720.58, 172.02], [680.74, 171.96], [659.17, 171.98]]}, {\"category\": \"car\", \"corners_3d\": [[745.53, 193.72], [761.96, 194.14], [732.11, 195.19], [715.51, 194.73], [745.53, 172.7], [761.96, 172.69], [732.11, 172.69], [715.51, 172.69]]}]\n```" - } - ] - }, - { - "id": 2063, - "data_source": "Mono3DRefer", - "scene_name": "001897", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[649.93, 253.01], [722.52, 251.45], [795.29, 274.23], [703.0, 276.84], [649.93, 177.82], [722.52, 177.72], [795.29, 179.13], [703.0, 179.29]]}, {\"category\": \"car\", \"corners_3d\": [[550.8, 215.05], [587.95, 214.46], [610.3, 219.7], [568.75, 220.44], [550.8, 180.54], [587.95, 180.43], [610.3, 181.39], [568.75, 181.52]]}, {\"category\": \"car\", \"corners_3d\": [[1267.39, 405.37], [1146.49, 351.09], [1641.08, 371.8], [1955.29, 441.91], [1267.39, 182.98], [1146.49, 180.61], [1641.08, 181.52], [1955.29, 184.57]]}, {\"category\": \"car\", \"corners_3d\": [[510.49, 201.87], [496.51, 201.17], [547.35, 199.54], [561.7, 200.16], [510.49, 174.26], [496.51, 174.23], [547.35, 174.15], [561.7, 174.18]]}, {\"category\": \"cyclist\", \"corners_3d\": [[390.53, 276.92], [350.87, 277.59], [369.89, 265.95], [405.06, 265.42], [390.53, 178.22], [350.87, 178.25], [369.89, 177.65], [405.06, 177.62]]}, {\"category\": \"car\", \"corners_3d\": [[652.15, 199.48], [674.45, 200.27], [627.47, 202.21], [605.08, 201.31], [652.15, 172.33], [674.45, 172.31], [627.47, 172.27], [605.08, 172.29]]}, {\"category\": \"car\", \"corners_3d\": [[699.55, 196.37], [720.58, 196.92], [680.74, 198.71], [659.17, 198.07], [699.55, 172.04], [720.58, 172.02], [680.74, 171.96], [659.17, 171.98]]}, {\"category\": \"car\", \"corners_3d\": [[745.53, 193.72], [761.96, 194.14], [732.11, 195.19], [715.51, 194.73], [745.53, 172.7], [761.96, 172.69], [732.11, 172.69], [715.51, 172.69]]}]\n```" - } - ] - }, - { - "id": 2064, - "data_source": "Mono3DRefer", - "scene_name": "001899", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[571.41, 195.52], [549.74, 195.52], [553.71, 194.01], [573.92, 194.01], [571.41, 174.11], [549.74, 174.11], [553.71, 174.03], [573.92, 174.03]]}]\n```" - } - ] - }, - { - "id": 2065, - "data_source": "Mono3DRefer", - "scene_name": "001907", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[582.05, 245.92], [648.6, 245.92], [657.5, 262.83], [575.55, 262.82], [582.05, 182.64], [648.6, 182.64], [657.5, 184.9], [575.55, 184.9]]}, {\"category\": \"car\", \"corners_3d\": [[59.35, 268.87], [-9.45, 268.79], [98.68, 252.21], [155.67, 252.26], [59.35, 216.66], [-9.45, 216.62], [98.68, 209.06], [155.67, 209.09]]}, {\"category\": \"car\", \"corners_3d\": [[322.12, 234.31], [278.4, 234.28], [326.26, 225.58], [363.8, 225.6], [322.12, 199.19], [278.4, 199.17], [326.26, 195.44], [363.8, 195.46]]}, {\"category\": \"car\", \"corners_3d\": [[1195.49, 254.29], [1147.3, 246.52], [1335.94, 248.51], [1406.07, 256.73], [1195.49, 183.9], [1147.3, 182.85], [1335.94, 183.12], [1406.07, 184.24]]}, {\"category\": \"car\", \"corners_3d\": [[245.27, 234.0], [207.91, 234.03], [249.6, 227.56], [282.99, 227.53], [245.27, 204.04], [207.91, 204.05], [249.6, 200.75], [282.99, 200.74]]}, {\"category\": \"car\", \"corners_3d\": [[398.36, 216.57], [370.28, 216.61], [385.86, 213.57], [411.98, 213.54], [398.36, 192.15], [370.28, 192.16], [385.86, 190.82], [411.98, 190.81]]}]\n```" - } - ] - }, - { - "id": 2066, - "data_source": "Mono3DRefer", - "scene_name": "001911", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[608.29, 305.92], [541.14, 306.41], [544.73, 293.16], [604.83, 292.77], [608.29, 149.39], [541.14, 149.27], [544.73, 152.55], [604.83, 152.65]]}]\n```" - } - ] - }, - { - "id": 2067, - "data_source": "Mono3DRefer", - "scene_name": "001917", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[642.6, 197.04], [664.94, 197.09], [665.8, 198.9], [641.79, 198.84], [642.6, 175.99], [664.94, 176.0], [665.8, 176.23], [641.79, 176.23]]}, {\"category\": \"car\", \"corners_3d\": [[753.61, 258.77], [831.19, 258.96], [901.16, 288.13], [797.19, 287.78], [753.61, 166.45], [831.19, 166.44], [901.16, 164.27], [797.19, 164.29]]}, {\"category\": \"car\", \"corners_3d\": [[834.82, 307.16], [966.54, 310.26], [1152.28, 407.64], [923.7, 398.74], [834.82, 187.22], [966.54, 187.55], [1152.28, 197.97], [923.7, 197.01]]}, {\"category\": \"car\", \"corners_3d\": [[708.41, 225.55], [756.47, 225.69], [777.81, 235.29], [720.98, 235.09], [708.41, 175.49], [756.47, 175.5], [777.81, 175.98], [720.98, 175.97]]}, {\"category\": \"car\", \"corners_3d\": [[695.28, 208.57], [726.29, 208.7], [732.7, 212.64], [698.27, 212.49], [695.28, 180.47], [726.29, 180.5], [732.7, 181.34], [698.27, 181.31]]}]\n```" - } - ] - }, - { - "id": 2068, - "data_source": "Mono3DRefer", - "scene_name": "001932", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[537.1, 216.03], [493.72, 216.0], [509.4, 210.48], [547.23, 210.5], [537.1, 168.66], [493.72, 168.66], [509.4, 169.2], [547.23, 169.19]]}, {\"category\": \"van\", \"corners_3d\": [[103.06, 460.75], [-165.92, 459.76], [201.34, 325.04], [344.46, 325.32], [103.06, 179.3], [-165.92, 179.28], [201.34, 176.26], [344.46, 176.27]]}]\n```" - } - ] - }, - { - "id": 2069, - "data_source": "Mono3DRefer", - "scene_name": "001932", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[516.58, 231.4], [467.05, 231.4], [488.13, 222.71], [530.3, 222.71], [516.58, 184.89], [467.05, 184.89], [488.13, 183.1], [530.3, 183.1]]}, {\"category\": \"car\", \"corners_3d\": [[576.18, 197.15], [548.34, 197.17], [552.28, 195.1], [577.75, 195.08], [576.18, 173.96], [548.34, 173.96], [552.28, 173.87], [577.75, 173.87]]}]\n```" - } - ] - }, - { - "id": 2070, - "data_source": "Mono3DRefer", - "scene_name": "001932", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[516.58, 231.4], [467.05, 231.4], [488.13, 222.71], [530.3, 222.71], [516.58, 184.89], [467.05, 184.89], [488.13, 183.1], [530.3, 183.1]]}, {\"category\": \"van\", \"corners_3d\": [[537.1, 216.03], [493.72, 216.0], [509.4, 210.48], [547.23, 210.5], [537.1, 168.66], [493.72, 168.66], [509.4, 169.2], [547.23, 169.19]]}, {\"category\": \"car\", \"corners_3d\": [[576.18, 197.15], [548.34, 197.17], [552.28, 195.1], [577.75, 195.08], [576.18, 173.96], [548.34, 173.96], [552.28, 173.87], [577.75, 173.87]]}, {\"category\": \"van\", \"corners_3d\": [[103.06, 460.75], [-165.92, 459.76], [201.34, 325.04], [344.46, 325.32], [103.06, 179.3], [-165.92, 179.28], [201.34, 176.26], [344.46, 176.27]]}]\n```" - } - ] - }, - { - "id": 2071, - "data_source": "Mono3DRefer", - "scene_name": "001933", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[587.02, 186.77], [607.09, 186.77], [607.37, 187.69], [585.97, 187.69], [587.02, 170.97], [607.09, 170.97], [607.37, 170.84], [585.97, 170.84]]}, {\"category\": \"car\", \"corners_3d\": [[562.28, 184.0], [579.43, 183.99], [578.93, 184.6], [560.84, 184.61], [562.28, 169.59], [579.43, 169.59], [578.93, 169.41], [560.84, 169.41]]}]\n```" - } - ] - }, - { - "id": 2072, - "data_source": "Mono3DRefer", - "scene_name": "001934", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[583.49, 191.79], [598.87, 191.78], [599.39, 192.72], [583.24, 192.73], [583.49, 179.68], [598.87, 179.67], [599.39, 180.01], [583.24, 180.02]]}, {\"category\": \"car\", \"corners_3d\": [[460.01, 214.42], [428.78, 214.38], [449.54, 209.95], [477.45, 209.98], [460.01, 183.03], [428.78, 183.02], [449.54, 181.94], [477.45, 181.94]]}, {\"category\": \"car\", \"corners_3d\": [[515.15, 196.97], [493.8, 196.97], [501.62, 195.43], [521.6, 195.43], [515.15, 178.35], [493.8, 178.35], [501.62, 178.0], [521.6, 178.0]]}]\n```" - } - ] - }, - { - "id": 2073, - "data_source": "Mono3DRefer", - "scene_name": "001935", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[566.58, 215.38], [611.49, 215.25], [617.22, 221.73], [565.44, 221.91], [566.58, 172.03], [611.49, 172.03], [617.22, 171.91], [565.44, 171.91]]}]\n```" - } - ] - }, - { - "id": 2074, - "data_source": "Mono3DRefer", - "scene_name": "001943", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[581.03, 189.73], [597.5, 189.72], [597.98, 190.62], [580.63, 190.63], [581.03, 176.76], [597.5, 176.76], [597.98, 176.97], [580.63, 176.97]]}, {\"category\": \"car\", \"corners_3d\": [[-230.69, 384.33], [-446.79, 385.01], [-7.93, 296.21], [117.08, 295.99], [-230.69, 192.49], [-446.79, 192.55], [-7.93, 184.31], [117.08, 184.28]]}, {\"category\": \"car\", \"corners_3d\": [[492.69, 204.84], [463.37, 204.84], [475.26, 202.23], [502.19, 202.23], [492.69, 178.39], [463.37, 178.39], [475.26, 177.94], [502.19, 177.94]]}]\n```" - } - ] - }, - { - "id": 2075, - "data_source": "Mono3DRefer", - "scene_name": "001946", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (motorcyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"motorcyclist\", \"corners_3d\": [[512.83, 316.33], [447.16, 314.05], [496.52, 289.65], [551.35, 291.21], [512.83, 150.05], [447.16, 150.42], [496.52, 154.29], [551.35, 154.05]]}]\n```" - } - ] - }, - { - "id": 2076, - "data_source": "Mono3DRefer", - "scene_name": "001946", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[773.72, 153.08], [747.05, 153.35], [759.14, 154.92], [783.79, 154.7], [773.72, 128.38], [747.05, 128.98], [759.14, 132.51], [783.79, 132.01]]}]\n```" - } - ] - }, - { - "id": 2077, - "data_source": "Mono3DRefer", - "scene_name": "001946", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (motorcyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"motorcyclist\", \"corners_3d\": [[512.83, 316.33], [447.16, 314.05], [496.52, 289.65], [551.35, 291.21], [512.83, 150.05], [447.16, 150.42], [496.52, 154.29], [551.35, 154.05]]}, {\"category\": \"car\", \"corners_3d\": [[773.72, 153.08], [747.05, 153.35], [759.14, 154.92], [783.79, 154.7], [773.72, 128.38], [747.05, 128.98], [759.14, 132.51], [783.79, 132.01]]}]\n```" - } - ] - }, - { - "id": 2078, - "data_source": "Mono3DRefer", - "scene_name": "001948", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-167.86, 311.05], [-45.57, 289.54], [220.85, 290.05], [147.3, 311.76], [-167.86, 169.09], [-45.57, 169.68], [220.85, 169.66], [147.3, 169.07]]}, {\"category\": \"car\", \"corners_3d\": [[457.68, 214.2], [449.73, 216.59], [354.99, 216.7], [368.14, 214.3], [457.68, 177.65], [449.73, 177.93], [354.99, 177.94], [368.14, 177.67]]}]\n```" - } - ] - }, - { - "id": 2079, - "data_source": "Mono3DRefer", - "scene_name": "001948", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[983.72, 228.25], [1005.3, 228.25], [1053.38, 234.99], [1029.17, 234.99], [983.72, 151.27], [1005.3, 151.27], [1053.38, 148.64], [1029.17, 148.64]]}, {\"category\": \"cyclist\", \"corners_3d\": [[896.19, 213.42], [911.69, 213.49], [933.54, 217.13], [916.61, 217.05], [896.19, 156.75], [911.69, 156.72], [933.54, 155.28], [916.61, 155.31]]}, {\"category\": \"cyclist\", \"corners_3d\": [[866.03, 208.67], [878.84, 208.77], [889.42, 211.4], [875.64, 211.28], [866.03, 163.68], [878.84, 163.65], [889.42, 162.98], [875.64, 163.01]]}]\n```" - } - ] - }, - { - "id": 2080, - "data_source": "Mono3DRefer", - "scene_name": "001948", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-167.86, 311.05], [-45.57, 289.54], [220.85, 290.05], [147.3, 311.76], [-167.86, 169.09], [-45.57, 169.68], [220.85, 169.66], [147.3, 169.07]]}, {\"category\": \"car\", \"corners_3d\": [[457.68, 214.2], [449.73, 216.59], [354.99, 216.7], [368.14, 214.3], [457.68, 177.65], [449.73, 177.93], [354.99, 177.94], [368.14, 177.67]]}, {\"category\": \"cyclist\", \"corners_3d\": [[983.72, 228.25], [1005.3, 228.25], [1053.38, 234.99], [1029.17, 234.99], [983.72, 151.27], [1005.3, 151.27], [1053.38, 148.64], [1029.17, 148.64]]}, {\"category\": \"cyclist\", \"corners_3d\": [[896.19, 213.42], [911.69, 213.49], [933.54, 217.13], [916.61, 217.05], [896.19, 156.75], [911.69, 156.72], [933.54, 155.28], [916.61, 155.31]]}, {\"category\": \"cyclist\", \"corners_3d\": [[866.03, 208.67], [878.84, 208.77], [889.42, 211.4], [875.64, 211.28], [866.03, 163.68], [878.84, 163.65], [889.42, 162.98], [875.64, 163.01]]}]\n```" - } - ] - }, - { - "id": 2081, - "data_source": "Mono3DRefer", - "scene_name": "001953", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1065.51, 234.94], [1004.05, 231.27], [1057.97, 224.91], [1115.36, 227.8], [1065.51, 185.68], [1004.05, 184.92], [1057.97, 183.61], [1115.36, 184.2]]}, {\"category\": \"car\", \"corners_3d\": [[1146.66, 221.96], [1101.61, 219.99], [1133.87, 216.16], [1176.36, 217.82], [1146.66, 184.22], [1101.61, 183.77], [1133.87, 182.88], [1176.36, 183.26]]}, {\"category\": \"car\", \"corners_3d\": [[976.74, 245.31], [916.06, 240.75], [989.77, 231.89], [1046.34, 235.31], [976.74, 183.67], [916.06, 182.99], [989.77, 181.67], [1046.34, 182.18]]}]\n```" - } - ] - }, - { - "id": 2082, - "data_source": "Mono3DRefer", - "scene_name": "001957", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[214.82, 256.57], [247.25, 249.13], [399.02, 248.48], [381.58, 255.79], [214.82, 192.06], [247.25, 190.35], [399.02, 190.2], [381.58, 191.88]]}]\n```" - } - ] - }, - { - "id": 2083, - "data_source": "Mono3DRefer", - "scene_name": "001957", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[616.43, 222.43], [634.88, 222.39], [639.25, 227.06], [619.06, 227.1], [616.43, 172.26], [634.88, 172.26], [639.25, 172.2], [619.06, 172.2]]}]\n```" - } - ] - }, - { - "id": 2084, - "data_source": "Mono3DRefer", - "scene_name": "001957", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[616.43, 222.43], [634.88, 222.39], [639.25, 227.06], [619.06, 227.1], [616.43, 172.26], [634.88, 172.26], [639.25, 172.2], [619.06, 172.2]]}, {\"category\": \"car\", \"corners_3d\": [[214.82, 256.57], [247.25, 249.13], [399.02, 248.48], [381.58, 255.79], [214.82, 192.06], [247.25, 190.35], [399.02, 190.2], [381.58, 191.88]]}]\n```" - } - ] - }, - { - "id": 2085, - "data_source": "Mono3DRefer", - "scene_name": "001960", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[456.73, 262.18], [364.01, 262.87], [413.54, 240.76], [483.34, 240.36], [456.73, 181.71], [364.01, 181.78], [413.54, 179.59], [483.34, 179.55]]}, {\"category\": \"car\", \"corners_3d\": [[699.78, 221.88], [652.01, 221.95], [643.94, 215.06], [685.01, 215.01], [699.78, 172.84], [652.01, 172.84], [643.94, 172.84], [685.01, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[533.21, 211.5], [493.32, 211.44], [510.93, 206.49], [545.72, 206.54], [533.21, 176.69], [493.32, 176.68], [510.93, 176.19], [545.72, 176.19]]}, {\"category\": \"car\", \"corners_3d\": [[628.81, 208.33], [661.84, 208.18], [675.48, 212.31], [638.63, 212.49], [628.81, 177.47], [661.84, 177.45], [675.48, 177.98], [638.63, 178.01]]}]\n```" - } - ] - }, - { - "id": 2086, - "data_source": "Mono3DRefer", - "scene_name": "001961", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[737.95, 198.89], [758.74, 198.47], [803.59, 200.7], [781.75, 201.2], [737.95, 174.36], [758.74, 174.34], [803.59, 174.47], [781.75, 174.5]]}]\n```" - } - ] - }, - { - "id": 2087, - "data_source": "Mono3DRefer", - "scene_name": "001966", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[468.17, 219.03], [519.9, 218.97], [504.56, 228.3], [442.31, 228.4], [468.17, 165.9], [519.9, 165.91], [504.56, 164.5], [442.31, 164.49]]}, {\"category\": \"van\", \"corners_3d\": [[674.75, 203.71], [716.67, 203.69], [733.96, 208.38], [685.68, 208.4], [674.75, 164.41], [716.67, 164.42], [733.96, 163.13], [685.68, 163.13]]}, {\"category\": \"van\", \"corners_3d\": [[316.12, 278.01], [452.81, 277.63], [316.51, 377.58], [47.52, 379.03], [316.12, 139.28], [452.81, 139.4], [316.51, 107.48], [47.52, 107.02]]}]\n```" - } - ] - }, - { - "id": 2088, - "data_source": "Mono3DRefer", - "scene_name": "001966", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[732.34, 247.6], [845.44, 247.61], [2420.51, 748.03], [1549.54, 747.48], [732.34, 96.22], [845.44, 96.21], [2420.51, -416.78], [1549.54, -416.22]]}]\n```" - } - ] - }, - { - "id": 2089, - "data_source": "Mono3DRefer", - "scene_name": "001966", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[425.4, 239.68], [496.77, 239.68], [461.05, 260.74], [367.2, 260.73], [425.4, 179.44], [496.77, 179.44], [461.05, 181.52], [367.2, 181.52]]}, {\"category\": \"car\", \"corners_3d\": [[702.56, 214.2], [747.34, 214.17], [774.14, 221.84], [721.05, 221.87], [702.56, 175.82], [747.34, 175.82], [774.14, 176.37], [721.05, 176.37]]}, {\"category\": \"car\", \"corners_3d\": [[577.25, 299.94], [700.67, 300.17], [758.48, 388.51], [549.51, 387.84], [577.25, 185.15], [700.67, 185.17], [758.48, 193.72], [549.51, 193.66]]}]\n```" - } - ] - }, - { - "id": 2090, - "data_source": "Mono3DRefer", - "scene_name": "001966", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[468.17, 219.03], [519.9, 218.97], [504.56, 228.3], [442.31, 228.4], [468.17, 165.9], [519.9, 165.91], [504.56, 164.5], [442.31, 164.49]]}, {\"category\": \"van\", \"corners_3d\": [[674.75, 203.71], [716.67, 203.69], [733.96, 208.38], [685.68, 208.4], [674.75, 164.41], [716.67, 164.42], [733.96, 163.13], [685.68, 163.13]]}, {\"category\": \"truck\", \"corners_3d\": [[732.34, 247.6], [845.44, 247.61], [2420.51, 748.03], [1549.54, 747.48], [732.34, 96.22], [845.44, 96.21], [2420.51, -416.78], [1549.54, -416.22]]}, {\"category\": \"van\", \"corners_3d\": [[316.12, 278.01], [452.81, 277.63], [316.51, 377.58], [47.52, 379.03], [316.12, 139.28], [452.81, 139.4], [316.51, 107.48], [47.52, 107.02]]}]\n```" - } - ] - }, - { - "id": 2091, - "data_source": "Mono3DRefer", - "scene_name": "001966", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[425.4, 239.68], [496.77, 239.68], [461.05, 260.74], [367.2, 260.73], [425.4, 179.44], [496.77, 179.44], [461.05, 181.52], [367.2, 181.52]]}, {\"category\": \"van\", \"corners_3d\": [[468.17, 219.03], [519.9, 218.97], [504.56, 228.3], [442.31, 228.4], [468.17, 165.9], [519.9, 165.91], [504.56, 164.5], [442.31, 164.49]]}, {\"category\": \"car\", \"corners_3d\": [[702.56, 214.2], [747.34, 214.17], [774.14, 221.84], [721.05, 221.87], [702.56, 175.82], [747.34, 175.82], [774.14, 176.37], [721.05, 176.37]]}, {\"category\": \"van\", \"corners_3d\": [[674.75, 203.71], [716.67, 203.69], [733.96, 208.38], [685.68, 208.4], [674.75, 164.41], [716.67, 164.42], [733.96, 163.13], [685.68, 163.13]]}, {\"category\": \"car\", \"corners_3d\": [[577.25, 299.94], [700.67, 300.17], [758.48, 388.51], [549.51, 387.84], [577.25, 185.15], [700.67, 185.17], [758.48, 193.72], [549.51, 193.66]]}, {\"category\": \"van\", \"corners_3d\": [[316.12, 278.01], [452.81, 277.63], [316.51, 377.58], [47.52, 379.03], [316.12, 139.28], [452.81, 139.4], [316.51, 107.48], [47.52, 107.02]]}]\n```" - } - ] - }, - { - "id": 2092, - "data_source": "Mono3DRefer", - "scene_name": "001966", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[425.4, 239.68], [496.77, 239.68], [461.05, 260.74], [367.2, 260.73], [425.4, 179.44], [496.77, 179.44], [461.05, 181.52], [367.2, 181.52]]}, {\"category\": \"car\", \"corners_3d\": [[702.56, 214.2], [747.34, 214.17], [774.14, 221.84], [721.05, 221.87], [702.56, 175.82], [747.34, 175.82], [774.14, 176.37], [721.05, 176.37]]}, {\"category\": \"car\", \"corners_3d\": [[577.25, 299.94], [700.67, 300.17], [758.48, 388.51], [549.51, 387.84], [577.25, 185.15], [700.67, 185.17], [758.48, 193.72], [549.51, 193.66]]}, {\"category\": \"truck\", \"corners_3d\": [[732.34, 247.6], [845.44, 247.61], [2420.51, 748.03], [1549.54, 747.48], [732.34, 96.22], [845.44, 96.21], [2420.51, -416.78], [1549.54, -416.22]]}]\n```" - } - ] - }, - { - "id": 2093, - "data_source": "Mono3DRefer", - "scene_name": "001966", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[425.4, 239.68], [496.77, 239.68], [461.05, 260.74], [367.2, 260.73], [425.4, 179.44], [496.77, 179.44], [461.05, 181.52], [367.2, 181.52]]}, {\"category\": \"van\", \"corners_3d\": [[468.17, 219.03], [519.9, 218.97], [504.56, 228.3], [442.31, 228.4], [468.17, 165.9], [519.9, 165.91], [504.56, 164.5], [442.31, 164.49]]}, {\"category\": \"car\", \"corners_3d\": [[702.56, 214.2], [747.34, 214.17], [774.14, 221.84], [721.05, 221.87], [702.56, 175.82], [747.34, 175.82], [774.14, 176.37], [721.05, 176.37]]}, {\"category\": \"van\", \"corners_3d\": [[674.75, 203.71], [716.67, 203.69], [733.96, 208.38], [685.68, 208.4], [674.75, 164.41], [716.67, 164.42], [733.96, 163.13], [685.68, 163.13]]}, {\"category\": \"car\", \"corners_3d\": [[577.25, 299.94], [700.67, 300.17], [758.48, 388.51], [549.51, 387.84], [577.25, 185.15], [700.67, 185.17], [758.48, 193.72], [549.51, 193.66]]}, {\"category\": \"truck\", \"corners_3d\": [[732.34, 247.6], [845.44, 247.61], [2420.51, 748.03], [1549.54, 747.48], [732.34, 96.22], [845.44, 96.21], [2420.51, -416.78], [1549.54, -416.22]]}, {\"category\": \"van\", \"corners_3d\": [[316.12, 278.01], [452.81, 277.63], [316.51, 377.58], [47.52, 379.03], [316.12, 139.28], [452.81, 139.4], [316.51, 107.48], [47.52, 107.02]]}]\n```" - } - ] - }, - { - "id": 2094, - "data_source": "Mono3DRefer", - "scene_name": "001970", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[258.83, 300.2], [232.28, 309.6], [122.25, 309.8], [156.86, 300.37], [258.83, 146.59], [232.28, 143.92], [122.25, 143.87], [156.86, 146.54]]}]\n```" - } - ] - }, - { - "id": 2095, - "data_source": "Mono3DRefer", - "scene_name": "001971", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[745.72, 326.07], [685.95, 328.97], [663.29, 314.99], [717.93, 312.61], [745.72, 131.42], [685.95, 130.44], [663.29, 135.16], [717.93, 135.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[809.21, 319.9], [753.03, 322.04], [734.18, 312.44], [786.86, 310.58], [809.21, 150.64], [753.03, 150.18], [734.18, 152.24], [786.86, 152.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[932.5, 284.0], [903.93, 286.2], [862.99, 278.99], [890.45, 277.07], [932.5, 153.22], [903.93, 152.64], [862.99, 154.54], [890.45, 155.05]]}]\n```" - } - ] - }, - { - "id": 2096, - "data_source": "Mono3DRefer", - "scene_name": "001974", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[884.93, 203.61], [896.21, 203.61], [914.61, 205.59], [902.59, 205.59], [884.93, 157.97], [896.21, 157.97], [914.61, 157.02], [902.59, 157.02]]}]\n```" - } - ] - }, - { - "id": 2097, - "data_source": "Mono3DRefer", - "scene_name": "001974", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[637.31, 185.88], [657.2, 185.84], [707.3, 190.69], [680.18, 190.77], [637.31, 157.52], [657.2, 157.57], [707.3, 151.86], [680.18, 151.76]]}]\n```" - } - ] - }, - { - "id": 2098, - "data_source": "Mono3DRefer", - "scene_name": "001974", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-24.21, 323.32], [-145.92, 323.34], [73.92, 279.51], [160.17, 279.5], [-24.21, 210.83], [-145.92, 210.83], [73.92, 199.77], [160.17, 199.77]]}, {\"category\": \"car\", \"corners_3d\": [[398.24, 225.24], [356.86, 225.19], [393.13, 218.07], [428.91, 218.11], [398.24, 185.07], [356.86, 185.06], [393.13, 183.4], [428.91, 183.41]]}]\n```" - } - ] - }, - { - "id": 2099, - "data_source": "Mono3DRefer", - "scene_name": "001974", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[637.31, 185.88], [657.2, 185.84], [707.3, 190.69], [680.18, 190.77], [637.31, 157.52], [657.2, 157.57], [707.3, 151.86], [680.18, 151.76]]}, {\"category\": \"cyclist\", \"corners_3d\": [[884.93, 203.61], [896.21, 203.61], [914.61, 205.59], [902.59, 205.59], [884.93, 157.97], [896.21, 157.97], [914.61, 157.02], [902.59, 157.02]]}]\n```" - } - ] - }, - { - "id": 2100, - "data_source": "Mono3DRefer", - "scene_name": "001974", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-24.21, 323.32], [-145.92, 323.34], [73.92, 279.51], [160.17, 279.5], [-24.21, 210.83], [-145.92, 210.83], [73.92, 199.77], [160.17, 199.77]]}, {\"category\": \"car\", \"corners_3d\": [[398.24, 225.24], [356.86, 225.19], [393.13, 218.07], [428.91, 218.11], [398.24, 185.07], [356.86, 185.06], [393.13, 183.4], [428.91, 183.41]]}, {\"category\": \"cyclist\", \"corners_3d\": [[884.93, 203.61], [896.21, 203.61], [914.61, 205.59], [902.59, 205.59], [884.93, 157.97], [896.21, 157.97], [914.61, 157.02], [902.59, 157.02]]}]\n```" - } - ] - }, - { - "id": 2101, - "data_source": "Mono3DRefer", - "scene_name": "001974", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[637.31, 185.88], [657.2, 185.84], [707.3, 190.69], [680.18, 190.77], [637.31, 157.52], [657.2, 157.57], [707.3, 151.86], [680.18, 151.76]]}, {\"category\": \"car\", \"corners_3d\": [[-24.21, 323.32], [-145.92, 323.34], [73.92, 279.51], [160.17, 279.5], [-24.21, 210.83], [-145.92, 210.83], [73.92, 199.77], [160.17, 199.77]]}, {\"category\": \"car\", \"corners_3d\": [[398.24, 225.24], [356.86, 225.19], [393.13, 218.07], [428.91, 218.11], [398.24, 185.07], [356.86, 185.06], [393.13, 183.4], [428.91, 183.41]]}]\n```" - } - ] - }, - { - "id": 2102, - "data_source": "Mono3DRefer", - "scene_name": "001974", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[637.31, 185.88], [657.2, 185.84], [707.3, 190.69], [680.18, 190.77], [637.31, 157.52], [657.2, 157.57], [707.3, 151.86], [680.18, 151.76]]}, {\"category\": \"car\", \"corners_3d\": [[-24.21, 323.32], [-145.92, 323.34], [73.92, 279.51], [160.17, 279.5], [-24.21, 210.83], [-145.92, 210.83], [73.92, 199.77], [160.17, 199.77]]}, {\"category\": \"car\", \"corners_3d\": [[398.24, 225.24], [356.86, 225.19], [393.13, 218.07], [428.91, 218.11], [398.24, 185.07], [356.86, 185.06], [393.13, 183.4], [428.91, 183.41]]}, {\"category\": \"cyclist\", \"corners_3d\": [[884.93, 203.61], [896.21, 203.61], [914.61, 205.59], [902.59, 205.59], [884.93, 157.97], [896.21, 157.97], [914.61, 157.02], [902.59, 157.02]]}]\n```" - } - ] - }, - { - "id": 2103, - "data_source": "Mono3DRefer", - "scene_name": "001975", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[653.23, 193.49], [633.75, 193.37], [642.64, 192.03], [660.9, 192.14], [653.23, 176.51], [633.75, 176.48], [642.64, 176.25], [660.9, 176.27]]}]\n```" - } - ] - }, - { - "id": 2104, - "data_source": "Mono3DRefer", - "scene_name": "001977", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[102.81, 277.32], [174.49, 271.26], [217.45, 286.49], [136.96, 294.64], [102.81, 216.08], [174.49, 213.58], [217.45, 219.88], [136.96, 223.26]]}, {\"category\": \"car\", \"corners_3d\": [[32.55, 261.22], [83.61, 257.84], [102.51, 267.97], [45.95, 272.21], [32.55, 225.82], [83.61, 223.79], [102.51, 229.86], [45.95, 232.41]]}, {\"category\": \"car\", \"corners_3d\": [[368.25, 339.25], [483.08, 320.46], [687.34, 375.82], [559.07, 413.12], [368.25, 213.79], [483.08, 209.17], [687.34, 222.79], [559.07, 231.96]]}, {\"category\": \"car\", \"corners_3d\": [[239.68, 298.4], [338.26, 287.59], [463.38, 323.5], [345.95, 342.69], [239.68, 218.01], [338.26, 214.13], [463.38, 227.04], [345.95, 233.94]]}, {\"category\": \"car\", \"corners_3d\": [[172.71, 285.6], [249.97, 278.42], [314.54, 300.3], [225.32, 310.92], [172.71, 220.9], [249.97, 217.84], [314.54, 227.16], [225.32, 231.69]]}]\n```" - } - ] - }, - { - "id": 2105, - "data_source": "Mono3DRefer", - "scene_name": "001983", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[647.85, 214.84], [688.41, 214.77], [706.79, 222.47], [658.79, 222.57], [647.85, 178.51], [688.41, 178.5], [706.79, 179.54], [658.79, 179.55]]}]\n```" - } - ] - }, - { - "id": 2106, - "data_source": "Mono3DRefer", - "scene_name": "002009", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[905.02, 246.78], [868.18, 238.42], [1069.86, 237.64], [1131.61, 245.8], [905.02, 170.23], [868.18, 170.53], [1069.86, 170.56], [1131.61, 170.27]]}, {\"category\": \"car\", \"corners_3d\": [[565.61, 205.59], [588.06, 205.5], [592.11, 208.57], [567.55, 208.68], [565.61, 184.5], [588.06, 184.47], [592.11, 185.56], [567.55, 185.6]]}, {\"category\": \"car\", \"corners_3d\": [[500.4, 210.69], [473.75, 210.78], [482.04, 207.35], [506.28, 207.28], [500.4, 188.11], [473.75, 188.15], [482.04, 186.77], [506.28, 186.74]]}]\n```" - } - ] - }, - { - "id": 2107, - "data_source": "Mono3DRefer", - "scene_name": "002016", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[124.61, 289.18], [30.99, 289.04], [184.79, 258.51], [253.92, 258.58], [124.61, 198.25], [30.99, 198.22], [184.79, 191.55], [253.92, 191.57]]}]\n```" - } - ] - }, - { - "id": 2108, - "data_source": "Mono3DRefer", - "scene_name": "002016", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[977.06, 216.18], [992.32, 216.18], [1026.31, 220.03], [1009.69, 220.03], [977.06, 154.47], [992.32, 154.47], [1026.31, 152.83], [1009.69, 152.83]]}]\n```" - } - ] - }, - { - "id": 2109, - "data_source": "Mono3DRefer", - "scene_name": "002016", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[621.71, 187.56], [641.93, 187.51], [690.0, 193.09], [662.28, 193.19], [621.71, 158.75], [641.93, 158.8], [690.0, 153.44], [662.28, 153.34]]}]\n```" - } - ] - }, - { - "id": 2110, - "data_source": "Mono3DRefer", - "scene_name": "002016", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[977.06, 216.18], [992.32, 216.18], [1026.31, 220.03], [1009.69, 220.03], [977.06, 154.47], [992.32, 154.47], [1026.31, 152.83], [1009.69, 152.83]]}, {\"category\": \"car\", \"corners_3d\": [[124.61, 289.18], [30.99, 289.04], [184.79, 258.51], [253.92, 258.58], [124.61, 198.25], [30.99, 198.22], [184.79, 191.55], [253.92, 191.57]]}]\n```" - } - ] - }, - { - "id": 2111, - "data_source": "Mono3DRefer", - "scene_name": "002016", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[621.71, 187.56], [641.93, 187.51], [690.0, 193.09], [662.28, 193.19], [621.71, 158.75], [641.93, 158.8], [690.0, 153.44], [662.28, 153.34]]}, {\"category\": \"car\", \"corners_3d\": [[124.61, 289.18], [30.99, 289.04], [184.79, 258.51], [253.92, 258.58], [124.61, 198.25], [30.99, 198.22], [184.79, 191.55], [253.92, 191.57]]}]\n```" - } - ] - }, - { - "id": 2112, - "data_source": "Mono3DRefer", - "scene_name": "002016", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[977.06, 216.18], [992.32, 216.18], [1026.31, 220.03], [1009.69, 220.03], [977.06, 154.47], [992.32, 154.47], [1026.31, 152.83], [1009.69, 152.83]]}, {\"category\": \"tram\", \"corners_3d\": [[621.71, 187.56], [641.93, 187.51], [690.0, 193.09], [662.28, 193.19], [621.71, 158.75], [641.93, 158.8], [690.0, 153.44], [662.28, 153.34]]}]\n```" - } - ] - }, - { - "id": 2113, - "data_source": "Mono3DRefer", - "scene_name": "002016", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[977.06, 216.18], [992.32, 216.18], [1026.31, 220.03], [1009.69, 220.03], [977.06, 154.47], [992.32, 154.47], [1026.31, 152.83], [1009.69, 152.83]]}, {\"category\": \"tram\", \"corners_3d\": [[621.71, 187.56], [641.93, 187.51], [690.0, 193.09], [662.28, 193.19], [621.71, 158.75], [641.93, 158.8], [690.0, 153.44], [662.28, 153.34]]}, {\"category\": \"car\", \"corners_3d\": [[124.61, 289.18], [30.99, 289.04], [184.79, 258.51], [253.92, 258.58], [124.61, 198.25], [30.99, 198.22], [184.79, 191.55], [253.92, 191.57]]}]\n```" - } - ] - }, - { - "id": 2114, - "data_source": "Mono3DRefer", - "scene_name": "002024", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[539.02, 258.57], [580.17, 257.85], [591.22, 269.73], [544.38, 270.67], [539.02, 170.82], [580.17, 170.84], [591.22, 170.55], [544.38, 170.53]]}]\n```" - } - ] - }, - { - "id": 2115, - "data_source": "Mono3DRefer", - "scene_name": "002024", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[754.28, 318.43], [893.65, 315.28], [1244.37, 452.8], [980.02, 465.25], [754.28, 183.02], [893.65, 182.8], [1244.37, 192.41], [980.02, 193.28]]}, {\"category\": \"car\", \"corners_3d\": [[73.8, 374.69], [-122.22, 378.38], [138.02, 299.8], [257.02, 298.38], [73.8, 211.01], [-122.22, 211.71], [138.02, 196.85], [257.02, 196.58]]}, {\"category\": \"car\", \"corners_3d\": [[644.86, 252.79], [727.28, 251.78], [789.61, 276.77], [681.7, 278.53], [644.86, 173.33], [727.28, 173.33], [789.61, 173.48], [681.7, 173.49]]}, {\"category\": \"car\", \"corners_3d\": [[602.1, 232.28], [663.77, 231.52], [702.41, 245.45], [626.47, 246.62], [602.1, 176.85], [663.77, 176.8], [702.41, 177.74], [626.47, 177.82]]}, {\"category\": \"car\", \"corners_3d\": [[361.87, 254.18], [290.53, 255.41], [327.21, 240.42], [385.3, 239.59], [361.87, 193.49], [290.53, 193.8], [327.21, 190.0], [385.3, 189.79]]}, {\"category\": \"car\", \"corners_3d\": [[390.1, 233.94], [340.27, 234.51], [363.64, 225.89], [406.37, 225.46], [390.1, 188.99], [340.27, 189.14], [363.64, 186.86], [406.37, 186.75]]}, {\"category\": \"car\", \"corners_3d\": [[582.37, 222.21], [625.59, 221.77], [645.47, 229.58], [595.48, 230.17], [582.37, 182.43], [625.59, 182.34], [645.47, 183.86], [595.48, 183.97]]}]\n```" - } - ] - }, - { - "id": 2116, - "data_source": "Mono3DRefer", - "scene_name": "002024", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[510.97, 201.68], [540.9, 201.53], [543.67, 204.69], [510.43, 204.88], [510.97, 168.92], [540.9, 168.94], [543.67, 168.51], [510.43, 168.48]]}, {\"category\": \"van\", \"corners_3d\": [[281.56, 289.67], [170.56, 291.57], [267.26, 259.57], [347.55, 258.55], [281.56, 181.4], [170.56, 181.54], [267.26, 179.19], [347.55, 179.12]]}]\n```" - } - ] - }, - { - "id": 2117, - "data_source": "Mono3DRefer", - "scene_name": "002024", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[910.99, 254.74], [866.95, 255.32], [844.74, 249.9], [886.05, 249.39], [910.99, 157.26], [866.95, 157.15], [844.74, 158.18], [886.05, 158.28]]}]\n```" - } - ] - }, - { - "id": 2118, - "data_source": "Mono3DRefer", - "scene_name": "002024", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[539.02, 258.57], [580.17, 257.85], [591.22, 269.73], [544.38, 270.67], [539.02, 170.82], [580.17, 170.84], [591.22, 170.55], [544.38, 170.53]]}, {\"category\": \"car\", \"corners_3d\": [[754.28, 318.43], [893.65, 315.28], [1244.37, 452.8], [980.02, 465.25], [754.28, 183.02], [893.65, 182.8], [1244.37, 192.41], [980.02, 193.28]]}, {\"category\": \"car\", \"corners_3d\": [[73.8, 374.69], [-122.22, 378.38], [138.02, 299.8], [257.02, 298.38], [73.8, 211.01], [-122.22, 211.71], [138.02, 196.85], [257.02, 196.58]]}, {\"category\": \"car\", \"corners_3d\": [[644.86, 252.79], [727.28, 251.78], [789.61, 276.77], [681.7, 278.53], [644.86, 173.33], [727.28, 173.33], [789.61, 173.48], [681.7, 173.49]]}, {\"category\": \"car\", \"corners_3d\": [[602.1, 232.28], [663.77, 231.52], [702.41, 245.45], [626.47, 246.62], [602.1, 176.85], [663.77, 176.8], [702.41, 177.74], [626.47, 177.82]]}, {\"category\": \"car\", \"corners_3d\": [[361.87, 254.18], [290.53, 255.41], [327.21, 240.42], [385.3, 239.59], [361.87, 193.49], [290.53, 193.8], [327.21, 190.0], [385.3, 189.79]]}, {\"category\": \"car\", \"corners_3d\": [[390.1, 233.94], [340.27, 234.51], [363.64, 225.89], [406.37, 225.46], [390.1, 188.99], [340.27, 189.14], [363.64, 186.86], [406.37, 186.75]]}, {\"category\": \"car\", \"corners_3d\": [[582.37, 222.21], [625.59, 221.77], [645.47, 229.58], [595.48, 230.17], [582.37, 182.43], [625.59, 182.34], [645.47, 183.86], [595.48, 183.97]]}]\n```" - } - ] - }, - { - "id": 2119, - "data_source": "Mono3DRefer", - "scene_name": "002024", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[510.97, 201.68], [540.9, 201.53], [543.67, 204.69], [510.43, 204.88], [510.97, 168.92], [540.9, 168.94], [543.67, 168.51], [510.43, 168.48]]}, {\"category\": \"cyclist\", \"corners_3d\": [[539.02, 258.57], [580.17, 257.85], [591.22, 269.73], [544.38, 270.67], [539.02, 170.82], [580.17, 170.84], [591.22, 170.55], [544.38, 170.53]]}, {\"category\": \"van\", \"corners_3d\": [[281.56, 289.67], [170.56, 291.57], [267.26, 259.57], [347.55, 258.55], [281.56, 181.4], [170.56, 181.54], [267.26, 179.19], [347.55, 179.12]]}]\n```" - } - ] - }, - { - "id": 2120, - "data_source": "Mono3DRefer", - "scene_name": "002024", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[539.02, 258.57], [580.17, 257.85], [591.22, 269.73], [544.38, 270.67], [539.02, 170.82], [580.17, 170.84], [591.22, 170.55], [544.38, 170.53]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[910.99, 254.74], [866.95, 255.32], [844.74, 249.9], [886.05, 249.39], [910.99, 157.26], [866.95, 157.15], [844.74, 158.18], [886.05, 158.28]]}]\n```" - } - ] - }, - { - "id": 2121, - "data_source": "Mono3DRefer", - "scene_name": "002024", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[510.97, 201.68], [540.9, 201.53], [543.67, 204.69], [510.43, 204.88], [510.97, 168.92], [540.9, 168.94], [543.67, 168.51], [510.43, 168.48]]}, {\"category\": \"car\", \"corners_3d\": [[754.28, 318.43], [893.65, 315.28], [1244.37, 452.8], [980.02, 465.25], [754.28, 183.02], [893.65, 182.8], [1244.37, 192.41], [980.02, 193.28]]}, {\"category\": \"car\", \"corners_3d\": [[73.8, 374.69], [-122.22, 378.38], [138.02, 299.8], [257.02, 298.38], [73.8, 211.01], [-122.22, 211.71], [138.02, 196.85], [257.02, 196.58]]}, {\"category\": \"van\", \"corners_3d\": [[281.56, 289.67], [170.56, 291.57], [267.26, 259.57], [347.55, 258.55], [281.56, 181.4], [170.56, 181.54], [267.26, 179.19], [347.55, 179.12]]}, {\"category\": \"car\", \"corners_3d\": [[644.86, 252.79], [727.28, 251.78], [789.61, 276.77], [681.7, 278.53], [644.86, 173.33], [727.28, 173.33], [789.61, 173.48], [681.7, 173.49]]}, {\"category\": \"car\", \"corners_3d\": [[602.1, 232.28], [663.77, 231.52], [702.41, 245.45], [626.47, 246.62], [602.1, 176.85], [663.77, 176.8], [702.41, 177.74], [626.47, 177.82]]}, {\"category\": \"car\", \"corners_3d\": [[361.87, 254.18], [290.53, 255.41], [327.21, 240.42], [385.3, 239.59], [361.87, 193.49], [290.53, 193.8], [327.21, 190.0], [385.3, 189.79]]}, {\"category\": \"car\", \"corners_3d\": [[390.1, 233.94], [340.27, 234.51], [363.64, 225.89], [406.37, 225.46], [390.1, 188.99], [340.27, 189.14], [363.64, 186.86], [406.37, 186.75]]}, {\"category\": \"car\", \"corners_3d\": [[582.37, 222.21], [625.59, 221.77], [645.47, 229.58], [595.48, 230.17], [582.37, 182.43], [625.59, 182.34], [645.47, 183.86], [595.48, 183.97]]}]\n```" - } - ] - }, - { - "id": 2122, - "data_source": "Mono3DRefer", - "scene_name": "002024", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[754.28, 318.43], [893.65, 315.28], [1244.37, 452.8], [980.02, 465.25], [754.28, 183.02], [893.65, 182.8], [1244.37, 192.41], [980.02, 193.28]]}, {\"category\": \"car\", \"corners_3d\": [[73.8, 374.69], [-122.22, 378.38], [138.02, 299.8], [257.02, 298.38], [73.8, 211.01], [-122.22, 211.71], [138.02, 196.85], [257.02, 196.58]]}, {\"category\": \"car\", \"corners_3d\": [[644.86, 252.79], [727.28, 251.78], [789.61, 276.77], [681.7, 278.53], [644.86, 173.33], [727.28, 173.33], [789.61, 173.48], [681.7, 173.49]]}, {\"category\": \"car\", \"corners_3d\": [[602.1, 232.28], [663.77, 231.52], [702.41, 245.45], [626.47, 246.62], [602.1, 176.85], [663.77, 176.8], [702.41, 177.74], [626.47, 177.82]]}, {\"category\": \"car\", \"corners_3d\": [[361.87, 254.18], [290.53, 255.41], [327.21, 240.42], [385.3, 239.59], [361.87, 193.49], [290.53, 193.8], [327.21, 190.0], [385.3, 189.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[910.99, 254.74], [866.95, 255.32], [844.74, 249.9], [886.05, 249.39], [910.99, 157.26], [866.95, 157.15], [844.74, 158.18], [886.05, 158.28]]}, {\"category\": \"car\", \"corners_3d\": [[390.1, 233.94], [340.27, 234.51], [363.64, 225.89], [406.37, 225.46], [390.1, 188.99], [340.27, 189.14], [363.64, 186.86], [406.37, 186.75]]}, {\"category\": \"car\", \"corners_3d\": [[582.37, 222.21], [625.59, 221.77], [645.47, 229.58], [595.48, 230.17], [582.37, 182.43], [625.59, 182.34], [645.47, 183.86], [595.48, 183.97]]}]\n```" - } - ] - }, - { - "id": 2123, - "data_source": "Mono3DRefer", - "scene_name": "002024", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[510.97, 201.68], [540.9, 201.53], [543.67, 204.69], [510.43, 204.88], [510.97, 168.92], [540.9, 168.94], [543.67, 168.51], [510.43, 168.48]]}, {\"category\": \"van\", \"corners_3d\": [[281.56, 289.67], [170.56, 291.57], [267.26, 259.57], [347.55, 258.55], [281.56, 181.4], [170.56, 181.54], [267.26, 179.19], [347.55, 179.12]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[910.99, 254.74], [866.95, 255.32], [844.74, 249.9], [886.05, 249.39], [910.99, 157.26], [866.95, 157.15], [844.74, 158.18], [886.05, 158.28]]}]\n```" - } - ] - }, - { - "id": 2124, - "data_source": "Mono3DRefer", - "scene_name": "002024", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[510.97, 201.68], [540.9, 201.53], [543.67, 204.69], [510.43, 204.88], [510.97, 168.92], [540.9, 168.94], [543.67, 168.51], [510.43, 168.48]]}, {\"category\": \"cyclist\", \"corners_3d\": [[539.02, 258.57], [580.17, 257.85], [591.22, 269.73], [544.38, 270.67], [539.02, 170.82], [580.17, 170.84], [591.22, 170.55], [544.38, 170.53]]}, {\"category\": \"car\", \"corners_3d\": [[754.28, 318.43], [893.65, 315.28], [1244.37, 452.8], [980.02, 465.25], [754.28, 183.02], [893.65, 182.8], [1244.37, 192.41], [980.02, 193.28]]}, {\"category\": \"car\", \"corners_3d\": [[73.8, 374.69], [-122.22, 378.38], [138.02, 299.8], [257.02, 298.38], [73.8, 211.01], [-122.22, 211.71], [138.02, 196.85], [257.02, 196.58]]}, {\"category\": \"van\", \"corners_3d\": [[281.56, 289.67], [170.56, 291.57], [267.26, 259.57], [347.55, 258.55], [281.56, 181.4], [170.56, 181.54], [267.26, 179.19], [347.55, 179.12]]}, {\"category\": \"car\", \"corners_3d\": [[644.86, 252.79], [727.28, 251.78], [789.61, 276.77], [681.7, 278.53], [644.86, 173.33], [727.28, 173.33], [789.61, 173.48], [681.7, 173.49]]}, {\"category\": \"car\", \"corners_3d\": [[602.1, 232.28], [663.77, 231.52], [702.41, 245.45], [626.47, 246.62], [602.1, 176.85], [663.77, 176.8], [702.41, 177.74], [626.47, 177.82]]}, {\"category\": \"car\", \"corners_3d\": [[361.87, 254.18], [290.53, 255.41], [327.21, 240.42], [385.3, 239.59], [361.87, 193.49], [290.53, 193.8], [327.21, 190.0], [385.3, 189.79]]}, {\"category\": \"car\", \"corners_3d\": [[390.1, 233.94], [340.27, 234.51], [363.64, 225.89], [406.37, 225.46], [390.1, 188.99], [340.27, 189.14], [363.64, 186.86], [406.37, 186.75]]}, {\"category\": \"car\", \"corners_3d\": [[582.37, 222.21], [625.59, 221.77], [645.47, 229.58], [595.48, 230.17], [582.37, 182.43], [625.59, 182.34], [645.47, 183.86], [595.48, 183.97]]}]\n```" - } - ] - }, - { - "id": 2125, - "data_source": "Mono3DRefer", - "scene_name": "002024", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[539.02, 258.57], [580.17, 257.85], [591.22, 269.73], [544.38, 270.67], [539.02, 170.82], [580.17, 170.84], [591.22, 170.55], [544.38, 170.53]]}, {\"category\": \"car\", \"corners_3d\": [[754.28, 318.43], [893.65, 315.28], [1244.37, 452.8], [980.02, 465.25], [754.28, 183.02], [893.65, 182.8], [1244.37, 192.41], [980.02, 193.28]]}, {\"category\": \"car\", \"corners_3d\": [[73.8, 374.69], [-122.22, 378.38], [138.02, 299.8], [257.02, 298.38], [73.8, 211.01], [-122.22, 211.71], [138.02, 196.85], [257.02, 196.58]]}, {\"category\": \"car\", \"corners_3d\": [[644.86, 252.79], [727.28, 251.78], [789.61, 276.77], [681.7, 278.53], [644.86, 173.33], [727.28, 173.33], [789.61, 173.48], [681.7, 173.49]]}, {\"category\": \"car\", \"corners_3d\": [[602.1, 232.28], [663.77, 231.52], [702.41, 245.45], [626.47, 246.62], [602.1, 176.85], [663.77, 176.8], [702.41, 177.74], [626.47, 177.82]]}, {\"category\": \"car\", \"corners_3d\": [[361.87, 254.18], [290.53, 255.41], [327.21, 240.42], [385.3, 239.59], [361.87, 193.49], [290.53, 193.8], [327.21, 190.0], [385.3, 189.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[910.99, 254.74], [866.95, 255.32], [844.74, 249.9], [886.05, 249.39], [910.99, 157.26], [866.95, 157.15], [844.74, 158.18], [886.05, 158.28]]}, {\"category\": \"car\", \"corners_3d\": [[390.1, 233.94], [340.27, 234.51], [363.64, 225.89], [406.37, 225.46], [390.1, 188.99], [340.27, 189.14], [363.64, 186.86], [406.37, 186.75]]}, {\"category\": \"car\", \"corners_3d\": [[582.37, 222.21], [625.59, 221.77], [645.47, 229.58], [595.48, 230.17], [582.37, 182.43], [625.59, 182.34], [645.47, 183.86], [595.48, 183.97]]}]\n```" - } - ] - }, - { - "id": 2126, - "data_source": "Mono3DRefer", - "scene_name": "002024", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[510.97, 201.68], [540.9, 201.53], [543.67, 204.69], [510.43, 204.88], [510.97, 168.92], [540.9, 168.94], [543.67, 168.51], [510.43, 168.48]]}, {\"category\": \"cyclist\", \"corners_3d\": [[539.02, 258.57], [580.17, 257.85], [591.22, 269.73], [544.38, 270.67], [539.02, 170.82], [580.17, 170.84], [591.22, 170.55], [544.38, 170.53]]}, {\"category\": \"van\", \"corners_3d\": [[281.56, 289.67], [170.56, 291.57], [267.26, 259.57], [347.55, 258.55], [281.56, 181.4], [170.56, 181.54], [267.26, 179.19], [347.55, 179.12]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[910.99, 254.74], [866.95, 255.32], [844.74, 249.9], [886.05, 249.39], [910.99, 157.26], [866.95, 157.15], [844.74, 158.18], [886.05, 158.28]]}]\n```" - } - ] - }, - { - "id": 2127, - "data_source": "Mono3DRefer", - "scene_name": "002024", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[510.97, 201.68], [540.9, 201.53], [543.67, 204.69], [510.43, 204.88], [510.97, 168.92], [540.9, 168.94], [543.67, 168.51], [510.43, 168.48]]}, {\"category\": \"car\", \"corners_3d\": [[754.28, 318.43], [893.65, 315.28], [1244.37, 452.8], [980.02, 465.25], [754.28, 183.02], [893.65, 182.8], [1244.37, 192.41], [980.02, 193.28]]}, {\"category\": \"car\", \"corners_3d\": [[73.8, 374.69], [-122.22, 378.38], [138.02, 299.8], [257.02, 298.38], [73.8, 211.01], [-122.22, 211.71], [138.02, 196.85], [257.02, 196.58]]}, {\"category\": \"van\", \"corners_3d\": [[281.56, 289.67], [170.56, 291.57], [267.26, 259.57], [347.55, 258.55], [281.56, 181.4], [170.56, 181.54], [267.26, 179.19], [347.55, 179.12]]}, {\"category\": \"car\", \"corners_3d\": [[644.86, 252.79], [727.28, 251.78], [789.61, 276.77], [681.7, 278.53], [644.86, 173.33], [727.28, 173.33], [789.61, 173.48], [681.7, 173.49]]}, {\"category\": \"car\", \"corners_3d\": [[602.1, 232.28], [663.77, 231.52], [702.41, 245.45], [626.47, 246.62], [602.1, 176.85], [663.77, 176.8], [702.41, 177.74], [626.47, 177.82]]}, {\"category\": \"car\", \"corners_3d\": [[361.87, 254.18], [290.53, 255.41], [327.21, 240.42], [385.3, 239.59], [361.87, 193.49], [290.53, 193.8], [327.21, 190.0], [385.3, 189.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[910.99, 254.74], [866.95, 255.32], [844.74, 249.9], [886.05, 249.39], [910.99, 157.26], [866.95, 157.15], [844.74, 158.18], [886.05, 158.28]]}, {\"category\": \"car\", \"corners_3d\": [[390.1, 233.94], [340.27, 234.51], [363.64, 225.89], [406.37, 225.46], [390.1, 188.99], [340.27, 189.14], [363.64, 186.86], [406.37, 186.75]]}, {\"category\": \"car\", \"corners_3d\": [[582.37, 222.21], [625.59, 221.77], [645.47, 229.58], [595.48, 230.17], [582.37, 182.43], [625.59, 182.34], [645.47, 183.86], [595.48, 183.97]]}]\n```" - } - ] - }, - { - "id": 2128, - "data_source": "Mono3DRefer", - "scene_name": "002024", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[510.97, 201.68], [540.9, 201.53], [543.67, 204.69], [510.43, 204.88], [510.97, 168.92], [540.9, 168.94], [543.67, 168.51], [510.43, 168.48]]}, {\"category\": \"cyclist\", \"corners_3d\": [[539.02, 258.57], [580.17, 257.85], [591.22, 269.73], [544.38, 270.67], [539.02, 170.82], [580.17, 170.84], [591.22, 170.55], [544.38, 170.53]]}, {\"category\": \"car\", \"corners_3d\": [[754.28, 318.43], [893.65, 315.28], [1244.37, 452.8], [980.02, 465.25], [754.28, 183.02], [893.65, 182.8], [1244.37, 192.41], [980.02, 193.28]]}, {\"category\": \"car\", \"corners_3d\": [[73.8, 374.69], [-122.22, 378.38], [138.02, 299.8], [257.02, 298.38], [73.8, 211.01], [-122.22, 211.71], [138.02, 196.85], [257.02, 196.58]]}, {\"category\": \"van\", \"corners_3d\": [[281.56, 289.67], [170.56, 291.57], [267.26, 259.57], [347.55, 258.55], [281.56, 181.4], [170.56, 181.54], [267.26, 179.19], [347.55, 179.12]]}, {\"category\": \"car\", \"corners_3d\": [[644.86, 252.79], [727.28, 251.78], [789.61, 276.77], [681.7, 278.53], [644.86, 173.33], [727.28, 173.33], [789.61, 173.48], [681.7, 173.49]]}, {\"category\": \"car\", \"corners_3d\": [[602.1, 232.28], [663.77, 231.52], [702.41, 245.45], [626.47, 246.62], [602.1, 176.85], [663.77, 176.8], [702.41, 177.74], [626.47, 177.82]]}, {\"category\": \"car\", \"corners_3d\": [[361.87, 254.18], [290.53, 255.41], [327.21, 240.42], [385.3, 239.59], [361.87, 193.49], [290.53, 193.8], [327.21, 190.0], [385.3, 189.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[910.99, 254.74], [866.95, 255.32], [844.74, 249.9], [886.05, 249.39], [910.99, 157.26], [866.95, 157.15], [844.74, 158.18], [886.05, 158.28]]}, {\"category\": \"car\", \"corners_3d\": [[390.1, 233.94], [340.27, 234.51], [363.64, 225.89], [406.37, 225.46], [390.1, 188.99], [340.27, 189.14], [363.64, 186.86], [406.37, 186.75]]}, {\"category\": \"car\", \"corners_3d\": [[582.37, 222.21], [625.59, 221.77], [645.47, 229.58], [595.48, 230.17], [582.37, 182.43], [625.59, 182.34], [645.47, 183.86], [595.48, 183.97]]}]\n```" - } - ] - }, - { - "id": 2129, - "data_source": "Mono3DRefer", - "scene_name": "002029", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[652.4, 215.19], [693.65, 215.17], [711.24, 223.37], [662.0, 223.4], [652.4, 165.55], [693.65, 165.55], [711.24, 164.13], [662.0, 164.13]]}]\n```" - } - ] - }, - { - "id": 2130, - "data_source": "Mono3DRefer", - "scene_name": "002029", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[686.88, 237.92], [739.76, 237.92], [768.02, 252.11], [703.61, 252.1], [686.88, 186.74], [739.76, 186.74], [768.02, 189.77], [703.61, 189.76]]}, {\"category\": \"car\", \"corners_3d\": [[670.29, 226.42], [715.84, 226.35], [736.08, 235.37], [682.88, 235.46], [670.29, 183.72], [715.84, 183.71], [736.08, 185.54], [682.88, 185.56]]}, {\"category\": \"car\", \"corners_3d\": [[561.92, 212.78], [530.9, 212.88], [534.34, 208.92], [562.29, 208.83], [561.92, 181.49], [530.9, 181.51], [534.34, 180.66], [562.29, 180.64]]}]\n```" - } - ] - }, - { - "id": 2131, - "data_source": "Mono3DRefer", - "scene_name": "002029", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[686.88, 237.92], [739.76, 237.92], [768.02, 252.11], [703.61, 252.1], [686.88, 186.74], [739.76, 186.74], [768.02, 189.77], [703.61, 189.76]]}, {\"category\": \"car\", \"corners_3d\": [[670.29, 226.42], [715.84, 226.35], [736.08, 235.37], [682.88, 235.46], [670.29, 183.72], [715.84, 183.71], [736.08, 185.54], [682.88, 185.56]]}, {\"category\": \"van\", \"corners_3d\": [[652.4, 215.19], [693.65, 215.17], [711.24, 223.37], [662.0, 223.4], [652.4, 165.55], [693.65, 165.55], [711.24, 164.13], [662.0, 164.13]]}, {\"category\": \"car\", \"corners_3d\": [[561.92, 212.78], [530.9, 212.88], [534.34, 208.92], [562.29, 208.83], [561.92, 181.49], [530.9, 181.51], [534.34, 180.66], [562.29, 180.64]]}]\n```" - } - ] - }, - { - "id": 2132, - "data_source": "Mono3DRefer", - "scene_name": "002031", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[400.12, 215.21], [408.17, 213.23], [491.47, 213.0], [487.53, 214.95], [400.12, 185.27], [408.17, 184.69], [491.47, 184.62], [487.53, 185.2]]}]\n```" - } - ] - }, - { - "id": 2133, - "data_source": "Mono3DRefer", - "scene_name": "002034", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[598.33, 229.97], [621.66, 229.86], [626.32, 234.7], [601.01, 234.83], [598.33, 203.59], [621.66, 203.53], [626.32, 206.14], [601.01, 206.2]]}]\n```" - } - ] - }, - { - "id": 2134, - "data_source": "Mono3DRefer", - "scene_name": "002034", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[490.19, 287.27], [383.56, 287.62], [438.95, 257.49], [517.53, 257.3], [490.19, 198.87], [383.56, 198.95], [438.95, 192.1], [517.53, 192.05]]}]\n```" - } - ] - }, - { - "id": 2135, - "data_source": "Mono3DRefer", - "scene_name": "002034", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[490.19, 287.27], [383.56, 287.62], [438.95, 257.49], [517.53, 257.3], [490.19, 198.87], [383.56, 198.95], [438.95, 192.1], [517.53, 192.05]]}, {\"category\": \"van\", \"corners_3d\": [[598.33, 229.97], [621.66, 229.86], [626.32, 234.7], [601.01, 234.83], [598.33, 203.59], [621.66, 203.53], [626.32, 206.14], [601.01, 206.2]]}]\n```" - } - ] - }, - { - "id": 2136, - "data_source": "Mono3DRefer", - "scene_name": "002036", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[812.01, 149.74], [784.99, 150.09], [797.23, 151.89], [822.26, 151.59], [812.01, 124.98], [784.99, 125.7], [797.23, 129.43], [822.26, 128.82]]}]\n```" - } - ] - }, - { - "id": 2137, - "data_source": "Mono3DRefer", - "scene_name": "002036", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (motorcyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"motorcyclist\", \"corners_3d\": [[559.47, 302.7], [499.34, 300.13], [543.81, 279.81], [594.93, 281.62], [559.47, 149.04], [499.34, 149.51], [543.81, 153.24], [594.93, 152.91]]}]\n```" - } - ] - }, - { - "id": 2138, - "data_source": "Mono3DRefer", - "scene_name": "002036", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, motorcyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"motorcyclist\", \"corners_3d\": [[559.47, 302.7], [499.34, 300.13], [543.81, 279.81], [594.93, 281.62], [559.47, 149.04], [499.34, 149.51], [543.81, 153.24], [594.93, 152.91]]}, {\"category\": \"car\", \"corners_3d\": [[812.01, 149.74], [784.99, 150.09], [797.23, 151.89], [822.26, 151.59], [812.01, 124.98], [784.99, 125.7], [797.23, 129.43], [822.26, 128.82]]}]\n```" - } - ] - }, - { - "id": 2139, - "data_source": "Mono3DRefer", - "scene_name": "002037", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[182.6, 238.49], [130.33, 238.5], [187.02, 230.72], [233.1, 230.72], [182.6, 192.96], [130.33, 192.96], [187.02, 190.58], [233.1, 190.58]]}, {\"category\": \"car\", \"corners_3d\": [[399.09, 205.91], [375.74, 205.89], [391.65, 203.77], [413.51, 203.79], [399.09, 187.74], [375.74, 187.73], [391.65, 186.78], [413.51, 186.78]]}]\n```" - } - ] - }, - { - "id": 2140, - "data_source": "Mono3DRefer", - "scene_name": "002043", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[293.07, 227.4], [314.16, 223.94], [411.36, 224.09], [396.82, 227.57], [293.07, 177.21], [314.16, 176.93], [411.36, 176.94], [396.82, 177.22]]}, {\"category\": \"car\", \"corners_3d\": [[742.86, 204.94], [708.05, 205.04], [692.41, 201.5], [723.45, 201.43], [742.86, 168.35], [708.05, 168.34], [692.41, 168.83], [723.45, 168.84]]}, {\"category\": \"car\", \"corners_3d\": [[582.18, 207.84], [552.81, 207.9], [555.54, 204.4], [581.97, 204.35], [582.18, 179.33], [552.81, 179.35], [555.54, 178.7], [581.97, 178.69]]}]\n```" - } - ] - }, - { - "id": 2141, - "data_source": "Mono3DRefer", - "scene_name": "002047", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[292.23, 201.81], [268.26, 201.84], [291.09, 199.76], [313.33, 199.74], [292.23, 181.15], [268.26, 181.15], [291.09, 180.56], [313.33, 180.55]]}, {\"category\": \"car\", \"corners_3d\": [[813.59, 298.24], [934.89, 297.62], [1230.31, 404.02], [1007.34, 406.16], [813.59, 185.86], [934.89, 185.79], [1230.31, 196.82], [1007.34, 197.05]]}, {\"category\": \"car\", \"corners_3d\": [[262.28, 205.41], [236.26, 205.46], [263.71, 202.86], [287.61, 202.82], [262.28, 181.12], [236.26, 181.13], [263.71, 180.47], [287.61, 180.46]]}]\n```" - } - ] - }, - { - "id": 2142, - "data_source": "Mono3DRefer", - "scene_name": "002048", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[862.8, 209.43], [899.59, 209.93], [909.05, 214.67], [867.48, 214.03], [862.8, 169.78], [899.59, 169.74], [909.05, 169.34], [867.48, 169.4]]}]\n```" - } - ] - }, - { - "id": 2143, - "data_source": "Mono3DRefer", - "scene_name": "002048", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[888.81, 234.37], [944.02, 235.46], [982.22, 251.45], [912.39, 249.73], [888.81, 184.09], [944.02, 184.29], [982.22, 187.21], [912.39, 186.89]]}, {\"category\": \"car\", \"corners_3d\": [[336.53, 271.79], [389.02, 261.83], [541.68, 269.07], [503.05, 280.83], [336.53, 178.13], [389.02, 177.6], [541.68, 177.99], [503.05, 178.62]]}]\n```" - } - ] - }, - { - "id": 2144, - "data_source": "Mono3DRefer", - "scene_name": "002048", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[445.67, 282.98], [476.76, 280.23], [491.16, 284.65], [459.14, 287.64], [445.67, 175.39], [476.76, 175.33], [491.16, 175.43], [459.14, 175.5]]}]\n```" - } - ] - }, - { - "id": 2145, - "data_source": "Mono3DRefer", - "scene_name": "002048", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[888.81, 234.37], [944.02, 235.46], [982.22, 251.45], [912.39, 249.73], [888.81, 184.09], [944.02, 184.29], [982.22, 187.21], [912.39, 186.89]]}, {\"category\": \"car\", \"corners_3d\": [[336.53, 271.79], [389.02, 261.83], [541.68, 269.07], [503.05, 280.83], [336.53, 178.13], [389.02, 177.6], [541.68, 177.99], [503.05, 178.62]]}, {\"category\": \"van\", \"corners_3d\": [[862.8, 209.43], [899.59, 209.93], [909.05, 214.67], [867.48, 214.03], [862.8, 169.78], [899.59, 169.74], [909.05, 169.34], [867.48, 169.4]]}]\n```" - } - ] - }, - { - "id": 2146, - "data_source": "Mono3DRefer", - "scene_name": "002048", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[445.67, 282.98], [476.76, 280.23], [491.16, 284.65], [459.14, 287.64], [445.67, 175.39], [476.76, 175.33], [491.16, 175.43], [459.14, 175.5]]}, {\"category\": \"van\", \"corners_3d\": [[862.8, 209.43], [899.59, 209.93], [909.05, 214.67], [867.48, 214.03], [862.8, 169.78], [899.59, 169.74], [909.05, 169.34], [867.48, 169.4]]}]\n```" - } - ] - }, - { - "id": 2147, - "data_source": "Mono3DRefer", - "scene_name": "002048", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[888.81, 234.37], [944.02, 235.46], [982.22, 251.45], [912.39, 249.73], [888.81, 184.09], [944.02, 184.29], [982.22, 187.21], [912.39, 186.89]]}, {\"category\": \"car\", \"corners_3d\": [[336.53, 271.79], [389.02, 261.83], [541.68, 269.07], [503.05, 280.83], [336.53, 178.13], [389.02, 177.6], [541.68, 177.99], [503.05, 178.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[445.67, 282.98], [476.76, 280.23], [491.16, 284.65], [459.14, 287.64], [445.67, 175.39], [476.76, 175.33], [491.16, 175.43], [459.14, 175.5]]}]\n```" - } - ] - }, - { - "id": 2148, - "data_source": "Mono3DRefer", - "scene_name": "002048", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[888.81, 234.37], [944.02, 235.46], [982.22, 251.45], [912.39, 249.73], [888.81, 184.09], [944.02, 184.29], [982.22, 187.21], [912.39, 186.89]]}, {\"category\": \"car\", \"corners_3d\": [[336.53, 271.79], [389.02, 261.83], [541.68, 269.07], [503.05, 280.83], [336.53, 178.13], [389.02, 177.6], [541.68, 177.99], [503.05, 178.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[445.67, 282.98], [476.76, 280.23], [491.16, 284.65], [459.14, 287.64], [445.67, 175.39], [476.76, 175.33], [491.16, 175.43], [459.14, 175.5]]}, {\"category\": \"van\", \"corners_3d\": [[862.8, 209.43], [899.59, 209.93], [909.05, 214.67], [867.48, 214.03], [862.8, 169.78], [899.59, 169.74], [909.05, 169.34], [867.48, 169.4]]}]\n```" - } - ] - }, - { - "id": 2149, - "data_source": "Mono3DRefer", - "scene_name": "002053", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[557.02, 207.74], [531.38, 207.9], [530.08, 205.18], [553.74, 205.04], [557.02, 181.45], [531.38, 181.49], [530.08, 180.82], [553.74, 180.79]]}]\n```" - } - ] - }, - { - "id": 2150, - "data_source": "Mono3DRefer", - "scene_name": "002054", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[918.39, 313.67], [781.19, 312.38], [740.09, 270.24], [835.33, 270.87], [918.39, 188.0], [781.19, 187.87], [740.09, 183.33], [835.33, 183.4]]}, {\"category\": \"car\", \"corners_3d\": [[619.81, 205.3], [587.1, 205.2], [595.11, 201.62], [624.21, 201.7], [619.81, 175.87], [587.1, 175.86], [595.11, 175.53], [624.21, 175.54]]}, {\"category\": \"car\", \"corners_3d\": [[400.26, 219.64], [413.76, 216.64], [531.07, 216.65], [525.61, 219.65], [400.26, 176.38], [413.76, 176.15], [531.07, 176.15], [525.61, 176.38]]}]\n```" - } - ] - }, - { - "id": 2151, - "data_source": "Mono3DRefer", - "scene_name": "002054", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[592.43, 215.62], [544.14, 215.36], [560.89, 209.88], [603.02, 210.08], [592.43, 166.47], [544.14, 166.51], [560.89, 167.32], [603.02, 167.3]]}, {\"category\": \"van\", \"corners_3d\": [[693.37, 211.04], [732.22, 211.1], [746.41, 216.5], [702.05, 216.42], [693.37, 169.05], [732.22, 169.04], [746.41, 168.51], [702.05, 168.51]]}]\n```" - } - ] - }, - { - "id": 2152, - "data_source": "Mono3DRefer", - "scene_name": "002054", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[918.39, 313.67], [781.19, 312.38], [740.09, 270.24], [835.33, 270.87], [918.39, 188.0], [781.19, 187.87], [740.09, 183.33], [835.33, 183.4]]}, {\"category\": \"van\", \"corners_3d\": [[592.43, 215.62], [544.14, 215.36], [560.89, 209.88], [603.02, 210.08], [592.43, 166.47], [544.14, 166.51], [560.89, 167.32], [603.02, 167.3]]}, {\"category\": \"van\", \"corners_3d\": [[693.37, 211.04], [732.22, 211.1], [746.41, 216.5], [702.05, 216.42], [693.37, 169.05], [732.22, 169.04], [746.41, 168.51], [702.05, 168.51]]}, {\"category\": \"car\", \"corners_3d\": [[619.81, 205.3], [587.1, 205.2], [595.11, 201.62], [624.21, 201.7], [619.81, 175.87], [587.1, 175.86], [595.11, 175.53], [624.21, 175.54]]}, {\"category\": \"car\", \"corners_3d\": [[400.26, 219.64], [413.76, 216.64], [531.07, 216.65], [525.61, 219.65], [400.26, 176.38], [413.76, 176.15], [531.07, 176.15], [525.61, 176.38]]}]\n```" - } - ] - }, - { - "id": 2153, - "data_source": "Mono3DRefer", - "scene_name": "002055", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[571.06, 197.67], [548.8, 197.68], [552.43, 195.98], [573.16, 195.97], [571.06, 175.69], [548.8, 175.69], [552.43, 175.5], [573.16, 175.5]]}]\n```" - } - ] - }, - { - "id": 2154, - "data_source": "Mono3DRefer", - "scene_name": "002057", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[898.45, 263.55], [788.78, 265.01], [716.64, 238.65], [796.11, 237.9], [898.45, 145.88], [788.78, 145.44], [716.64, 153.28], [796.11, 153.51]]}, {\"category\": \"car\", \"corners_3d\": [[397.65, 282.66], [304.24, 283.65], [369.63, 255.14], [438.73, 254.6], [397.65, 194.19], [304.24, 194.38], [369.63, 188.84], [438.73, 188.74]]}, {\"category\": \"car\", \"corners_3d\": [[480.12, 231.21], [424.94, 231.61], [442.65, 222.86], [489.56, 222.57], [480.12, 186.55], [424.94, 186.65], [442.65, 184.59], [489.56, 184.53]]}, {\"category\": \"car\", \"corners_3d\": [[500.43, 217.85], [455.4, 218.05], [468.55, 212.28], [507.8, 212.13], [500.43, 173.92], [455.4, 173.93], [468.55, 173.79], [507.8, 173.79]]}, {\"category\": \"car\", \"corners_3d\": [[580.97, 194.56], [605.0, 194.51], [607.83, 196.11], [582.02, 196.16], [580.97, 171.59], [605.0, 171.59], [607.83, 171.5], [582.02, 171.5]]}]\n```" - } - ] - }, - { - "id": 2155, - "data_source": "Mono3DRefer", - "scene_name": "002061", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[462.2, 252.17], [380.91, 253.23], [410.29, 236.49], [474.52, 235.83], [462.2, 170.9], [380.91, 170.88], [410.29, 171.29], [474.52, 171.3]]}, {\"category\": \"car\", \"corners_3d\": [[478.88, 212.17], [435.66, 212.79], [432.27, 207.66], [470.06, 207.19], [478.88, 175.81], [435.66, 175.86], [432.27, 175.47], [470.06, 175.44]]}]\n```" - } - ] - }, - { - "id": 2156, - "data_source": "Mono3DRefer", - "scene_name": "002062", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[388.1, 312.62], [243.13, 312.92], [359.82, 267.34], [457.52, 267.21], [388.1, 182.49], [243.13, 182.51], [359.82, 179.37], [457.52, 179.36]]}, {\"category\": \"car\", \"corners_3d\": [[718.95, 267.53], [814.84, 267.54], [907.3, 310.31], [768.09, 310.29], [718.95, 188.62], [814.84, 188.62], [907.3, 195.74], [768.09, 195.73]]}, {\"category\": \"car\", \"corners_3d\": [[740.4, 233.5], [679.5, 233.6], [663.99, 222.52], [713.82, 222.44], [740.4, 178.98], [679.5, 178.99], [663.99, 177.87], [713.82, 177.87]]}, {\"category\": \"car\", \"corners_3d\": [[513.88, 228.85], [452.97, 229.04], [472.89, 220.23], [524.23, 220.1], [513.88, 176.53], [452.97, 176.54], [472.89, 175.96], [524.23, 175.95]]}, {\"category\": \"car\", \"corners_3d\": [[654.61, 211.0], [694.49, 210.98], [707.95, 216.58], [662.21, 216.61], [654.61, 176.13], [694.49, 176.13], [707.95, 176.61], [662.21, 176.61]]}]\n```" - } - ] - }, - { - "id": 2157, - "data_source": "Mono3DRefer", - "scene_name": "002066", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[575.52, 198.78], [554.32, 198.73], [560.81, 197.02], [580.61, 197.06], [575.52, 180.91], [554.32, 180.9], [560.81, 180.37], [580.61, 180.38]]}]\n```" - } - ] - }, - { - "id": 2158, - "data_source": "Mono3DRefer", - "scene_name": "002069", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[564.62, 188.76], [549.08, 188.77], [551.11, 187.93], [565.82, 187.92], [564.62, 174.07], [549.08, 174.07], [551.11, 174.0], [565.82, 174.0]]}, {\"category\": \"car\", \"corners_3d\": [[549.28, 198.47], [525.34, 198.47], [531.87, 196.47], [553.94, 196.47], [549.28, 173.16], [525.34, 173.16], [531.87, 173.13], [553.94, 173.13]]}]\n```" - } - ] - }, - { - "id": 2159, - "data_source": "Mono3DRefer", - "scene_name": "002075", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[459.92, 212.98], [442.65, 212.97], [445.54, 212.32], [462.54, 212.34], [459.92, 179.18], [442.65, 179.17], [445.54, 179.07], [462.54, 179.07]]}]\n```" - } - ] - }, - { - "id": 2160, - "data_source": "Mono3DRefer", - "scene_name": "002080", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[300.51, 242.31], [333.77, 236.27], [491.86, 237.74], [473.19, 244.08], [300.51, 181.52], [333.77, 180.77], [491.86, 180.95], [473.19, 181.74]]}, {\"category\": \"car\", \"corners_3d\": [[381.37, 225.28], [392.61, 221.64], [525.1, 220.84], [523.77, 224.36], [381.37, 180.61], [392.61, 180.07], [525.1, 179.95], [523.77, 180.47]]}, {\"category\": \"car\", \"corners_3d\": [[972.6, 248.24], [1017.62, 257.78], [832.15, 257.56], [808.02, 248.07], [972.6, 169.25], [1017.62, 168.79], [832.15, 168.8], [808.02, 169.25]]}, {\"category\": \"car\", \"corners_3d\": [[1075.64, 262.47], [1153.69, 277.71], [865.14, 277.3], [829.16, 262.17], [1075.64, 170.39], [1153.69, 169.97], [865.14, 169.99], [829.16, 170.4]]}, {\"category\": \"car\", \"corners_3d\": [[370.21, 324.72], [292.4, 363.54], [-148.03, 352.6], [11.68, 317.7], [370.21, 202.14], [292.4, 209.62], [-148.03, 207.51], [11.68, 200.78]]}, {\"category\": \"car\", \"corners_3d\": [[1934.89, 456.71], [3201.56, 737.17], [1482.91, 670.76], [1096.34, 438.86], [1934.89, 170.85], [3201.56, 168.88], [1482.91, 169.34], [1096.34, 170.98]]}, {\"category\": \"car\", \"corners_3d\": [[1321.15, 329.26], [1538.18, 376.96], [958.86, 376.96], [877.22, 329.26], [1321.15, 152.5], [1538.18, 146.29], [958.86, 146.29], [877.22, 152.5]]}]\n```" - } - ] - }, - { - "id": 2161, - "data_source": "Mono3DRefer", - "scene_name": "002080", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[527.26, 208.38], [541.29, 208.04], [551.53, 209.22], [537.13, 209.58], [527.26, 166.1], [541.29, 166.16], [551.53, 165.94], [537.13, 165.87]]}]\n```" - } - ] - }, - { - "id": 2162, - "data_source": "Mono3DRefer", - "scene_name": "002080", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[300.51, 242.31], [333.77, 236.27], [491.86, 237.74], [473.19, 244.08], [300.51, 181.52], [333.77, 180.77], [491.86, 180.95], [473.19, 181.74]]}, {\"category\": \"car\", \"corners_3d\": [[381.37, 225.28], [392.61, 221.64], [525.1, 220.84], [523.77, 224.36], [381.37, 180.61], [392.61, 180.07], [525.1, 179.95], [523.77, 180.47]]}, {\"category\": \"car\", \"corners_3d\": [[972.6, 248.24], [1017.62, 257.78], [832.15, 257.56], [808.02, 248.07], [972.6, 169.25], [1017.62, 168.79], [832.15, 168.8], [808.02, 169.25]]}, {\"category\": \"car\", \"corners_3d\": [[1075.64, 262.47], [1153.69, 277.71], [865.14, 277.3], [829.16, 262.17], [1075.64, 170.39], [1153.69, 169.97], [865.14, 169.99], [829.16, 170.4]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[527.26, 208.38], [541.29, 208.04], [551.53, 209.22], [537.13, 209.58], [527.26, 166.1], [541.29, 166.16], [551.53, 165.94], [537.13, 165.87]]}, {\"category\": \"car\", \"corners_3d\": [[370.21, 324.72], [292.4, 363.54], [-148.03, 352.6], [11.68, 317.7], [370.21, 202.14], [292.4, 209.62], [-148.03, 207.51], [11.68, 200.78]]}, {\"category\": \"car\", \"corners_3d\": [[1934.89, 456.71], [3201.56, 737.17], [1482.91, 670.76], [1096.34, 438.86], [1934.89, 170.85], [3201.56, 168.88], [1482.91, 169.34], [1096.34, 170.98]]}, {\"category\": \"car\", \"corners_3d\": [[1321.15, 329.26], [1538.18, 376.96], [958.86, 376.96], [877.22, 329.26], [1321.15, 152.5], [1538.18, 146.29], [958.86, 146.29], [877.22, 152.5]]}]\n```" - } - ] - }, - { - "id": 2163, - "data_source": "Mono3DRefer", - "scene_name": "002093", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[346.76, 342.14], [164.37, 345.91], [310.48, 279.28], [421.62, 277.84], [346.76, 153.44], [164.37, 153.0], [310.48, 160.65], [421.62, 160.81]]}]\n```" - } - ] - }, - { - "id": 2164, - "data_source": "Mono3DRefer", - "scene_name": "002093", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-518.61, 1052.42], [-1763.54, 1207.25], [62.74, 380.7], [284.48, 373.6], [-518.61, 249.64], [-1763.54, 263.15], [62.74, 191.0], [284.48, 190.38]]}, {\"category\": \"car\", \"corners_3d\": [[446.67, 268.44], [339.39, 268.16], [415.36, 242.67], [494.04, 242.82], [446.67, 167.49], [339.39, 167.5], [415.36, 168.93], [494.04, 168.93]]}, {\"category\": \"car\", \"corners_3d\": [[663.3, 201.85], [693.55, 202.21], [683.22, 204.38], [650.9, 203.96], [663.3, 171.08], [693.55, 171.06], [683.22, 170.93], [650.9, 170.95]]}]\n```" - } - ] - }, - { - "id": 2165, - "data_source": "Mono3DRefer", - "scene_name": "002093", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-518.61, 1052.42], [-1763.54, 1207.25], [62.74, 380.7], [284.48, 373.6], [-518.61, 249.64], [-1763.54, 263.15], [62.74, 191.0], [284.48, 190.38]]}, {\"category\": \"van\", \"corners_3d\": [[346.76, 342.14], [164.37, 345.91], [310.48, 279.28], [421.62, 277.84], [346.76, 153.44], [164.37, 153.0], [310.48, 160.65], [421.62, 160.81]]}, {\"category\": \"car\", \"corners_3d\": [[446.67, 268.44], [339.39, 268.16], [415.36, 242.67], [494.04, 242.82], [446.67, 167.49], [339.39, 167.5], [415.36, 168.93], [494.04, 168.93]]}, {\"category\": \"car\", \"corners_3d\": [[663.3, 201.85], [693.55, 202.21], [683.22, 204.38], [650.9, 203.96], [663.3, 171.08], [693.55, 171.06], [683.22, 170.93], [650.9, 170.95]]}]\n```" - } - ] - }, - { - "id": 2166, - "data_source": "Mono3DRefer", - "scene_name": "002121", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[319.63, 220.38], [282.88, 220.36], [318.28, 215.31], [351.14, 215.33], [319.63, 190.24], [282.88, 190.23], [318.28, 188.38], [351.14, 188.39]]}]\n```" - } - ] - }, - { - "id": 2167, - "data_source": "Mono3DRefer", - "scene_name": "002121", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[679.72, 222.22], [756.73, 222.28], [896.22, 271.74], [742.01, 271.51], [679.72, 138.95], [756.73, 138.91], [896.22, 104.95], [742.01, 105.1]]}]\n```" - } - ] - }, - { - "id": 2168, - "data_source": "Mono3DRefer", - "scene_name": "002121", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[679.72, 222.22], [756.73, 222.28], [896.22, 271.74], [742.01, 271.51], [679.72, 138.95], [756.73, 138.91], [896.22, 104.95], [742.01, 105.1]]}, {\"category\": \"car\", \"corners_3d\": [[319.63, 220.38], [282.88, 220.36], [318.28, 215.31], [351.14, 215.33], [319.63, 190.24], [282.88, 190.23], [318.28, 188.38], [351.14, 188.39]]}]\n```" - } - ] - }, - { - "id": 2169, - "data_source": "Mono3DRefer", - "scene_name": "002122", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[331.73, 254.91], [384.35, 259.59], [208.44, 282.76], [154.54, 275.36], [331.73, 189.16], [384.35, 190.09], [208.44, 194.69], [154.54, 193.22]]}, {\"category\": \"car\", \"corners_3d\": [[249.56, 248.97], [304.86, 253.36], [88.58, 278.25], [32.58, 270.85], [249.56, 174.93], [304.86, 175.05], [88.58, 175.73], [32.58, 175.53]]}]\n```" - } - ] - }, - { - "id": 2170, - "data_source": "Mono3DRefer", - "scene_name": "002127", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[293.97, 348.44], [98.98, 350.34], [294.86, 278.08], [409.72, 277.41], [293.97, 172.81], [98.98, 172.81], [294.86, 172.83], [409.72, 172.83]]}, {\"category\": \"car\", \"corners_3d\": [[642.79, 207.82], [663.87, 207.81], [669.21, 210.88], [646.29, 210.89], [642.79, 189.0], [663.87, 189.0], [669.21, 190.41], [646.29, 190.42]]}]\n```" - } - ] - }, - { - "id": 2171, - "data_source": "Mono3DRefer", - "scene_name": "002131", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-391.48, 347.81], [-235.82, 316.46], [142.98, 305.07], [76.96, 331.19], [-391.48, 183.93], [-235.82, 181.95], [142.98, 181.22], [76.96, 182.88]]}, {\"category\": \"car\", \"corners_3d\": [[410.53, 253.15], [477.88, 252.86], [452.39, 272.58], [368.24, 273.04], [410.53, 185.36], [477.88, 185.31], [452.39, 188.39], [368.24, 188.46]]}, {\"category\": \"car\", \"corners_3d\": [[501.72, 242.7], [440.26, 242.7], [468.17, 231.15], [519.46, 231.14], [501.72, 185.64], [440.26, 185.64], [468.17, 183.53], [519.46, 183.53]]}, {\"category\": \"car\", \"corners_3d\": [[529.42, 209.45], [494.27, 209.5], [504.53, 205.46], [535.8, 205.42], [529.42, 180.85], [494.27, 180.86], [504.53, 179.98], [535.8, 179.97]]}]\n```" - } - ] - }, - { - "id": 2172, - "data_source": "Mono3DRefer", - "scene_name": "002137", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[884.43, 406.5], [1098.0, 400.1], [3470.91, 1375.26], [2587.94, 1579.2], [884.43, 171.44], [1098.0, 171.48], [3470.91, 165.57], [2587.94, 164.33]]}, {\"category\": \"car\", \"corners_3d\": [[363.18, 389.53], [167.69, 393.6], [321.77, 306.55], [439.3, 305.04], [363.18, 216.16], [167.69, 216.97], [321.77, 199.57], [439.3, 199.27]]}, {\"category\": \"car\", \"corners_3d\": [[477.37, 226.15], [519.42, 225.82], [517.4, 235.54], [467.55, 236.01], [477.37, 185.77], [519.42, 185.69], [517.4, 188.04], [467.55, 188.16]]}, {\"category\": \"car\", \"corners_3d\": [[618.7, 224.2], [662.73, 223.98], [678.95, 232.02], [628.04, 232.31], [618.7, 175.26], [662.73, 175.25], [678.95, 175.63], [628.04, 175.64]]}]\n```" - } - ] - }, - { - "id": 2173, - "data_source": "Mono3DRefer", - "scene_name": "002141", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[802.9, 220.19], [852.04, 220.91], [864.24, 228.94], [806.81, 227.98], [802.9, 175.96], [852.04, 176.01], [864.24, 176.54], [806.81, 176.47]]}, {\"category\": \"car\", \"corners_3d\": [[-280.08, 463.92], [-472.99, 441.54], [139.4, 308.39], [257.56, 313.86], [-280.08, 217.85], [-472.99, 214.39], [139.4, 193.81], [257.56, 194.65]]}]\n```" - } - ] - }, - { - "id": 2174, - "data_source": "Mono3DRefer", - "scene_name": "002148", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[498.43, 215.11], [450.83, 214.97], [473.82, 210.97], [515.79, 211.07], [498.43, 176.59], [450.83, 176.61], [473.82, 177.2], [515.79, 177.18]]}]\n```" - } - ] - }, - { - "id": 2175, - "data_source": "Mono3DRefer", - "scene_name": "002163", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-88.95, 276.49], [-169.64, 276.5], [-12.74, 255.62], [51.68, 255.61], [-88.95, 214.92], [-169.64, 214.92], [-12.74, 206.45], [51.68, 206.44]]}, {\"category\": \"car\", \"corners_3d\": [[289.54, 227.58], [249.28, 227.58], [288.62, 221.59], [324.48, 221.59], [289.54, 191.99], [249.28, 191.99], [288.62, 189.9], [324.48, 189.9]]}, {\"category\": \"car\", \"corners_3d\": [[429.31, 204.39], [403.02, 204.39], [418.23, 202.06], [442.58, 202.06], [429.31, 182.67], [403.02, 182.67], [418.23, 181.95], [442.58, 181.95]]}]\n```" - } - ] - }, - { - "id": 2176, - "data_source": "Mono3DRefer", - "scene_name": "002163", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[631.24, 197.16], [669.17, 197.21], [679.03, 205.17], [628.71, 205.08], [631.24, 156.16], [669.17, 156.13], [679.03, 150.66], [628.71, 150.72]]}]\n```" - } - ] - }, - { - "id": 2177, - "data_source": "Mono3DRefer", - "scene_name": "002163", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-88.95, 276.49], [-169.64, 276.5], [-12.74, 255.62], [51.68, 255.61], [-88.95, 214.92], [-169.64, 214.92], [-12.74, 206.45], [51.68, 206.44]]}, {\"category\": \"car\", \"corners_3d\": [[289.54, 227.58], [249.28, 227.58], [288.62, 221.59], [324.48, 221.59], [289.54, 191.99], [249.28, 191.99], [288.62, 189.9], [324.48, 189.9]]}, {\"category\": \"car\", \"corners_3d\": [[429.31, 204.39], [403.02, 204.39], [418.23, 202.06], [442.58, 202.06], [429.31, 182.67], [403.02, 182.67], [418.23, 181.95], [442.58, 181.95]]}, {\"category\": \"bus\", \"corners_3d\": [[631.24, 197.16], [669.17, 197.21], [679.03, 205.17], [628.71, 205.08], [631.24, 156.16], [669.17, 156.13], [679.03, 150.66], [628.71, 150.72]]}]\n```" - } - ] - }, - { - "id": 2178, - "data_source": "Mono3DRefer", - "scene_name": "002164", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[583.48, 211.45], [584.06, 211.05], [599.77, 211.08], [599.35, 211.49], [583.48, 180.61], [584.06, 180.53], [599.77, 180.53], [599.35, 180.61]]}]\n```" - } - ] - }, - { - "id": 2179, - "data_source": "Mono3DRefer", - "scene_name": "002164", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[717.46, 207.87], [746.54, 209.09], [693.51, 212.42], [663.8, 210.97], [717.46, 179.55], [746.54, 179.78], [693.51, 180.41], [663.8, 180.14]]}, {\"category\": \"car\", \"corners_3d\": [[684.57, 207.25], [655.28, 207.01], [662.04, 204.04], [688.8, 204.23], [684.57, 180.82], [655.28, 180.77], [662.04, 180.08], [688.8, 180.12]]}]\n```" - } - ] - }, - { - "id": 2180, - "data_source": "Mono3DRefer", - "scene_name": "002164", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[963.43, 254.39], [983.37, 258.2], [887.59, 259.4], [871.71, 255.48], [963.43, 163.95], [983.37, 163.53], [887.59, 163.4], [871.71, 163.83]]}]\n```" - } - ] - }, - { - "id": 2181, - "data_source": "Mono3DRefer", - "scene_name": "002164", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[717.46, 207.87], [746.54, 209.09], [693.51, 212.42], [663.8, 210.97], [717.46, 179.55], [746.54, 179.78], [693.51, 180.41], [663.8, 180.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[583.48, 211.45], [584.06, 211.05], [599.77, 211.08], [599.35, 211.49], [583.48, 180.61], [584.06, 180.53], [599.77, 180.53], [599.35, 180.61]]}, {\"category\": \"car\", \"corners_3d\": [[684.57, 207.25], [655.28, 207.01], [662.04, 204.04], [688.8, 204.23], [684.57, 180.82], [655.28, 180.77], [662.04, 180.08], [688.8, 180.12]]}]\n```" - } - ] - }, - { - "id": 2182, - "data_source": "Mono3DRefer", - "scene_name": "002164", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[963.43, 254.39], [983.37, 258.2], [887.59, 259.4], [871.71, 255.48], [963.43, 163.95], [983.37, 163.53], [887.59, 163.4], [871.71, 163.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[583.48, 211.45], [584.06, 211.05], [599.77, 211.08], [599.35, 211.49], [583.48, 180.61], [584.06, 180.53], [599.77, 180.53], [599.35, 180.61]]}]\n```" - } - ] - }, - { - "id": 2183, - "data_source": "Mono3DRefer", - "scene_name": "002164", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[963.43, 254.39], [983.37, 258.2], [887.59, 259.4], [871.71, 255.48], [963.43, 163.95], [983.37, 163.53], [887.59, 163.4], [871.71, 163.83]]}, {\"category\": \"car\", \"corners_3d\": [[717.46, 207.87], [746.54, 209.09], [693.51, 212.42], [663.8, 210.97], [717.46, 179.55], [746.54, 179.78], [693.51, 180.41], [663.8, 180.14]]}, {\"category\": \"car\", \"corners_3d\": [[684.57, 207.25], [655.28, 207.01], [662.04, 204.04], [688.8, 204.23], [684.57, 180.82], [655.28, 180.77], [662.04, 180.08], [688.8, 180.12]]}]\n```" - } - ] - }, - { - "id": 2184, - "data_source": "Mono3DRefer", - "scene_name": "002164", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[963.43, 254.39], [983.37, 258.2], [887.59, 259.4], [871.71, 255.48], [963.43, 163.95], [983.37, 163.53], [887.59, 163.4], [871.71, 163.83]]}, {\"category\": \"car\", \"corners_3d\": [[717.46, 207.87], [746.54, 209.09], [693.51, 212.42], [663.8, 210.97], [717.46, 179.55], [746.54, 179.78], [693.51, 180.41], [663.8, 180.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[583.48, 211.45], [584.06, 211.05], [599.77, 211.08], [599.35, 211.49], [583.48, 180.61], [584.06, 180.53], [599.77, 180.53], [599.35, 180.61]]}, {\"category\": \"car\", \"corners_3d\": [[684.57, 207.25], [655.28, 207.01], [662.04, 204.04], [688.8, 204.23], [684.57, 180.82], [655.28, 180.77], [662.04, 180.08], [688.8, 180.12]]}]\n```" - } - ] - }, - { - "id": 2185, - "data_source": "Mono3DRefer", - "scene_name": "002165", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[227.07, 223.82], [187.26, 223.82], [216.18, 220.33], [253.26, 220.32], [227.07, 190.7], [187.26, 190.7], [216.18, 189.47], [253.26, 189.47]]}]\n```" - } - ] - }, - { - "id": 2186, - "data_source": "Mono3DRefer", - "scene_name": "002165", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[591.45, 187.86], [614.52, 187.85], [617.85, 190.49], [590.7, 190.51], [591.45, 162.85], [614.52, 162.85], [617.85, 161.09], [590.7, 161.08]]}]\n```" - } - ] - }, - { - "id": 2187, - "data_source": "Mono3DRefer", - "scene_name": "002165", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[591.45, 187.86], [614.52, 187.85], [617.85, 190.49], [590.7, 190.51], [591.45, 162.85], [614.52, 162.85], [617.85, 161.09], [590.7, 161.08]]}, {\"category\": \"car\", \"corners_3d\": [[227.07, 223.82], [187.26, 223.82], [216.18, 220.33], [253.26, 220.32], [227.07, 190.7], [187.26, 190.7], [216.18, 189.47], [253.26, 189.47]]}]\n```" - } - ] - }, - { - "id": 2188, - "data_source": "Mono3DRefer", - "scene_name": "002168", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[155.7, 251.15], [193.66, 244.02], [362.59, 243.18], [341.84, 250.14], [155.7, 186.42], [193.66, 185.18], [362.59, 185.04], [341.84, 186.24]]}, {\"category\": \"car\", \"corners_3d\": [[393.99, 232.05], [375.66, 237.45], [215.31, 237.73], [247.16, 232.29], [393.99, 178.45], [375.66, 178.96], [215.31, 178.98], [247.16, 178.47]]}, {\"category\": \"car\", \"corners_3d\": [[296.38, 224.25], [310.72, 221.08], [427.67, 220.24], [421.15, 223.3], [296.38, 178.01], [310.72, 177.7], [427.67, 177.61], [421.15, 177.92]]}, {\"category\": \"car\", \"corners_3d\": [[739.0, 219.09], [753.78, 221.96], [660.68, 222.95], [651.15, 219.97], [739.0, 180.42], [753.78, 180.89], [660.68, 181.05], [651.15, 180.57]]}, {\"category\": \"car\", \"corners_3d\": [[460.05, 194.83], [478.05, 194.55], [493.41, 195.78], [474.59, 196.09], [460.05, 176.12], [478.05, 176.08], [493.41, 176.26], [474.59, 176.31]]}, {\"category\": \"car\", \"corners_3d\": [[-259.75, 338.75], [-115.86, 309.03], [168.94, 304.51], [90.5, 332.09], [-259.75, 190.29], [-115.86, 187.16], [168.94, 186.69], [90.5, 189.59]]}, {\"category\": \"car\", \"corners_3d\": [[764.85, 224.63], [779.98, 228.07], [673.78, 228.91], [665.15, 225.36], [764.85, 180.04], [779.98, 180.51], [673.78, 180.63], [665.15, 180.14]]}, {\"category\": \"car\", \"corners_3d\": [[919.98, 254.77], [972.12, 265.93], [757.48, 268.86], [730.34, 257.03], [919.98, 181.33], [972.12, 182.48], [757.48, 182.79], [730.34, 181.56]]}, {\"category\": \"car\", \"corners_3d\": [[318.74, 247.08], [295.0, 256.05], [78.02, 259.14], [126.72, 249.53], [318.74, 168.14], [295.0, 167.57], [78.02, 167.37], [126.72, 167.98]]}]\n```" - } - ] - }, - { - "id": 2189, - "data_source": "Mono3DRefer", - "scene_name": "002179", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[655.33, 228.03], [705.43, 227.88], [734.95, 240.96], [673.0, 241.19], [655.33, 183.05], [705.43, 183.02], [734.95, 185.44], [673.0, 185.48]]}]\n```" - } - ] - }, - { - "id": 2190, - "data_source": "Mono3DRefer", - "scene_name": "002192", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[510.3, 226.43], [465.17, 226.49], [481.95, 219.54], [521.22, 219.48], [510.3, 181.99], [465.17, 182.0], [481.95, 180.82], [521.22, 180.81]]}]\n```" - } - ] - }, - { - "id": 2191, - "data_source": "Mono3DRefer", - "scene_name": "002193", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[271.31, 481.17], [18.43, 480.17], [304.73, 332.98], [436.78, 333.25], [271.31, 250.7], [18.43, 250.45], [304.73, 213.28], [436.78, 213.35]]}, {\"category\": \"car\", \"corners_3d\": [[860.64, 419.39], [1098.43, 418.64], [3598.92, 1658.92], [2185.57, 1686.78], [860.64, 210.74], [1098.43, 210.62], [3598.92, 401.21], [2185.57, 405.49]]}, {\"category\": \"car\", \"corners_3d\": [[555.0, 226.28], [512.27, 226.37], [521.39, 219.87], [558.92, 219.8], [555.0, 187.87], [512.27, 187.9], [521.39, 186.07], [558.92, 186.05]]}, {\"category\": \"car\", \"corners_3d\": [[650.17, 212.51], [683.91, 212.51], [693.5, 217.66], [655.37, 217.66], [650.17, 182.05], [683.91, 182.05], [693.5, 183.24], [655.37, 183.24]]}, {\"category\": \"car\", \"corners_3d\": [[555.67, 205.19], [578.75, 205.16], [577.85, 208.12], [552.66, 208.16], [555.67, 183.03], [578.75, 183.02], [577.85, 183.95], [552.66, 183.96]]}]\n```" - } - ] - }, - { - "id": 2192, - "data_source": "Mono3DRefer", - "scene_name": "002196", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[402.01, 214.08], [368.81, 214.48], [376.11, 210.62], [406.2, 210.29], [402.01, 185.74], [368.81, 185.87], [376.11, 184.66], [406.2, 184.56]]}]\n```" - } - ] - }, - { - "id": 2193, - "data_source": "Mono3DRefer", - "scene_name": "002197", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[762.02, 209.86], [803.49, 210.24], [812.26, 215.52], [764.89, 215.03], [762.02, 171.87], [803.49, 171.86], [812.26, 171.72], [764.89, 171.73]]}]\n```" - } - ] - }, - { - "id": 2194, - "data_source": "Mono3DRefer", - "scene_name": "002200", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1040.52, 235.55], [1012.99, 229.77], [1179.78, 231.96], [1226.16, 238.22], [1040.52, 169.77], [1012.99, 170.05], [1179.78, 169.95], [1226.16, 169.64]]}, {\"category\": \"car\", \"corners_3d\": [[553.55, 192.31], [542.38, 191.94], [580.26, 191.11], [591.63, 191.45], [553.55, 171.37], [542.38, 171.4], [580.26, 171.46], [591.63, 171.44]]}, {\"category\": \"car\", \"corners_3d\": [[-516.02, 872.33], [-1134.34, 866.73], [45.04, 399.25], [248.78, 399.84], [-516.02, 322.91], [-1134.34, 321.71], [45.04, 221.42], [248.78, 221.55]]}, {\"category\": \"car\", \"corners_3d\": [[899.42, 366.99], [1092.52, 367.55], [1779.1, 648.86], [1304.15, 645.54], [899.42, 185.1], [1092.52, 185.13], [1779.1, 202.88], [1304.15, 202.67]]}, {\"category\": \"car\", \"corners_3d\": [[653.54, 213.61], [692.7, 213.18], [723.25, 218.75], [678.99, 219.31], [653.54, 173.38], [692.7, 173.37], [723.25, 173.44], [678.99, 173.45]]}, {\"category\": \"car\", \"corners_3d\": [[582.18, 198.52], [607.03, 198.17], [630.25, 200.24], [603.69, 200.64], [582.18, 174.41], [607.03, 174.39], [630.25, 174.51], [603.69, 174.54]]}, {\"category\": \"car\", \"corners_3d\": [[820.01, 234.82], [817.45, 229.96], [963.17, 232.62], [978.88, 237.96], [820.01, 172.02], [817.45, 172.08], [963.17, 172.05], [978.88, 171.98]]}]\n```" - } - ] - }, - { - "id": 2195, - "data_source": "Mono3DRefer", - "scene_name": "002200", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[488.02, 205.67], [475.25, 205.87], [470.34, 204.62], [482.66, 204.43], [488.02, 172.85], [475.25, 172.85], [470.34, 172.85], [482.66, 172.85]]}]\n```" - } - ] - }, - { - "id": 2196, - "data_source": "Mono3DRefer", - "scene_name": "002200", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1040.52, 235.55], [1012.99, 229.77], [1179.78, 231.96], [1226.16, 238.22], [1040.52, 169.77], [1012.99, 170.05], [1179.78, 169.95], [1226.16, 169.64]]}, {\"category\": \"car\", \"corners_3d\": [[553.55, 192.31], [542.38, 191.94], [580.26, 191.11], [591.63, 191.45], [553.55, 171.37], [542.38, 171.4], [580.26, 171.46], [591.63, 171.44]]}, {\"category\": \"cyclist\", \"corners_3d\": [[488.02, 205.67], [475.25, 205.87], [470.34, 204.62], [482.66, 204.43], [488.02, 172.85], [475.25, 172.85], [470.34, 172.85], [482.66, 172.85]]}, {\"category\": \"car\", \"corners_3d\": [[-516.02, 872.33], [-1134.34, 866.73], [45.04, 399.25], [248.78, 399.84], [-516.02, 322.91], [-1134.34, 321.71], [45.04, 221.42], [248.78, 221.55]]}, {\"category\": \"car\", \"corners_3d\": [[899.42, 366.99], [1092.52, 367.55], [1779.1, 648.86], [1304.15, 645.54], [899.42, 185.1], [1092.52, 185.13], [1779.1, 202.88], [1304.15, 202.67]]}, {\"category\": \"car\", \"corners_3d\": [[653.54, 213.61], [692.7, 213.18], [723.25, 218.75], [678.99, 219.31], [653.54, 173.38], [692.7, 173.37], [723.25, 173.44], [678.99, 173.45]]}, {\"category\": \"car\", \"corners_3d\": [[582.18, 198.52], [607.03, 198.17], [630.25, 200.24], [603.69, 200.64], [582.18, 174.41], [607.03, 174.39], [630.25, 174.51], [603.69, 174.54]]}, {\"category\": \"car\", \"corners_3d\": [[820.01, 234.82], [817.45, 229.96], [963.17, 232.62], [978.88, 237.96], [820.01, 172.02], [817.45, 172.08], [963.17, 172.05], [978.88, 171.98]]}]\n```" - } - ] - }, - { - "id": 2197, - "data_source": "Mono3DRefer", - "scene_name": "002204", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[554.73, 191.31], [540.2, 191.33], [541.08, 190.44], [554.91, 190.41], [554.73, 176.0], [540.2, 176.01], [541.08, 175.86], [554.91, 175.85]]}]\n```" - } - ] - }, - { - "id": 2198, - "data_source": "Mono3DRefer", - "scene_name": "002207", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1237.63, 252.5], [1351.53, 261.62], [1253.02, 274.51], [1135.93, 262.72], [1237.63, 170.35], [1351.53, 170.06], [1253.02, 169.65], [1135.93, 170.02]]}, {\"category\": \"car\", \"corners_3d\": [[1030.19, 226.56], [983.65, 223.08], [1070.86, 219.89], [1119.89, 222.93], [1030.19, 171.1], [983.65, 171.21], [1070.86, 171.32], [1119.89, 171.22]]}]\n```" - } - ] - }, - { - "id": 2199, - "data_source": "Mono3DRefer", - "scene_name": "002207", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[595.32, 245.03], [570.71, 238.57], [728.56, 232.93], [764.91, 238.27], [595.32, 152.61], [570.71, 154.42], [728.56, 156.0], [764.91, 154.5]]}]\n```" - } - ] - }, - { - "id": 2200, - "data_source": "Mono3DRefer", - "scene_name": "002207", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1237.63, 252.5], [1351.53, 261.62], [1253.02, 274.51], [1135.93, 262.72], [1237.63, 170.35], [1351.53, 170.06], [1253.02, 169.65], [1135.93, 170.02]]}, {\"category\": \"van\", \"corners_3d\": [[595.32, 245.03], [570.71, 238.57], [728.56, 232.93], [764.91, 238.27], [595.32, 152.61], [570.71, 154.42], [728.56, 156.0], [764.91, 154.5]]}, {\"category\": \"car\", \"corners_3d\": [[1030.19, 226.56], [983.65, 223.08], [1070.86, 219.89], [1119.89, 222.93], [1030.19, 171.1], [983.65, 171.21], [1070.86, 171.32], [1119.89, 171.22]]}]\n```" - } - ] - }, - { - "id": 2201, - "data_source": "Mono3DRefer", - "scene_name": "002208", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[415.66, 207.57], [423.54, 206.09], [492.77, 206.06], [487.97, 207.53], [415.66, 178.5], [423.54, 178.26], [492.77, 178.26], [487.97, 178.5]]}]\n```" - } - ] - }, - { - "id": 2202, - "data_source": "Mono3DRefer", - "scene_name": "002215", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1205.17, 250.46], [1309.7, 257.98], [1247.52, 272.0], [1133.84, 261.95], [1205.17, 174.67], [1309.7, 174.85], [1247.52, 175.18], [1133.84, 174.94]]}, {\"category\": \"car\", \"corners_3d\": [[861.33, 239.39], [814.69, 234.99], [904.5, 229.07], [952.11, 232.64], [861.33, 177.51], [814.69, 177.2], [904.5, 176.79], [952.11, 177.04]]}, {\"category\": \"car\", \"corners_3d\": [[1074.52, 225.47], [1027.3, 222.71], [1088.76, 218.79], [1135.2, 221.12], [1074.52, 181.76], [1027.3, 181.3], [1088.76, 180.63], [1135.2, 181.03]]}]\n```" - } - ] - }, - { - "id": 2203, - "data_source": "Mono3DRefer", - "scene_name": "002215", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[313.01, 275.48], [300.25, 264.39], [519.27, 251.58], [552.5, 259.65], [313.01, 153.4], [300.25, 155.5], [519.27, 157.93], [552.5, 156.4]]}]\n```" - } - ] - }, - { - "id": 2204, - "data_source": "Mono3DRefer", - "scene_name": "002215", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1205.17, 250.46], [1309.7, 257.98], [1247.52, 272.0], [1133.84, 261.95], [1205.17, 174.67], [1309.7, 174.85], [1247.52, 175.18], [1133.84, 174.94]]}, {\"category\": \"van\", \"corners_3d\": [[313.01, 275.48], [300.25, 264.39], [519.27, 251.58], [552.5, 259.65], [313.01, 153.4], [300.25, 155.5], [519.27, 157.93], [552.5, 156.4]]}, {\"category\": \"car\", \"corners_3d\": [[861.33, 239.39], [814.69, 234.99], [904.5, 229.07], [952.11, 232.64], [861.33, 177.51], [814.69, 177.2], [904.5, 176.79], [952.11, 177.04]]}, {\"category\": \"car\", \"corners_3d\": [[1074.52, 225.47], [1027.3, 222.71], [1088.76, 218.79], [1135.2, 221.12], [1074.52, 181.76], [1027.3, 181.3], [1088.76, 180.63], [1135.2, 181.03]]}]\n```" - } - ] - }, - { - "id": 2205, - "data_source": "Mono3DRefer", - "scene_name": "002222", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[554.2, 190.9], [538.59, 190.92], [540.38, 189.98], [555.18, 189.96], [554.2, 174.45], [538.59, 174.45], [540.38, 174.36], [555.18, 174.36]]}]\n```" - } - ] - }, - { - "id": 2206, - "data_source": "Mono3DRefer", - "scene_name": "002223", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[661.75, 226.61], [718.86, 227.08], [725.04, 238.63], [655.83, 237.94], [661.75, 173.18], [718.86, 173.19], [725.04, 173.26], [655.83, 173.25]]}, {\"category\": \"car\", \"corners_3d\": [[1192.22, 318.98], [1450.29, 350.0], [1463.49, 479.42], [1063.99, 397.05], [1192.22, 173.86], [1450.29, 174.07], [1463.49, 174.97], [1063.99, 174.4]]}, {\"category\": \"car\", \"corners_3d\": [[516.13, 227.6], [467.83, 226.92], [505.34, 219.48], [547.32, 219.98], [516.13, 170.91], [467.83, 170.93], [505.34, 171.2], [547.32, 171.18]]}, {\"category\": \"car\", \"corners_3d\": [[157.99, 273.13], [82.63, 271.13], [233.63, 248.3], [293.57, 249.48], [157.99, 189.01], [82.63, 188.69], [233.63, 185.01], [293.57, 185.2]]}, {\"category\": \"car\", \"corners_3d\": [[440.77, 224.22], [399.94, 223.29], [460.02, 216.14], [495.93, 216.83], [440.77, 184.57], [399.94, 184.36], [460.02, 182.73], [495.93, 182.88]]}, {\"category\": \"car\", \"corners_3d\": [[784.81, 192.92], [805.63, 193.53], [762.42, 194.6], [741.9, 193.92], [784.81, 168.19], [805.63, 168.05], [762.42, 167.8], [741.9, 167.96]]}, {\"category\": \"car\", \"corners_3d\": [[-123.62, 283.37], [-194.01, 281.3], [45.38, 253.07], [100.55, 254.19], [-123.62, 199.46], [-194.01, 198.96], [45.38, 192.16], [100.55, 192.43]]}]\n```" - } - ] - }, - { - "id": 2207, - "data_source": "Mono3DRefer", - "scene_name": "002227", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[90.14, 267.27], [143.01, 257.5], [312.22, 257.28], [278.98, 266.99], [90.14, 191.62], [143.01, 189.68], [312.22, 189.63], [278.98, 191.56]]}, {\"category\": \"car\", \"corners_3d\": [[176.11, 248.77], [211.79, 242.4], [369.48, 242.22], [348.3, 248.56], [176.11, 183.93], [211.79, 183.0], [369.48, 182.97], [348.3, 183.9]]}, {\"category\": \"car\", \"corners_3d\": [[348.09, 221.98], [364.38, 219.09], [465.23, 219.23], [455.2, 222.14], [348.09, 173.86], [364.38, 173.8], [465.23, 173.8], [455.2, 173.86]]}, {\"category\": \"car\", \"corners_3d\": [[522.59, 200.81], [519.82, 201.78], [463.06, 201.81], [467.75, 200.83], [522.59, 178.32], [519.82, 178.51], [463.06, 178.52], [467.75, 178.33]]}]\n```" - } - ] - }, - { - "id": 2208, - "data_source": "Mono3DRefer", - "scene_name": "002227", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[788.55, 204.8], [797.77, 204.8], [807.76, 206.5], [798.04, 206.5], [788.55, 170.06], [797.77, 170.06], [807.76, 169.91], [798.04, 169.91]]}, {\"category\": \"cyclist\", \"corners_3d\": [[770.03, 202.68], [778.17, 202.67], [786.61, 204.11], [778.08, 204.11], [770.03, 172.66], [778.17, 172.66], [786.61, 172.65], [778.08, 172.65]]}]\n```" - } - ] - }, - { - "id": 2209, - "data_source": "Mono3DRefer", - "scene_name": "002227", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[90.14, 267.27], [143.01, 257.5], [312.22, 257.28], [278.98, 266.99], [90.14, 191.62], [143.01, 189.68], [312.22, 189.63], [278.98, 191.56]]}, {\"category\": \"car\", \"corners_3d\": [[176.11, 248.77], [211.79, 242.4], [369.48, 242.22], [348.3, 248.56], [176.11, 183.93], [211.79, 183.0], [369.48, 182.97], [348.3, 183.9]]}, {\"category\": \"car\", \"corners_3d\": [[348.09, 221.98], [364.38, 219.09], [465.23, 219.23], [455.2, 222.14], [348.09, 173.86], [364.38, 173.8], [465.23, 173.8], [455.2, 173.86]]}, {\"category\": \"car\", \"corners_3d\": [[522.59, 200.81], [519.82, 201.78], [463.06, 201.81], [467.75, 200.83], [522.59, 178.32], [519.82, 178.51], [463.06, 178.52], [467.75, 178.33]]}, {\"category\": \"cyclist\", \"corners_3d\": [[788.55, 204.8], [797.77, 204.8], [807.76, 206.5], [798.04, 206.5], [788.55, 170.06], [797.77, 170.06], [807.76, 169.91], [798.04, 169.91]]}, {\"category\": \"cyclist\", \"corners_3d\": [[770.03, 202.68], [778.17, 202.67], [786.61, 204.11], [778.08, 204.11], [770.03, 172.66], [778.17, 172.66], [786.61, 172.65], [778.08, 172.65]]}]\n```" - } - ] - }, - { - "id": 2210, - "data_source": "Mono3DRefer", - "scene_name": "002234", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1139.24, 236.66], [1216.04, 245.43], [956.27, 246.51], [910.28, 237.49], [1139.24, 125.08], [1216.04, 118.51], [956.27, 117.7], [910.28, 124.45]]}]\n```" - } - ] - }, - { - "id": 2211, - "data_source": "Mono3DRefer", - "scene_name": "002234", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[274.84, 389.49], [80.94, 385.94], [303.26, 303.03], [423.2, 304.34], [274.84, 211.19], [80.94, 210.56], [303.26, 195.89], [423.2, 196.12]]}, {\"category\": \"car\", \"corners_3d\": [[440.29, 304.14], [319.4, 302.81], [418.74, 264.04], [504.0, 264.69], [440.29, 175.73], [319.4, 175.7], [418.74, 174.85], [504.0, 174.87]]}, {\"category\": \"car\", \"corners_3d\": [[715.08, 234.24], [769.37, 234.2], [804.23, 247.05], [738.59, 247.11], [715.08, 188.11], [769.37, 188.1], [804.23, 191.3], [738.59, 191.31]]}, {\"category\": \"car\", \"corners_3d\": [[537.93, 247.7], [476.95, 247.32], [506.67, 234.7], [557.4, 234.96], [537.93, 182.95], [476.95, 182.9], [506.67, 181.19], [557.4, 181.23]]}, {\"category\": \"car\", \"corners_3d\": [[692.48, 221.98], [739.54, 221.95], [763.84, 230.68], [708.42, 230.72], [692.48, 177.58], [739.54, 177.58], [763.84, 178.42], [708.42, 178.43]]}, {\"category\": \"car\", \"corners_3d\": [[559.07, 232.66], [514.76, 232.55], [528.46, 225.5], [567.54, 225.58], [559.07, 183.16], [514.76, 183.14], [528.46, 181.92], [567.54, 181.94]]}]\n```" - } - ] - }, - { - "id": 2212, - "data_source": "Mono3DRefer", - "scene_name": "002234", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[274.84, 389.49], [80.94, 385.94], [303.26, 303.03], [423.2, 304.34], [274.84, 211.19], [80.94, 210.56], [303.26, 195.89], [423.2, 196.12]]}, {\"category\": \"car\", \"corners_3d\": [[440.29, 304.14], [319.4, 302.81], [418.74, 264.04], [504.0, 264.69], [440.29, 175.73], [319.4, 175.7], [418.74, 174.85], [504.0, 174.87]]}, {\"category\": \"van\", \"corners_3d\": [[1139.24, 236.66], [1216.04, 245.43], [956.27, 246.51], [910.28, 237.49], [1139.24, 125.08], [1216.04, 118.51], [956.27, 117.7], [910.28, 124.45]]}, {\"category\": \"car\", \"corners_3d\": [[715.08, 234.24], [769.37, 234.2], [804.23, 247.05], [738.59, 247.11], [715.08, 188.11], [769.37, 188.1], [804.23, 191.3], [738.59, 191.31]]}, {\"category\": \"car\", \"corners_3d\": [[537.93, 247.7], [476.95, 247.32], [506.67, 234.7], [557.4, 234.96], [537.93, 182.95], [476.95, 182.9], [506.67, 181.19], [557.4, 181.23]]}, {\"category\": \"car\", \"corners_3d\": [[692.48, 221.98], [739.54, 221.95], [763.84, 230.68], [708.42, 230.72], [692.48, 177.58], [739.54, 177.58], [763.84, 178.42], [708.42, 178.43]]}, {\"category\": \"car\", \"corners_3d\": [[559.07, 232.66], [514.76, 232.55], [528.46, 225.5], [567.54, 225.58], [559.07, 183.16], [514.76, 183.14], [528.46, 181.92], [567.54, 181.94]]}]\n```" - } - ] - }, - { - "id": 2213, - "data_source": "Mono3DRefer", - "scene_name": "002237", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[250.73, 494.54], [-60.82, 500.1], [270.74, 330.81], [419.72, 329.5], [250.73, 169.24], [-60.82, 169.17], [270.74, 171.08], [419.72, 171.09]]}, {\"category\": \"car\", \"corners_3d\": [[556.23, 245.73], [499.03, 244.61], [540.27, 232.71], [588.4, 233.49], [556.23, 190.14], [499.03, 189.88], [540.27, 187.05], [588.4, 187.24]]}, {\"category\": \"car\", \"corners_3d\": [[599.24, 230.27], [561.99, 229.32], [593.9, 222.84], [627.29, 223.58], [599.24, 186.71], [561.99, 186.49], [593.9, 184.92], [627.29, 185.1]]}, {\"category\": \"car\", \"corners_3d\": [[651.64, 220.0], [618.79, 219.28], [641.04, 214.93], [671.08, 215.52], [651.64, 186.35], [618.79, 186.14], [641.04, 184.9], [671.08, 185.07]]}, {\"category\": \"car\", \"corners_3d\": [[817.06, 189.32], [838.0, 189.59], [820.92, 190.75], [798.83, 190.45], [817.06, 168.35], [838.0, 168.27], [820.92, 167.96], [798.83, 168.04]]}]\n```" - } - ] - }, - { - "id": 2214, - "data_source": "Mono3DRefer", - "scene_name": "002238", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[587.8, 192.99], [571.14, 192.99], [572.75, 191.93], [588.53, 191.93], [587.8, 179.05], [571.14, 179.05], [572.75, 178.72], [588.53, 178.72]]}]\n```" - } - ] - }, - { - "id": 2215, - "data_source": "Mono3DRefer", - "scene_name": "002266", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.35, 201.59], [615.21, 201.57], [616.79, 203.91], [589.91, 203.93], [590.35, 177.02], [615.21, 177.02], [616.79, 177.35], [589.91, 177.36]]}, {\"category\": \"car\", \"corners_3d\": [[435.66, 258.7], [511.04, 258.71], [466.23, 297.53], [356.78, 297.51], [435.66, 178.33], [511.04, 178.33], [466.23, 180.8], [356.78, 180.8]]}, {\"category\": \"car\", \"corners_3d\": [[139.21, 264.43], [67.61, 264.7], [156.54, 248.99], [215.7, 248.8], [139.21, 198.89], [67.61, 198.97], [156.54, 194.5], [215.7, 194.45]]}]\n```" - } - ] - }, - { - "id": 2216, - "data_source": "Mono3DRefer", - "scene_name": "002266", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[682.2, 230.47], [693.8, 229.45], [747.07, 233.24], [735.7, 234.4], [682.2, 163.04], [693.8, 163.21], [747.07, 162.57], [735.7, 162.37]]}]\n```" - } - ] - }, - { - "id": 2217, - "data_source": "Mono3DRefer", - "scene_name": "002266", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[692.53, 232.15], [689.87, 230.98], [704.32, 230.9], [707.26, 232.06], [692.53, 172.47], [689.87, 172.48], [704.32, 172.48], [707.26, 172.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[675.21, 224.65], [666.76, 223.84], [676.04, 223.39], [684.56, 224.18], [675.21, 173.18], [666.76, 173.18], [676.04, 173.18], [684.56, 173.18]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[401.76, 235.64], [398.0, 236.91], [386.39, 236.94], [390.38, 235.67], [401.76, 175.47], [398.0, 175.53], [386.39, 175.53], [390.38, 175.47]]}]\n```" - } - ] - }, - { - "id": 2218, - "data_source": "Mono3DRefer", - "scene_name": "002266", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.35, 201.59], [615.21, 201.57], [616.79, 203.91], [589.91, 203.93], [590.35, 177.02], [615.21, 177.02], [616.79, 177.35], [589.91, 177.36]]}, {\"category\": \"car\", \"corners_3d\": [[435.66, 258.7], [511.04, 258.71], [466.23, 297.53], [356.78, 297.51], [435.66, 178.33], [511.04, 178.33], [466.23, 180.8], [356.78, 180.8]]}, {\"category\": \"car\", \"corners_3d\": [[139.21, 264.43], [67.61, 264.7], [156.54, 248.99], [215.7, 248.8], [139.21, 198.89], [67.61, 198.97], [156.54, 194.5], [215.7, 194.45]]}, {\"category\": \"cyclist\", \"corners_3d\": [[682.2, 230.47], [693.8, 229.45], [747.07, 233.24], [735.7, 234.4], [682.2, 163.04], [693.8, 163.21], [747.07, 162.57], [735.7, 162.37]]}]\n```" - } - ] - }, - { - "id": 2219, - "data_source": "Mono3DRefer", - "scene_name": "002266", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.35, 201.59], [615.21, 201.57], [616.79, 203.91], [589.91, 203.93], [590.35, 177.02], [615.21, 177.02], [616.79, 177.35], [589.91, 177.36]]}, {\"category\": \"car\", \"corners_3d\": [[435.66, 258.7], [511.04, 258.71], [466.23, 297.53], [356.78, 297.51], [435.66, 178.33], [511.04, 178.33], [466.23, 180.8], [356.78, 180.8]]}, {\"category\": \"car\", \"corners_3d\": [[139.21, 264.43], [67.61, 264.7], [156.54, 248.99], [215.7, 248.8], [139.21, 198.89], [67.61, 198.97], [156.54, 194.5], [215.7, 194.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[692.53, 232.15], [689.87, 230.98], [704.32, 230.9], [707.26, 232.06], [692.53, 172.47], [689.87, 172.48], [704.32, 172.48], [707.26, 172.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[675.21, 224.65], [666.76, 223.84], [676.04, 223.39], [684.56, 224.18], [675.21, 173.18], [666.76, 173.18], [676.04, 173.18], [684.56, 173.18]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[401.76, 235.64], [398.0, 236.91], [386.39, 236.94], [390.38, 235.67], [401.76, 175.47], [398.0, 175.53], [386.39, 175.53], [390.38, 175.47]]}]\n```" - } - ] - }, - { - "id": 2220, - "data_source": "Mono3DRefer", - "scene_name": "002266", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[682.2, 230.47], [693.8, 229.45], [747.07, 233.24], [735.7, 234.4], [682.2, 163.04], [693.8, 163.21], [747.07, 162.57], [735.7, 162.37]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[692.53, 232.15], [689.87, 230.98], [704.32, 230.9], [707.26, 232.06], [692.53, 172.47], [689.87, 172.48], [704.32, 172.48], [707.26, 172.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[675.21, 224.65], [666.76, 223.84], [676.04, 223.39], [684.56, 224.18], [675.21, 173.18], [666.76, 173.18], [676.04, 173.18], [684.56, 173.18]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[401.76, 235.64], [398.0, 236.91], [386.39, 236.94], [390.38, 235.67], [401.76, 175.47], [398.0, 175.53], [386.39, 175.53], [390.38, 175.47]]}]\n```" - } - ] - }, - { - "id": 2221, - "data_source": "Mono3DRefer", - "scene_name": "002266", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.35, 201.59], [615.21, 201.57], [616.79, 203.91], [589.91, 203.93], [590.35, 177.02], [615.21, 177.02], [616.79, 177.35], [589.91, 177.36]]}, {\"category\": \"car\", \"corners_3d\": [[435.66, 258.7], [511.04, 258.71], [466.23, 297.53], [356.78, 297.51], [435.66, 178.33], [511.04, 178.33], [466.23, 180.8], [356.78, 180.8]]}, {\"category\": \"car\", \"corners_3d\": [[139.21, 264.43], [67.61, 264.7], [156.54, 248.99], [215.7, 248.8], [139.21, 198.89], [67.61, 198.97], [156.54, 194.5], [215.7, 194.45]]}, {\"category\": \"cyclist\", \"corners_3d\": [[682.2, 230.47], [693.8, 229.45], [747.07, 233.24], [735.7, 234.4], [682.2, 163.04], [693.8, 163.21], [747.07, 162.57], [735.7, 162.37]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[692.53, 232.15], [689.87, 230.98], [704.32, 230.9], [707.26, 232.06], [692.53, 172.47], [689.87, 172.48], [704.32, 172.48], [707.26, 172.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[675.21, 224.65], [666.76, 223.84], [676.04, 223.39], [684.56, 224.18], [675.21, 173.18], [666.76, 173.18], [676.04, 173.18], [684.56, 173.18]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[401.76, 235.64], [398.0, 236.91], [386.39, 236.94], [390.38, 235.67], [401.76, 175.47], [398.0, 175.53], [386.39, 175.53], [390.38, 175.47]]}]\n```" - } - ] - }, - { - "id": 2222, - "data_source": "Mono3DRefer", - "scene_name": "002269", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[543.06, 217.67], [507.14, 217.67], [518.82, 212.54], [550.61, 212.54], [543.06, 183.64], [507.14, 183.64], [518.82, 182.41], [550.61, 182.41]]}, {\"category\": \"car\", \"corners_3d\": [[334.94, 344.07], [189.82, 343.41], [337.57, 285.29], [433.5, 285.58], [334.94, 188.82], [189.82, 188.75], [337.57, 183.34], [433.5, 183.36]]}]\n```" - } - ] - }, - { - "id": 2223, - "data_source": "Mono3DRefer", - "scene_name": "002269", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[387.32, 219.59], [375.85, 219.52], [388.8, 217.49], [399.78, 217.55], [387.32, 184.48], [375.85, 184.46], [388.8, 183.96], [399.78, 183.97]]}]\n```" - } - ] - }, - { - "id": 2224, - "data_source": "Mono3DRefer", - "scene_name": "002269", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[543.06, 217.67], [507.14, 217.67], [518.82, 212.54], [550.61, 212.54], [543.06, 183.64], [507.14, 183.64], [518.82, 182.41], [550.61, 182.41]]}, {\"category\": \"car\", \"corners_3d\": [[334.94, 344.07], [189.82, 343.41], [337.57, 285.29], [433.5, 285.58], [334.94, 188.82], [189.82, 188.75], [337.57, 183.34], [433.5, 183.36]]}, {\"category\": \"cyclist\", \"corners_3d\": [[387.32, 219.59], [375.85, 219.52], [388.8, 217.49], [399.78, 217.55], [387.32, 184.48], [375.85, 184.46], [388.8, 183.96], [399.78, 183.97]]}]\n```" - } - ] - }, - { - "id": 2225, - "data_source": "Mono3DRefer", - "scene_name": "002273", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[458.46, 217.8], [446.46, 220.47], [334.83, 219.95], [352.9, 217.33], [458.46, 184.93], [446.46, 185.65], [334.83, 185.51], [352.9, 184.8]]}]\n```" - } - ] - }, - { - "id": 2226, - "data_source": "Mono3DRefer", - "scene_name": "002276", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[134.23, 253.9], [165.57, 246.49], [349.7, 243.39], [337.48, 250.16], [134.23, 178.38], [165.57, 177.88], [349.7, 177.67], [337.48, 178.13]]}, {\"category\": \"car\", \"corners_3d\": [[461.72, 227.85], [505.07, 227.65], [496.08, 235.64], [446.35, 235.89], [461.72, 184.23], [505.07, 184.19], [496.08, 185.85], [446.35, 185.9]]}, {\"category\": \"car\", \"corners_3d\": [[513.91, 222.29], [472.08, 222.33], [487.48, 216.45], [524.34, 216.43], [513.91, 183.53], [472.08, 183.54], [487.48, 182.27], [524.34, 182.26]]}]\n```" - } - ] - }, - { - "id": 2227, - "data_source": "Mono3DRefer", - "scene_name": "002277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[478.02, 273.6], [376.61, 274.17], [430.14, 247.52], [504.77, 247.21], [478.02, 164.58], [376.61, 164.54], [430.14, 166.72], [504.77, 166.75]]}]\n```" - } - ] - }, - { - "id": 2228, - "data_source": "Mono3DRefer", - "scene_name": "002277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[381.9, 367.52], [184.21, 370.71], [346.07, 286.75], [459.14, 285.69], [381.9, 196.27], [184.21, 196.65], [346.07, 186.55], [459.14, 186.43]]}, {\"category\": \"car\", \"corners_3d\": [[518.07, 240.19], [443.35, 239.84], [484.32, 226.24], [543.97, 226.46], [518.07, 169.51], [443.35, 169.53], [484.32, 170.21], [543.97, 170.19]]}, {\"category\": \"car\", \"corners_3d\": [[663.44, 198.45], [691.4, 198.88], [675.2, 200.63], [645.68, 200.14], [663.44, 169.11], [691.4, 169.05], [675.2, 168.79], [645.68, 168.86]]}]\n```" - } - ] - }, - { - "id": 2229, - "data_source": "Mono3DRefer", - "scene_name": "002277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[381.9, 367.52], [184.21, 370.71], [346.07, 286.75], [459.14, 285.69], [381.9, 196.27], [184.21, 196.65], [346.07, 186.55], [459.14, 186.43]]}, {\"category\": \"van\", \"corners_3d\": [[478.02, 273.6], [376.61, 274.17], [430.14, 247.52], [504.77, 247.21], [478.02, 164.58], [376.61, 164.54], [430.14, 166.72], [504.77, 166.75]]}, {\"category\": \"car\", \"corners_3d\": [[518.07, 240.19], [443.35, 239.84], [484.32, 226.24], [543.97, 226.46], [518.07, 169.51], [443.35, 169.53], [484.32, 170.21], [543.97, 170.19]]}, {\"category\": \"car\", \"corners_3d\": [[663.44, 198.45], [691.4, 198.88], [675.2, 200.63], [645.68, 200.14], [663.44, 169.11], [691.4, 169.05], [675.2, 168.79], [645.68, 168.86]]}]\n```" - } - ] - }, - { - "id": 2230, - "data_source": "Mono3DRefer", - "scene_name": "002278", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[636.56, 196.26], [651.53, 196.17], [661.38, 197.5], [645.59, 197.6], [636.56, 179.52], [651.53, 179.5], [661.38, 179.87], [645.59, 179.9]]}]\n```" - } - ] - }, - { - "id": 2231, - "data_source": "Mono3DRefer", - "scene_name": "002278", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[434.64, 206.74], [402.78, 206.78], [421.83, 203.28], [450.38, 203.24], [434.64, 170.84], [402.78, 170.84], [421.83, 171.04], [450.38, 171.05]]}]\n```" - } - ] - }, - { - "id": 2232, - "data_source": "Mono3DRefer", - "scene_name": "002278", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[434.64, 206.74], [402.78, 206.78], [421.83, 203.28], [450.38, 203.24], [434.64, 170.84], [402.78, 170.84], [421.83, 171.04], [450.38, 171.05]]}, {\"category\": \"car\", \"corners_3d\": [[636.56, 196.26], [651.53, 196.17], [661.38, 197.5], [645.59, 197.6], [636.56, 179.52], [651.53, 179.5], [661.38, 179.87], [645.59, 179.9]]}]\n```" - } - ] - }, - { - "id": 2233, - "data_source": "Mono3DRefer", - "scene_name": "002279", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[698.63, 209.59], [717.36, 209.64], [722.28, 212.18], [702.25, 212.13], [698.63, 163.66], [717.36, 163.65], [722.28, 163.01], [702.25, 163.02]]}, {\"category\": \"cyclist\", \"corners_3d\": [[680.18, 199.12], [688.61, 199.15], [689.44, 200.21], [680.67, 200.19], [680.18, 170.11], [688.61, 170.11], [689.44, 169.99], [680.67, 170.0]]}]\n```" - } - ] - }, - { - "id": 2234, - "data_source": "Mono3DRefer", - "scene_name": "002279", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[865.13, 240.74], [900.69, 240.77], [917.22, 244.74], [879.58, 244.7], [865.13, 157.8], [900.69, 157.79], [917.22, 156.91], [879.58, 156.92]]}]\n```" - } - ] - }, - { - "id": 2235, - "data_source": "Mono3DRefer", - "scene_name": "002279", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[865.13, 240.74], [900.69, 240.77], [917.22, 244.74], [879.58, 244.7], [865.13, 157.8], [900.69, 157.79], [917.22, 156.91], [879.58, 156.92]]}, {\"category\": \"cyclist\", \"corners_3d\": [[698.63, 209.59], [717.36, 209.64], [722.28, 212.18], [702.25, 212.13], [698.63, 163.66], [717.36, 163.65], [722.28, 163.01], [702.25, 163.02]]}, {\"category\": \"cyclist\", \"corners_3d\": [[680.18, 199.12], [688.61, 199.15], [689.44, 200.21], [680.67, 200.19], [680.18, 170.11], [688.61, 170.11], [689.44, 169.99], [680.67, 170.0]]}]\n```" - } - ] - }, - { - "id": 2236, - "data_source": "Mono3DRefer", - "scene_name": "002281", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[238.9, 218.2], [196.42, 218.33], [248.82, 212.0], [285.27, 211.91], [238.9, 173.65], [196.42, 173.65], [248.82, 173.54], [285.27, 173.54]]}, {\"category\": \"car\", \"corners_3d\": [[586.1, 218.57], [634.3, 218.45], [642.46, 225.47], [586.85, 225.63], [586.1, 179.7], [634.3, 179.68], [642.46, 180.74], [586.85, 180.76]]}, {\"category\": \"car\", \"corners_3d\": [[-194.19, 366.28], [57.5, 335.46], [132.83, 438.35], [-307.64, 531.73], [-194.19, 142.2], [57.5, 147.08], [132.83, 130.77], [-307.64, 115.98]]}, {\"category\": \"car\", \"corners_3d\": [[225.22, 266.37], [307.8, 258.21], [421.17, 273.7], [334.92, 285.29], [225.22, 183.35], [307.8, 182.43], [421.17, 184.17], [334.92, 185.47]]}, {\"category\": \"car\", \"corners_3d\": [[316.91, 247.48], [381.9, 242.18], [483.05, 253.69], [415.44, 260.99], [316.91, 167.03], [381.9, 167.44], [483.05, 166.54], [415.44, 165.97]]}, {\"category\": \"car\", \"corners_3d\": [[398.88, 225.51], [441.13, 222.92], [523.09, 229.38], [479.89, 232.7], [398.88, 181.83], [441.13, 181.39], [523.09, 182.49], [479.89, 183.05]]}, {\"category\": \"car\", \"corners_3d\": [[417.32, 219.33], [453.26, 217.15], [526.87, 221.49], [491.22, 224.13], [417.32, 178.59], [453.26, 178.32], [526.87, 178.85], [491.22, 179.18]]}, {\"category\": \"car\", \"corners_3d\": [[460.62, 208.44], [488.11, 207.25], [538.33, 209.9], [510.51, 211.29], [460.62, 179.16], [488.11, 178.95], [538.33, 179.42], [510.51, 179.67]]}, {\"category\": \"car\", \"corners_3d\": [[485.52, 202.68], [507.32, 201.73], [553.38, 203.34], [531.92, 204.4], [485.52, 175.74], [507.32, 175.65], [553.38, 175.8], [531.92, 175.91]]}, {\"category\": \"car\", \"corners_3d\": [[656.2, 200.39], [682.85, 200.42], [687.93, 203.16], [658.63, 203.13], [656.2, 177.97], [682.85, 177.98], [687.93, 178.49], [658.63, 178.48]]}, {\"category\": \"car\", \"corners_3d\": [[588.44, 203.88], [619.79, 203.81], [624.52, 207.02], [589.94, 207.1], [588.44, 175.29], [619.79, 175.29], [624.52, 175.54], [589.94, 175.55]]}, {\"category\": \"car\", \"corners_3d\": [[75.04, 288.27], [-46.56, 289.4], [111.85, 259.24], [201.07, 258.62], [75.04, 190.43], [-46.56, 190.6], [111.85, 186.01], [201.07, 185.91]]}, {\"category\": \"car\", \"corners_3d\": [[117.76, 233.9], [55.32, 233.85], [161.27, 222.33], [211.97, 222.36], [117.76, 180.52], [55.32, 180.51], [161.27, 179.07], [211.97, 179.07]]}]\n```" - } - ] - }, - { - "id": 2237, - "data_source": "Mono3DRefer", - "scene_name": "002281", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[331.21, 227.97], [271.36, 228.06], [329.1, 218.2], [378.19, 218.14], [331.21, 153.77], [271.36, 153.74], [329.1, 157.15], [378.19, 157.18]]}, {\"category\": \"van\", \"corners_3d\": [[691.36, 230.46], [760.18, 230.29], [821.94, 250.95], [728.51, 251.25], [691.36, 151.2], [760.18, 151.26], [821.94, 143.49], [728.51, 143.38]]}]\n```" - } - ] - }, - { - "id": 2238, - "data_source": "Mono3DRefer", - "scene_name": "002281", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[331.21, 227.97], [271.36, 228.06], [329.1, 218.2], [378.19, 218.14], [331.21, 153.77], [271.36, 153.74], [329.1, 157.15], [378.19, 157.18]]}, {\"category\": \"car\", \"corners_3d\": [[238.9, 218.2], [196.42, 218.33], [248.82, 212.0], [285.27, 211.91], [238.9, 173.65], [196.42, 173.65], [248.82, 173.54], [285.27, 173.54]]}, {\"category\": \"car\", \"corners_3d\": [[586.1, 218.57], [634.3, 218.45], [642.46, 225.47], [586.85, 225.63], [586.1, 179.7], [634.3, 179.68], [642.46, 180.74], [586.85, 180.76]]}, {\"category\": \"car\", \"corners_3d\": [[-194.19, 366.28], [57.5, 335.46], [132.83, 438.35], [-307.64, 531.73], [-194.19, 142.2], [57.5, 147.08], [132.83, 130.77], [-307.64, 115.98]]}, {\"category\": \"car\", \"corners_3d\": [[225.22, 266.37], [307.8, 258.21], [421.17, 273.7], [334.92, 285.29], [225.22, 183.35], [307.8, 182.43], [421.17, 184.17], [334.92, 185.47]]}, {\"category\": \"car\", \"corners_3d\": [[316.91, 247.48], [381.9, 242.18], [483.05, 253.69], [415.44, 260.99], [316.91, 167.03], [381.9, 167.44], [483.05, 166.54], [415.44, 165.97]]}, {\"category\": \"car\", \"corners_3d\": [[398.88, 225.51], [441.13, 222.92], [523.09, 229.38], [479.89, 232.7], [398.88, 181.83], [441.13, 181.39], [523.09, 182.49], [479.89, 183.05]]}, {\"category\": \"car\", \"corners_3d\": [[417.32, 219.33], [453.26, 217.15], [526.87, 221.49], [491.22, 224.13], [417.32, 178.59], [453.26, 178.32], [526.87, 178.85], [491.22, 179.18]]}, {\"category\": \"car\", \"corners_3d\": [[460.62, 208.44], [488.11, 207.25], [538.33, 209.9], [510.51, 211.29], [460.62, 179.16], [488.11, 178.95], [538.33, 179.42], [510.51, 179.67]]}, {\"category\": \"car\", \"corners_3d\": [[485.52, 202.68], [507.32, 201.73], [553.38, 203.34], [531.92, 204.4], [485.52, 175.74], [507.32, 175.65], [553.38, 175.8], [531.92, 175.91]]}, {\"category\": \"van\", \"corners_3d\": [[691.36, 230.46], [760.18, 230.29], [821.94, 250.95], [728.51, 251.25], [691.36, 151.2], [760.18, 151.26], [821.94, 143.49], [728.51, 143.38]]}, {\"category\": \"car\", \"corners_3d\": [[656.2, 200.39], [682.85, 200.42], [687.93, 203.16], [658.63, 203.13], [656.2, 177.97], [682.85, 177.98], [687.93, 178.49], [658.63, 178.48]]}, {\"category\": \"car\", \"corners_3d\": [[588.44, 203.88], [619.79, 203.81], [624.52, 207.02], [589.94, 207.1], [588.44, 175.29], [619.79, 175.29], [624.52, 175.54], [589.94, 175.55]]}, {\"category\": \"car\", \"corners_3d\": [[75.04, 288.27], [-46.56, 289.4], [111.85, 259.24], [201.07, 258.62], [75.04, 190.43], [-46.56, 190.6], [111.85, 186.01], [201.07, 185.91]]}, {\"category\": \"car\", \"corners_3d\": [[117.76, 233.9], [55.32, 233.85], [161.27, 222.33], [211.97, 222.36], [117.76, 180.52], [55.32, 180.51], [161.27, 179.07], [211.97, 179.07]]}]\n```" - } - ] - }, - { - "id": 2239, - "data_source": "Mono3DRefer", - "scene_name": "002283", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[81.65, 252.95], [75.96, 254.68], [57.88, 255.27], [63.99, 253.52], [81.65, 194.4], [75.96, 194.86], [57.88, 195.02], [63.99, 194.55]]}]\n```" - } - ] - }, - { - "id": 2240, - "data_source": "Mono3DRefer", - "scene_name": "002283", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[188.29, 244.43], [187.57, 245.38], [123.38, 247.9], [124.99, 246.89], [188.29, 188.48], [187.57, 188.69], [123.38, 189.24], [124.99, 189.02]]}]\n```" - } - ] - }, - { - "id": 2241, - "data_source": "Mono3DRefer", - "scene_name": "002283", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[597.5, 203.28], [625.59, 203.31], [625.68, 206.13], [594.99, 206.1], [597.5, 175.54], [625.59, 175.54], [625.68, 175.79], [594.99, 175.79]]}]\n```" - } - ] - }, - { - "id": 2242, - "data_source": "Mono3DRefer", - "scene_name": "002283", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[81.65, 252.95], [75.96, 254.68], [57.88, 255.27], [63.99, 253.52], [81.65, 194.4], [75.96, 194.86], [57.88, 195.02], [63.99, 194.55]]}, {\"category\": \"cyclist\", \"corners_3d\": [[188.29, 244.43], [187.57, 245.38], [123.38, 247.9], [124.99, 246.89], [188.29, 188.48], [187.57, 188.69], [123.38, 189.24], [124.99, 189.02]]}]\n```" - } - ] - }, - { - "id": 2243, - "data_source": "Mono3DRefer", - "scene_name": "002283", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[597.5, 203.28], [625.59, 203.31], [625.68, 206.13], [594.99, 206.1], [597.5, 175.54], [625.59, 175.54], [625.68, 175.79], [594.99, 175.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[81.65, 252.95], [75.96, 254.68], [57.88, 255.27], [63.99, 253.52], [81.65, 194.4], [75.96, 194.86], [57.88, 195.02], [63.99, 194.55]]}]\n```" - } - ] - }, - { - "id": 2244, - "data_source": "Mono3DRefer", - "scene_name": "002283", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[597.5, 203.28], [625.59, 203.31], [625.68, 206.13], [594.99, 206.1], [597.5, 175.54], [625.59, 175.54], [625.68, 175.79], [594.99, 175.79]]}, {\"category\": \"cyclist\", \"corners_3d\": [[188.29, 244.43], [187.57, 245.38], [123.38, 247.9], [124.99, 246.89], [188.29, 188.48], [187.57, 188.69], [123.38, 189.24], [124.99, 189.02]]}]\n```" - } - ] - }, - { - "id": 2245, - "data_source": "Mono3DRefer", - "scene_name": "002283", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[597.5, 203.28], [625.59, 203.31], [625.68, 206.13], [594.99, 206.1], [597.5, 175.54], [625.59, 175.54], [625.68, 175.79], [594.99, 175.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[81.65, 252.95], [75.96, 254.68], [57.88, 255.27], [63.99, 253.52], [81.65, 194.4], [75.96, 194.86], [57.88, 195.02], [63.99, 194.55]]}, {\"category\": \"cyclist\", \"corners_3d\": [[188.29, 244.43], [187.57, 245.38], [123.38, 247.9], [124.99, 246.89], [188.29, 188.48], [187.57, 188.69], [123.38, 189.24], [124.99, 189.02]]}]\n```" - } - ] - }, - { - "id": 2246, - "data_source": "Mono3DRefer", - "scene_name": "002284", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[421.52, 207.09], [395.96, 207.05], [412.11, 204.7], [435.92, 204.73], [421.52, 184.76], [395.96, 184.75], [412.11, 183.93], [435.92, 183.94]]}]\n```" - } - ] - }, - { - "id": 2247, - "data_source": "Mono3DRefer", - "scene_name": "002284", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[653.69, 207.31], [708.96, 207.31], [764.42, 226.65], [678.12, 226.64], [653.69, 147.42], [708.96, 147.42], [764.42, 133.14], [678.12, 133.15]]}]\n```" - } - ] - }, - { - "id": 2248, - "data_source": "Mono3DRefer", - "scene_name": "002284", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[653.69, 207.31], [708.96, 207.31], [764.42, 226.65], [678.12, 226.64], [653.69, 147.42], [708.96, 147.42], [764.42, 133.14], [678.12, 133.15]]}, {\"category\": \"car\", \"corners_3d\": [[421.52, 207.09], [395.96, 207.05], [412.11, 204.7], [435.92, 204.73], [421.52, 184.76], [395.96, 184.75], [412.11, 183.93], [435.92, 183.94]]}]\n```" - } - ] - }, - { - "id": 2249, - "data_source": "Mono3DRefer", - "scene_name": "002285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[738.21, 201.77], [736.35, 201.13], [774.2, 201.22], [776.93, 201.87], [738.21, 169.8], [736.35, 169.87], [774.2, 169.86], [776.93, 169.79]]}]\n```" - } - ] - }, - { - "id": 2250, - "data_source": "Mono3DRefer", - "scene_name": "002285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[861.03, 217.1], [844.97, 217.12], [836.5, 215.62], [852.02, 215.6], [861.03, 166.89], [844.97, 166.88], [836.5, 167.09], [852.02, 167.09]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[552.1, 205.09], [552.7, 205.61], [540.75, 205.68], [540.35, 205.16], [552.1, 171.41], [552.7, 171.38], [540.75, 171.38], [540.35, 171.4]]}]\n```" - } - ] - }, - { - "id": 2251, - "data_source": "Mono3DRefer", - "scene_name": "002285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.3, 221.06], [655.21, 221.06], [662.78, 229.17], [602.15, 229.16], [603.3, 179.29], [655.21, 179.29], [662.78, 180.38], [602.15, 180.37]]}, {\"category\": \"car\", \"corners_3d\": [[964.89, 207.95], [946.28, 206.25], [1034.2, 206.08], [1057.14, 207.76], [964.89, 176.14], [946.28, 175.98], [1034.2, 175.96], [1057.14, 176.12]]}, {\"category\": \"car\", \"corners_3d\": [[436.14, 206.07], [408.73, 206.06], [425.06, 203.45], [450.31, 203.46], [436.14, 180.67], [408.73, 180.67], [425.06, 180.06], [450.31, 180.06]]}, {\"category\": \"car\", \"corners_3d\": [[585.99, 209.18], [622.09, 209.15], [625.29, 213.54], [584.81, 213.59], [585.99, 176.31], [622.09, 176.3], [625.29, 176.72], [584.81, 176.73]]}]\n```" - } - ] - }, - { - "id": 2252, - "data_source": "Mono3DRefer", - "scene_name": "002285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[643.42, 199.13], [671.11, 199.17], [674.32, 201.8], [643.86, 201.75], [643.42, 167.65], [671.11, 167.64], [674.32, 167.12], [643.86, 167.13]]}]\n```" - } - ] - }, - { - "id": 2253, - "data_source": "Mono3DRefer", - "scene_name": "002285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[861.03, 217.1], [844.97, 217.12], [836.5, 215.62], [852.02, 215.6], [861.03, 166.89], [844.97, 166.88], [836.5, 167.09], [852.02, 167.09]]}, {\"category\": \"cyclist\", \"corners_3d\": [[738.21, 201.77], [736.35, 201.13], [774.2, 201.22], [776.93, 201.87], [738.21, 169.8], [736.35, 169.87], [774.2, 169.86], [776.93, 169.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[552.1, 205.09], [552.7, 205.61], [540.75, 205.68], [540.35, 205.16], [552.1, 171.41], [552.7, 171.38], [540.75, 171.38], [540.35, 171.4]]}]\n```" - } - ] - }, - { - "id": 2254, - "data_source": "Mono3DRefer", - "scene_name": "002285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.3, 221.06], [655.21, 221.06], [662.78, 229.17], [602.15, 229.16], [603.3, 179.29], [655.21, 179.29], [662.78, 180.38], [602.15, 180.37]]}, {\"category\": \"car\", \"corners_3d\": [[964.89, 207.95], [946.28, 206.25], [1034.2, 206.08], [1057.14, 207.76], [964.89, 176.14], [946.28, 175.98], [1034.2, 175.96], [1057.14, 176.12]]}, {\"category\": \"car\", \"corners_3d\": [[436.14, 206.07], [408.73, 206.06], [425.06, 203.45], [450.31, 203.46], [436.14, 180.67], [408.73, 180.67], [425.06, 180.06], [450.31, 180.06]]}, {\"category\": \"cyclist\", \"corners_3d\": [[738.21, 201.77], [736.35, 201.13], [774.2, 201.22], [776.93, 201.87], [738.21, 169.8], [736.35, 169.87], [774.2, 169.86], [776.93, 169.79]]}, {\"category\": \"car\", \"corners_3d\": [[585.99, 209.18], [622.09, 209.15], [625.29, 213.54], [584.81, 213.59], [585.99, 176.31], [622.09, 176.3], [625.29, 176.72], [584.81, 176.73]]}]\n```" - } - ] - }, - { - "id": 2255, - "data_source": "Mono3DRefer", - "scene_name": "002285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[738.21, 201.77], [736.35, 201.13], [774.2, 201.22], [776.93, 201.87], [738.21, 169.8], [736.35, 169.87], [774.2, 169.86], [776.93, 169.79]]}, {\"category\": \"van\", \"corners_3d\": [[643.42, 199.13], [671.11, 199.17], [674.32, 201.8], [643.86, 201.75], [643.42, 167.65], [671.11, 167.64], [674.32, 167.12], [643.86, 167.13]]}]\n```" - } - ] - }, - { - "id": 2256, - "data_source": "Mono3DRefer", - "scene_name": "002285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.3, 221.06], [655.21, 221.06], [662.78, 229.17], [602.15, 229.16], [603.3, 179.29], [655.21, 179.29], [662.78, 180.38], [602.15, 180.37]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[861.03, 217.1], [844.97, 217.12], [836.5, 215.62], [852.02, 215.6], [861.03, 166.89], [844.97, 166.88], [836.5, 167.09], [852.02, 167.09]]}, {\"category\": \"car\", \"corners_3d\": [[964.89, 207.95], [946.28, 206.25], [1034.2, 206.08], [1057.14, 207.76], [964.89, 176.14], [946.28, 175.98], [1034.2, 175.96], [1057.14, 176.12]]}, {\"category\": \"car\", \"corners_3d\": [[436.14, 206.07], [408.73, 206.06], [425.06, 203.45], [450.31, 203.46], [436.14, 180.67], [408.73, 180.67], [425.06, 180.06], [450.31, 180.06]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[552.1, 205.09], [552.7, 205.61], [540.75, 205.68], [540.35, 205.16], [552.1, 171.41], [552.7, 171.38], [540.75, 171.38], [540.35, 171.4]]}, {\"category\": \"car\", \"corners_3d\": [[585.99, 209.18], [622.09, 209.15], [625.29, 213.54], [584.81, 213.59], [585.99, 176.31], [622.09, 176.3], [625.29, 176.72], [584.81, 176.73]]}]\n```" - } - ] - }, - { - "id": 2257, - "data_source": "Mono3DRefer", - "scene_name": "002285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[861.03, 217.1], [844.97, 217.12], [836.5, 215.62], [852.02, 215.6], [861.03, 166.89], [844.97, 166.88], [836.5, 167.09], [852.02, 167.09]]}, {\"category\": \"van\", \"corners_3d\": [[643.42, 199.13], [671.11, 199.17], [674.32, 201.8], [643.86, 201.75], [643.42, 167.65], [671.11, 167.64], [674.32, 167.12], [643.86, 167.13]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[552.1, 205.09], [552.7, 205.61], [540.75, 205.68], [540.35, 205.16], [552.1, 171.41], [552.7, 171.38], [540.75, 171.38], [540.35, 171.4]]}]\n```" - } - ] - }, - { - "id": 2258, - "data_source": "Mono3DRefer", - "scene_name": "002285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.3, 221.06], [655.21, 221.06], [662.78, 229.17], [602.15, 229.16], [603.3, 179.29], [655.21, 179.29], [662.78, 180.38], [602.15, 180.37]]}, {\"category\": \"car\", \"corners_3d\": [[964.89, 207.95], [946.28, 206.25], [1034.2, 206.08], [1057.14, 207.76], [964.89, 176.14], [946.28, 175.98], [1034.2, 175.96], [1057.14, 176.12]]}, {\"category\": \"car\", \"corners_3d\": [[436.14, 206.07], [408.73, 206.06], [425.06, 203.45], [450.31, 203.46], [436.14, 180.67], [408.73, 180.67], [425.06, 180.06], [450.31, 180.06]]}, {\"category\": \"van\", \"corners_3d\": [[643.42, 199.13], [671.11, 199.17], [674.32, 201.8], [643.86, 201.75], [643.42, 167.65], [671.11, 167.64], [674.32, 167.12], [643.86, 167.13]]}, {\"category\": \"car\", \"corners_3d\": [[585.99, 209.18], [622.09, 209.15], [625.29, 213.54], [584.81, 213.59], [585.99, 176.31], [622.09, 176.3], [625.29, 176.72], [584.81, 176.73]]}]\n```" - } - ] - }, - { - "id": 2259, - "data_source": "Mono3DRefer", - "scene_name": "002285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.3, 221.06], [655.21, 221.06], [662.78, 229.17], [602.15, 229.16], [603.3, 179.29], [655.21, 179.29], [662.78, 180.38], [602.15, 180.37]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[861.03, 217.1], [844.97, 217.12], [836.5, 215.62], [852.02, 215.6], [861.03, 166.89], [844.97, 166.88], [836.5, 167.09], [852.02, 167.09]]}, {\"category\": \"car\", \"corners_3d\": [[964.89, 207.95], [946.28, 206.25], [1034.2, 206.08], [1057.14, 207.76], [964.89, 176.14], [946.28, 175.98], [1034.2, 175.96], [1057.14, 176.12]]}, {\"category\": \"car\", \"corners_3d\": [[436.14, 206.07], [408.73, 206.06], [425.06, 203.45], [450.31, 203.46], [436.14, 180.67], [408.73, 180.67], [425.06, 180.06], [450.31, 180.06]]}, {\"category\": \"cyclist\", \"corners_3d\": [[738.21, 201.77], [736.35, 201.13], [774.2, 201.22], [776.93, 201.87], [738.21, 169.8], [736.35, 169.87], [774.2, 169.86], [776.93, 169.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[552.1, 205.09], [552.7, 205.61], [540.75, 205.68], [540.35, 205.16], [552.1, 171.41], [552.7, 171.38], [540.75, 171.38], [540.35, 171.4]]}, {\"category\": \"car\", \"corners_3d\": [[585.99, 209.18], [622.09, 209.15], [625.29, 213.54], [584.81, 213.59], [585.99, 176.31], [622.09, 176.3], [625.29, 176.72], [584.81, 176.73]]}]\n```" - } - ] - }, - { - "id": 2260, - "data_source": "Mono3DRefer", - "scene_name": "002285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[861.03, 217.1], [844.97, 217.12], [836.5, 215.62], [852.02, 215.6], [861.03, 166.89], [844.97, 166.88], [836.5, 167.09], [852.02, 167.09]]}, {\"category\": \"cyclist\", \"corners_3d\": [[738.21, 201.77], [736.35, 201.13], [774.2, 201.22], [776.93, 201.87], [738.21, 169.8], [736.35, 169.87], [774.2, 169.86], [776.93, 169.79]]}, {\"category\": \"van\", \"corners_3d\": [[643.42, 199.13], [671.11, 199.17], [674.32, 201.8], [643.86, 201.75], [643.42, 167.65], [671.11, 167.64], [674.32, 167.12], [643.86, 167.13]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[552.1, 205.09], [552.7, 205.61], [540.75, 205.68], [540.35, 205.16], [552.1, 171.41], [552.7, 171.38], [540.75, 171.38], [540.35, 171.4]]}]\n```" - } - ] - }, - { - "id": 2261, - "data_source": "Mono3DRefer", - "scene_name": "002285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.3, 221.06], [655.21, 221.06], [662.78, 229.17], [602.15, 229.16], [603.3, 179.29], [655.21, 179.29], [662.78, 180.38], [602.15, 180.37]]}, {\"category\": \"car\", \"corners_3d\": [[964.89, 207.95], [946.28, 206.25], [1034.2, 206.08], [1057.14, 207.76], [964.89, 176.14], [946.28, 175.98], [1034.2, 175.96], [1057.14, 176.12]]}, {\"category\": \"car\", \"corners_3d\": [[436.14, 206.07], [408.73, 206.06], [425.06, 203.45], [450.31, 203.46], [436.14, 180.67], [408.73, 180.67], [425.06, 180.06], [450.31, 180.06]]}, {\"category\": \"cyclist\", \"corners_3d\": [[738.21, 201.77], [736.35, 201.13], [774.2, 201.22], [776.93, 201.87], [738.21, 169.8], [736.35, 169.87], [774.2, 169.86], [776.93, 169.79]]}, {\"category\": \"van\", \"corners_3d\": [[643.42, 199.13], [671.11, 199.17], [674.32, 201.8], [643.86, 201.75], [643.42, 167.65], [671.11, 167.64], [674.32, 167.12], [643.86, 167.13]]}, {\"category\": \"car\", \"corners_3d\": [[585.99, 209.18], [622.09, 209.15], [625.29, 213.54], [584.81, 213.59], [585.99, 176.31], [622.09, 176.3], [625.29, 176.72], [584.81, 176.73]]}]\n```" - } - ] - }, - { - "id": 2262, - "data_source": "Mono3DRefer", - "scene_name": "002285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.3, 221.06], [655.21, 221.06], [662.78, 229.17], [602.15, 229.16], [603.3, 179.29], [655.21, 179.29], [662.78, 180.38], [602.15, 180.37]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[861.03, 217.1], [844.97, 217.12], [836.5, 215.62], [852.02, 215.6], [861.03, 166.89], [844.97, 166.88], [836.5, 167.09], [852.02, 167.09]]}, {\"category\": \"car\", \"corners_3d\": [[964.89, 207.95], [946.28, 206.25], [1034.2, 206.08], [1057.14, 207.76], [964.89, 176.14], [946.28, 175.98], [1034.2, 175.96], [1057.14, 176.12]]}, {\"category\": \"car\", \"corners_3d\": [[436.14, 206.07], [408.73, 206.06], [425.06, 203.45], [450.31, 203.46], [436.14, 180.67], [408.73, 180.67], [425.06, 180.06], [450.31, 180.06]]}, {\"category\": \"van\", \"corners_3d\": [[643.42, 199.13], [671.11, 199.17], [674.32, 201.8], [643.86, 201.75], [643.42, 167.65], [671.11, 167.64], [674.32, 167.12], [643.86, 167.13]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[552.1, 205.09], [552.7, 205.61], [540.75, 205.68], [540.35, 205.16], [552.1, 171.41], [552.7, 171.38], [540.75, 171.38], [540.35, 171.4]]}, {\"category\": \"car\", \"corners_3d\": [[585.99, 209.18], [622.09, 209.15], [625.29, 213.54], [584.81, 213.59], [585.99, 176.31], [622.09, 176.3], [625.29, 176.72], [584.81, 176.73]]}]\n```" - } - ] - }, - { - "id": 2263, - "data_source": "Mono3DRefer", - "scene_name": "002285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.3, 221.06], [655.21, 221.06], [662.78, 229.17], [602.15, 229.16], [603.3, 179.29], [655.21, 179.29], [662.78, 180.38], [602.15, 180.37]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[861.03, 217.1], [844.97, 217.12], [836.5, 215.62], [852.02, 215.6], [861.03, 166.89], [844.97, 166.88], [836.5, 167.09], [852.02, 167.09]]}, {\"category\": \"car\", \"corners_3d\": [[964.89, 207.95], [946.28, 206.25], [1034.2, 206.08], [1057.14, 207.76], [964.89, 176.14], [946.28, 175.98], [1034.2, 175.96], [1057.14, 176.12]]}, {\"category\": \"car\", \"corners_3d\": [[436.14, 206.07], [408.73, 206.06], [425.06, 203.45], [450.31, 203.46], [436.14, 180.67], [408.73, 180.67], [425.06, 180.06], [450.31, 180.06]]}, {\"category\": \"cyclist\", \"corners_3d\": [[738.21, 201.77], [736.35, 201.13], [774.2, 201.22], [776.93, 201.87], [738.21, 169.8], [736.35, 169.87], [774.2, 169.86], [776.93, 169.79]]}, {\"category\": \"van\", \"corners_3d\": [[643.42, 199.13], [671.11, 199.17], [674.32, 201.8], [643.86, 201.75], [643.42, 167.65], [671.11, 167.64], [674.32, 167.12], [643.86, 167.13]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[552.1, 205.09], [552.7, 205.61], [540.75, 205.68], [540.35, 205.16], [552.1, 171.41], [552.7, 171.38], [540.75, 171.38], [540.35, 171.4]]}, {\"category\": \"car\", \"corners_3d\": [[585.99, 209.18], [622.09, 209.15], [625.29, 213.54], [584.81, 213.59], [585.99, 176.31], [622.09, 176.3], [625.29, 176.72], [584.81, 176.73]]}]\n```" - } - ] - }, - { - "id": 2264, - "data_source": "Mono3DRefer", - "scene_name": "002288", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1160.81, 248.11], [1193.7, 248.37], [1290.63, 262.3], [1251.31, 261.93], [1160.81, 137.62], [1193.7, 137.5], [1290.63, 130.98], [1251.31, 131.15]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1006.57, 228.97], [1027.36, 229.07], [1073.66, 236.03], [1050.22, 235.91], [1006.57, 153.19], [1027.36, 153.16], [1073.66, 150.72], [1050.22, 150.76]]}, {\"category\": \"cyclist\", \"corners_3d\": [[968.23, 220.04], [985.67, 220.57], [980.99, 224.48], [962.18, 223.87], [968.23, 162.98], [985.67, 162.87], [980.99, 162.05], [962.18, 162.18]]}]\n```" - } - ] - }, - { - "id": 2265, - "data_source": "Mono3DRefer", - "scene_name": "002288", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[401.93, 231.11], [387.34, 235.68], [257.78, 236.01], [281.91, 231.4], [401.93, 182.45], [387.34, 183.2], [257.78, 183.25], [281.91, 182.49]]}, {\"category\": \"car\", \"corners_3d\": [[321.84, 225.12], [342.46, 222.08], [436.07, 222.69], [421.04, 225.81], [321.84, 183.63], [342.46, 183.0], [436.07, 183.13], [421.04, 183.77]]}, {\"category\": \"car\", \"corners_3d\": [[910.89, 185.63], [884.61, 185.52], [877.99, 184.52], [902.13, 184.62], [910.89, 163.7], [884.61, 163.78], [877.99, 164.49], [902.13, 164.43]]}]\n```" - } - ] - }, - { - "id": 2266, - "data_source": "Mono3DRefer", - "scene_name": "002288", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[401.93, 231.11], [387.34, 235.68], [257.78, 236.01], [281.91, 231.4], [401.93, 182.45], [387.34, 183.2], [257.78, 183.25], [281.91, 182.49]]}, {\"category\": \"car\", \"corners_3d\": [[321.84, 225.12], [342.46, 222.08], [436.07, 222.69], [421.04, 225.81], [321.84, 183.63], [342.46, 183.0], [436.07, 183.13], [421.04, 183.77]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1160.81, 248.11], [1193.7, 248.37], [1290.63, 262.3], [1251.31, 261.93], [1160.81, 137.62], [1193.7, 137.5], [1290.63, 130.98], [1251.31, 131.15]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1006.57, 228.97], [1027.36, 229.07], [1073.66, 236.03], [1050.22, 235.91], [1006.57, 153.19], [1027.36, 153.16], [1073.66, 150.72], [1050.22, 150.76]]}, {\"category\": \"cyclist\", \"corners_3d\": [[968.23, 220.04], [985.67, 220.57], [980.99, 224.48], [962.18, 223.87], [968.23, 162.98], [985.67, 162.87], [980.99, 162.05], [962.18, 162.18]]}, {\"category\": \"car\", \"corners_3d\": [[910.89, 185.63], [884.61, 185.52], [877.99, 184.52], [902.13, 184.62], [910.89, 163.7], [884.61, 163.78], [877.99, 164.49], [902.13, 164.43]]}]\n```" - } - ] - }, - { - "id": 2267, - "data_source": "Mono3DRefer", - "scene_name": "002294", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[680.04, 203.89], [708.91, 203.99], [713.32, 207.35], [681.34, 207.23], [680.04, 170.77], [708.91, 170.76], [713.32, 170.54], [681.34, 170.54]]}]\n```" - } - ] - }, - { - "id": 2268, - "data_source": "Mono3DRefer", - "scene_name": "002294", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[708.63, 293.61], [824.83, 294.97], [905.81, 355.24], [731.88, 352.21], [708.63, 184.76], [824.83, 184.89], [905.81, 190.83], [731.88, 190.53]]}, {\"category\": \"car\", \"corners_3d\": [[272.13, 259.02], [308.13, 252.21], [427.01, 254.49], [400.42, 261.72], [272.13, 190.16], [308.13, 188.79], [427.01, 189.25], [400.42, 190.7]]}]\n```" - } - ] - }, - { - "id": 2269, - "data_source": "Mono3DRefer", - "scene_name": "002294", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[684.56, 317.62], [716.82, 312.75], [742.58, 318.66], [709.85, 323.95], [684.56, 170.19], [716.82, 170.28], [742.58, 170.17], [709.85, 170.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[357.87, 265.3], [380.22, 263.35], [390.86, 265.67], [368.16, 267.73], [357.87, 189.89], [380.22, 189.54], [390.86, 189.96], [368.16, 190.34]]}]\n```" - } - ] - }, - { - "id": 2270, - "data_source": "Mono3DRefer", - "scene_name": "002294", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[708.63, 293.61], [824.83, 294.97], [905.81, 355.24], [731.88, 352.21], [708.63, 184.76], [824.83, 184.89], [905.81, 190.83], [731.88, 190.53]]}, {\"category\": \"car\", \"corners_3d\": [[272.13, 259.02], [308.13, 252.21], [427.01, 254.49], [400.42, 261.72], [272.13, 190.16], [308.13, 188.79], [427.01, 189.25], [400.42, 190.7]]}, {\"category\": \"van\", \"corners_3d\": [[680.04, 203.89], [708.91, 203.99], [713.32, 207.35], [681.34, 207.23], [680.04, 170.77], [708.91, 170.76], [713.32, 170.54], [681.34, 170.54]]}]\n```" - } - ] - }, - { - "id": 2271, - "data_source": "Mono3DRefer", - "scene_name": "002294", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[684.56, 317.62], [716.82, 312.75], [742.58, 318.66], [709.85, 323.95], [684.56, 170.19], [716.82, 170.28], [742.58, 170.17], [709.85, 170.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[357.87, 265.3], [380.22, 263.35], [390.86, 265.67], [368.16, 267.73], [357.87, 189.89], [380.22, 189.54], [390.86, 189.96], [368.16, 190.34]]}, {\"category\": \"van\", \"corners_3d\": [[680.04, 203.89], [708.91, 203.99], [713.32, 207.35], [681.34, 207.23], [680.04, 170.77], [708.91, 170.76], [713.32, 170.54], [681.34, 170.54]]}]\n```" - } - ] - }, - { - "id": 2272, - "data_source": "Mono3DRefer", - "scene_name": "002294", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[708.63, 293.61], [824.83, 294.97], [905.81, 355.24], [731.88, 352.21], [708.63, 184.76], [824.83, 184.89], [905.81, 190.83], [731.88, 190.53]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[684.56, 317.62], [716.82, 312.75], [742.58, 318.66], [709.85, 323.95], [684.56, 170.19], [716.82, 170.28], [742.58, 170.17], [709.85, 170.07]]}, {\"category\": \"car\", \"corners_3d\": [[272.13, 259.02], [308.13, 252.21], [427.01, 254.49], [400.42, 261.72], [272.13, 190.16], [308.13, 188.79], [427.01, 189.25], [400.42, 190.7]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[357.87, 265.3], [380.22, 263.35], [390.86, 265.67], [368.16, 267.73], [357.87, 189.89], [380.22, 189.54], [390.86, 189.96], [368.16, 190.34]]}]\n```" - } - ] - }, - { - "id": 2273, - "data_source": "Mono3DRefer", - "scene_name": "002294", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[708.63, 293.61], [824.83, 294.97], [905.81, 355.24], [731.88, 352.21], [708.63, 184.76], [824.83, 184.89], [905.81, 190.83], [731.88, 190.53]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[684.56, 317.62], [716.82, 312.75], [742.58, 318.66], [709.85, 323.95], [684.56, 170.19], [716.82, 170.28], [742.58, 170.17], [709.85, 170.07]]}, {\"category\": \"car\", \"corners_3d\": [[272.13, 259.02], [308.13, 252.21], [427.01, 254.49], [400.42, 261.72], [272.13, 190.16], [308.13, 188.79], [427.01, 189.25], [400.42, 190.7]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[357.87, 265.3], [380.22, 263.35], [390.86, 265.67], [368.16, 267.73], [357.87, 189.89], [380.22, 189.54], [390.86, 189.96], [368.16, 190.34]]}, {\"category\": \"van\", \"corners_3d\": [[680.04, 203.89], [708.91, 203.99], [713.32, 207.35], [681.34, 207.23], [680.04, 170.77], [708.91, 170.76], [713.32, 170.54], [681.34, 170.54]]}]\n```" - } - ] - }, - { - "id": 2274, - "data_source": "Mono3DRefer", - "scene_name": "002304", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[107.56, 212.97], [102.82, 214.89], [-169.22, 221.26], [-149.53, 218.72], [107.56, 166.92], [102.82, 166.63], [-169.22, 165.69], [-149.53, 166.06]]}]\n```" - } - ] - }, - { - "id": 2275, - "data_source": "Mono3DRefer", - "scene_name": "002304", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[592.33, 197.99], [620.9, 198.01], [620.56, 200.39], [589.29, 200.37], [592.33, 169.77], [620.9, 169.77], [620.56, 169.47], [589.29, 169.48]]}]\n```" - } - ] - }, - { - "id": 2276, - "data_source": "Mono3DRefer", - "scene_name": "002304", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[592.33, 197.99], [620.9, 198.01], [620.56, 200.39], [589.29, 200.37], [592.33, 169.77], [620.9, 169.77], [620.56, 169.47], [589.29, 169.48]]}, {\"category\": \"truck\", \"corners_3d\": [[107.56, 212.97], [102.82, 214.89], [-169.22, 221.26], [-149.53, 218.72], [107.56, 166.92], [102.82, 166.63], [-169.22, 165.69], [-149.53, 166.06]]}]\n```" - } - ] - }, - { - "id": 2277, - "data_source": "Mono3DRefer", - "scene_name": "002321", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[354.85, 246.4], [325.38, 253.67], [142.12, 252.41], [187.43, 245.35], [354.85, 193.75], [325.38, 195.81], [142.12, 195.45], [187.43, 193.45]]}]\n```" - } - ] - }, - { - "id": 2278, - "data_source": "Mono3DRefer", - "scene_name": "002337", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[81.99, 258.52], [13.21, 258.36], [118.46, 243.61], [175.51, 243.72], [81.99, 203.66], [13.21, 203.6], [118.46, 198.3], [175.51, 198.34]]}]\n```" - } - ] - }, - { - "id": 2279, - "data_source": "Mono3DRefer", - "scene_name": "002337", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[599.39, 190.35], [623.48, 190.35], [625.96, 193.6], [597.39, 193.6], [599.39, 164.24], [623.48, 164.24], [625.96, 162.64], [597.39, 162.64]]}]\n```" - } - ] - }, - { - "id": 2280, - "data_source": "Mono3DRefer", - "scene_name": "002337", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[599.39, 190.35], [623.48, 190.35], [625.96, 193.6], [597.39, 193.6], [599.39, 164.24], [623.48, 164.24], [625.96, 162.64], [597.39, 162.64]]}, {\"category\": \"car\", \"corners_3d\": [[81.99, 258.52], [13.21, 258.36], [118.46, 243.61], [175.51, 243.72], [81.99, 203.66], [13.21, 203.6], [118.46, 198.3], [175.51, 198.34]]}]\n```" - } - ] - }, - { - "id": 2281, - "data_source": "Mono3DRefer", - "scene_name": "002339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[510.66, 209.84], [479.79, 209.84], [492.43, 206.22], [520.27, 206.22], [510.66, 178.13], [479.79, 178.13], [492.43, 177.62], [520.27, 177.62]]}, {\"category\": \"car\", \"corners_3d\": [[413.4, 253.81], [345.86, 253.74], [395.82, 238.79], [450.91, 238.84], [413.4, 186.04], [345.86, 186.03], [395.82, 183.59], [450.91, 183.6]]}, {\"category\": \"car\", \"corners_3d\": [[688.88, 239.64], [750.1, 239.47], [785.09, 253.89], [710.68, 254.14], [688.88, 176.84], [750.1, 176.83], [785.09, 177.69], [710.68, 177.7]]}, {\"category\": \"car\", \"corners_3d\": [[660.45, 224.23], [709.46, 224.13], [733.23, 234.2], [674.62, 234.35], [660.45, 170.69], [709.46, 170.69], [733.23, 170.27], [674.62, 170.26]]}, {\"category\": \"car\", \"corners_3d\": [[649.93, 214.53], [691.61, 214.53], [706.19, 221.99], [657.05, 221.99], [649.93, 169.39], [691.61, 169.39], [706.19, 168.77], [657.05, 168.77]]}, {\"category\": \"car\", \"corners_3d\": [[555.92, 193.85], [535.92, 193.85], [540.09, 192.53], [558.82, 192.52], [555.92, 170.18], [535.92, 170.18], [540.09, 170.35], [558.82, 170.35]]}]\n```" - } - ] - }, - { - "id": 2282, - "data_source": "Mono3DRefer", - "scene_name": "002339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[467.73, 259.78], [439.99, 259.82], [456.69, 250.84], [481.55, 250.81], [467.73, 175.31], [439.99, 175.31], [456.69, 175.05], [481.55, 175.05]]}, {\"category\": \"cyclist\", \"corners_3d\": [[499.67, 231.04], [482.71, 231.04], [491.94, 226.79], [507.66, 226.79], [499.67, 170.85], [482.71, 170.85], [491.94, 170.99], [507.66, 170.99]]}]\n```" - } - ] - }, - { - "id": 2283, - "data_source": "Mono3DRefer", - "scene_name": "002339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[814.52, 239.91], [847.96, 239.89], [855.33, 241.9], [820.89, 241.93], [814.52, 159.9], [847.96, 159.91], [855.33, 159.52], [820.89, 159.51]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[625.97, 195.85], [622.01, 196.01], [609.79, 195.69], [613.78, 195.54], [625.97, 167.1], [622.01, 167.06], [609.79, 167.14], [613.78, 167.18]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[632.36, 196.18], [627.62, 196.37], [617.24, 196.08], [622.0, 195.9], [632.36, 168.21], [627.62, 168.18], [617.24, 168.23], [622.0, 168.27]]}]\n```" - } - ] - }, - { - "id": 2284, - "data_source": "Mono3DRefer", - "scene_name": "002339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[510.66, 209.84], [479.79, 209.84], [492.43, 206.22], [520.27, 206.22], [510.66, 178.13], [479.79, 178.13], [492.43, 177.62], [520.27, 177.62]]}, {\"category\": \"car\", \"corners_3d\": [[413.4, 253.81], [345.86, 253.74], [395.82, 238.79], [450.91, 238.84], [413.4, 186.04], [345.86, 186.03], [395.82, 183.59], [450.91, 183.6]]}, {\"category\": \"cyclist\", \"corners_3d\": [[467.73, 259.78], [439.99, 259.82], [456.69, 250.84], [481.55, 250.81], [467.73, 175.31], [439.99, 175.31], [456.69, 175.05], [481.55, 175.05]]}, {\"category\": \"cyclist\", \"corners_3d\": [[499.67, 231.04], [482.71, 231.04], [491.94, 226.79], [507.66, 226.79], [499.67, 170.85], [482.71, 170.85], [491.94, 170.99], [507.66, 170.99]]}, {\"category\": \"car\", \"corners_3d\": [[688.88, 239.64], [750.1, 239.47], [785.09, 253.89], [710.68, 254.14], [688.88, 176.84], [750.1, 176.83], [785.09, 177.69], [710.68, 177.7]]}, {\"category\": \"car\", \"corners_3d\": [[660.45, 224.23], [709.46, 224.13], [733.23, 234.2], [674.62, 234.35], [660.45, 170.69], [709.46, 170.69], [733.23, 170.27], [674.62, 170.26]]}, {\"category\": \"car\", \"corners_3d\": [[649.93, 214.53], [691.61, 214.53], [706.19, 221.99], [657.05, 221.99], [649.93, 169.39], [691.61, 169.39], [706.19, 168.77], [657.05, 168.77]]}, {\"category\": \"car\", \"corners_3d\": [[555.92, 193.85], [535.92, 193.85], [540.09, 192.53], [558.82, 192.52], [555.92, 170.18], [535.92, 170.18], [540.09, 170.35], [558.82, 170.35]]}]\n```" - } - ] - }, - { - "id": 2285, - "data_source": "Mono3DRefer", - "scene_name": "002339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[510.66, 209.84], [479.79, 209.84], [492.43, 206.22], [520.27, 206.22], [510.66, 178.13], [479.79, 178.13], [492.43, 177.62], [520.27, 177.62]]}, {\"category\": \"car\", \"corners_3d\": [[413.4, 253.81], [345.86, 253.74], [395.82, 238.79], [450.91, 238.84], [413.4, 186.04], [345.86, 186.03], [395.82, 183.59], [450.91, 183.6]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[814.52, 239.91], [847.96, 239.89], [855.33, 241.9], [820.89, 241.93], [814.52, 159.9], [847.96, 159.91], [855.33, 159.52], [820.89, 159.51]]}, {\"category\": \"car\", \"corners_3d\": [[688.88, 239.64], [750.1, 239.47], [785.09, 253.89], [710.68, 254.14], [688.88, 176.84], [750.1, 176.83], [785.09, 177.69], [710.68, 177.7]]}, {\"category\": \"car\", \"corners_3d\": [[660.45, 224.23], [709.46, 224.13], [733.23, 234.2], [674.62, 234.35], [660.45, 170.69], [709.46, 170.69], [733.23, 170.27], [674.62, 170.26]]}, {\"category\": \"car\", \"corners_3d\": [[649.93, 214.53], [691.61, 214.53], [706.19, 221.99], [657.05, 221.99], [649.93, 169.39], [691.61, 169.39], [706.19, 168.77], [657.05, 168.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[625.97, 195.85], [622.01, 196.01], [609.79, 195.69], [613.78, 195.54], [625.97, 167.1], [622.01, 167.06], [609.79, 167.14], [613.78, 167.18]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[632.36, 196.18], [627.62, 196.37], [617.24, 196.08], [622.0, 195.9], [632.36, 168.21], [627.62, 168.18], [617.24, 168.23], [622.0, 168.27]]}, {\"category\": \"car\", \"corners_3d\": [[555.92, 193.85], [535.92, 193.85], [540.09, 192.53], [558.82, 192.52], [555.92, 170.18], [535.92, 170.18], [540.09, 170.35], [558.82, 170.35]]}]\n```" - } - ] - }, - { - "id": 2286, - "data_source": "Mono3DRefer", - "scene_name": "002339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[814.52, 239.91], [847.96, 239.89], [855.33, 241.9], [820.89, 241.93], [814.52, 159.9], [847.96, 159.91], [855.33, 159.52], [820.89, 159.51]]}, {\"category\": \"cyclist\", \"corners_3d\": [[467.73, 259.78], [439.99, 259.82], [456.69, 250.84], [481.55, 250.81], [467.73, 175.31], [439.99, 175.31], [456.69, 175.05], [481.55, 175.05]]}, {\"category\": \"cyclist\", \"corners_3d\": [[499.67, 231.04], [482.71, 231.04], [491.94, 226.79], [507.66, 226.79], [499.67, 170.85], [482.71, 170.85], [491.94, 170.99], [507.66, 170.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[625.97, 195.85], [622.01, 196.01], [609.79, 195.69], [613.78, 195.54], [625.97, 167.1], [622.01, 167.06], [609.79, 167.14], [613.78, 167.18]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[632.36, 196.18], [627.62, 196.37], [617.24, 196.08], [622.0, 195.9], [632.36, 168.21], [627.62, 168.18], [617.24, 168.23], [622.0, 168.27]]}]\n```" - } - ] - }, - { - "id": 2287, - "data_source": "Mono3DRefer", - "scene_name": "002339", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[510.66, 209.84], [479.79, 209.84], [492.43, 206.22], [520.27, 206.22], [510.66, 178.13], [479.79, 178.13], [492.43, 177.62], [520.27, 177.62]]}, {\"category\": \"car\", \"corners_3d\": [[413.4, 253.81], [345.86, 253.74], [395.82, 238.79], [450.91, 238.84], [413.4, 186.04], [345.86, 186.03], [395.82, 183.59], [450.91, 183.6]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[814.52, 239.91], [847.96, 239.89], [855.33, 241.9], [820.89, 241.93], [814.52, 159.9], [847.96, 159.91], [855.33, 159.52], [820.89, 159.51]]}, {\"category\": \"cyclist\", \"corners_3d\": [[467.73, 259.78], [439.99, 259.82], [456.69, 250.84], [481.55, 250.81], [467.73, 175.31], [439.99, 175.31], [456.69, 175.05], [481.55, 175.05]]}, {\"category\": \"cyclist\", \"corners_3d\": [[499.67, 231.04], [482.71, 231.04], [491.94, 226.79], [507.66, 226.79], [499.67, 170.85], [482.71, 170.85], [491.94, 170.99], [507.66, 170.99]]}, {\"category\": \"car\", \"corners_3d\": [[688.88, 239.64], [750.1, 239.47], [785.09, 253.89], [710.68, 254.14], [688.88, 176.84], [750.1, 176.83], [785.09, 177.69], [710.68, 177.7]]}, {\"category\": \"car\", \"corners_3d\": [[660.45, 224.23], [709.46, 224.13], [733.23, 234.2], [674.62, 234.35], [660.45, 170.69], [709.46, 170.69], [733.23, 170.27], [674.62, 170.26]]}, {\"category\": \"car\", \"corners_3d\": [[649.93, 214.53], [691.61, 214.53], [706.19, 221.99], [657.05, 221.99], [649.93, 169.39], [691.61, 169.39], [706.19, 168.77], [657.05, 168.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[625.97, 195.85], [622.01, 196.01], [609.79, 195.69], [613.78, 195.54], [625.97, 167.1], [622.01, 167.06], [609.79, 167.14], [613.78, 167.18]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[632.36, 196.18], [627.62, 196.37], [617.24, 196.08], [622.0, 195.9], [632.36, 168.21], [627.62, 168.18], [617.24, 168.23], [622.0, 168.27]]}, {\"category\": \"car\", \"corners_3d\": [[555.92, 193.85], [535.92, 193.85], [540.09, 192.53], [558.82, 192.52], [555.92, 170.18], [535.92, 170.18], [540.09, 170.35], [558.82, 170.35]]}]\n```" - } - ] - }, - { - "id": 2288, - "data_source": "Mono3DRefer", - "scene_name": "002343", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[550.53, 195.28], [527.95, 195.28], [533.94, 193.62], [554.85, 193.62], [550.53, 171.4], [527.95, 171.4], [533.94, 171.51], [554.85, 171.51]]}]\n```" - } - ] - }, - { - "id": 2289, - "data_source": "Mono3DRefer", - "scene_name": "002350", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[550.64, 242.3], [492.42, 242.08], [515.62, 231.04], [564.59, 231.19], [550.64, 186.45], [492.42, 186.41], [515.62, 184.25], [564.59, 184.28]]}, {\"category\": \"car\", \"corners_3d\": [[599.35, 206.95], [575.2, 206.92], [579.44, 204.31], [601.75, 204.34], [599.35, 186.26], [575.2, 186.24], [579.44, 185.22], [601.75, 185.23]]}, {\"category\": \"car\", \"corners_3d\": [[1379.16, 259.45], [1428.59, 270.58], [1063.6, 254.15], [1056.92, 246.3], [1379.16, 125.06], [1428.59, 118.92], [1063.6, 127.98], [1056.92, 132.32]]}, {\"category\": \"car\", \"corners_3d\": [[422.15, 272.5], [510.4, 273.13], [443.26, 322.7], [312.42, 321.3], [422.15, 182.5], [510.4, 182.56], [443.26, 187.35], [312.42, 187.22]]}]\n```" - } - ] - }, - { - "id": 2290, - "data_source": "Mono3DRefer", - "scene_name": "002373", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[576.39, 204.63], [602.18, 204.63], [601.39, 207.77], [573.05, 207.77], [576.39, 180.83], [602.18, 180.83], [601.39, 181.63], [573.05, 181.62]]}, {\"category\": \"car\", \"corners_3d\": [[945.26, 303.18], [826.7, 302.77], [765.78, 263.84], [848.6, 264.04], [945.26, 191.87], [826.7, 191.81], [765.78, 186.13], [848.6, 186.16]]}, {\"category\": \"car\", \"corners_3d\": [[839.76, 233.95], [790.55, 233.99], [763.45, 225.21], [805.61, 225.18], [839.76, 186.57], [790.55, 186.58], [763.45, 184.61], [805.61, 184.6]]}, {\"category\": \"car\", \"corners_3d\": [[653.62, 209.64], [649.39, 208.02], [729.58, 207.74], [737.44, 209.33], [653.62, 179.11], [649.39, 178.84], [729.58, 178.79], [737.44, 179.06]]}]\n```" - } - ] - }, - { - "id": 2291, - "data_source": "Mono3DRefer", - "scene_name": "002374", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[567.71, 220.24], [535.88, 220.48], [535.26, 215.96], [564.08, 215.77], [567.71, 187.61], [535.88, 187.68], [535.26, 186.27], [564.08, 186.22]]}, {\"category\": \"car\", \"corners_3d\": [[530.57, 205.84], [510.26, 206.07], [504.61, 203.88], [523.62, 203.68], [530.57, 188.33], [510.26, 188.44], [504.61, 187.41], [523.62, 187.32]]}]\n```" - } - ] - }, - { - "id": 2292, - "data_source": "Mono3DRefer", - "scene_name": "002378", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[953.81, 465.58], [1217.69, 462.41], [4632.49, 2033.85], [3084.9, 2173.14], [953.81, 215.35], [1217.69, 214.89], [4632.49, 443.02], [3084.9, 463.24]]}, {\"category\": \"car\", \"corners_3d\": [[343.3, 345.59], [164.07, 344.76], [326.97, 283.8], [442.95, 284.14], [343.3, 177.13], [164.07, 177.11], [326.97, 175.6], [442.95, 175.61]]}, {\"category\": \"car\", \"corners_3d\": [[1061.08, 392.87], [857.24, 392.92], [748.8, 296.8], [863.63, 296.78], [1061.08, 213.84], [857.24, 213.85], [748.8, 195.94], [863.63, 195.94]]}, {\"category\": \"car\", \"corners_3d\": [[500.32, 238.52], [430.13, 238.65], [458.45, 227.32], [516.52, 227.23], [500.32, 170.99], [430.13, 170.98], [458.45, 171.31], [516.52, 171.31]]}, {\"category\": \"car\", \"corners_3d\": [[749.93, 227.9], [692.28, 227.86], [679.07, 218.3], [726.69, 218.33], [749.93, 177.91], [692.28, 177.91], [679.07, 177.03], [726.69, 177.03]]}, {\"category\": \"car\", \"corners_3d\": [[542.49, 217.92], [488.21, 218.02], [502.32, 211.59], [548.86, 211.52], [542.49, 173.8], [488.21, 173.8], [502.32, 173.67], [548.86, 173.67]]}, {\"category\": \"car\", \"corners_3d\": [[652.04, 197.03], [677.58, 197.05], [682.59, 199.34], [654.64, 199.32], [652.04, 172.36], [677.58, 172.36], [682.59, 172.31], [654.64, 172.31]]}]\n```" - } - ] - }, - { - "id": 2293, - "data_source": "Mono3DRefer", - "scene_name": "002378", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[642.85, 189.39], [667.7, 189.4], [671.59, 191.2], [644.04, 191.17], [642.85, 159.57], [667.7, 159.56], [671.59, 158.12], [644.04, 158.13]]}]\n```" - } - ] - }, - { - "id": 2294, - "data_source": "Mono3DRefer", - "scene_name": "002378", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[953.81, 465.58], [1217.69, 462.41], [4632.49, 2033.85], [3084.9, 2173.14], [953.81, 215.35], [1217.69, 214.89], [4632.49, 443.02], [3084.9, 463.24]]}, {\"category\": \"car\", \"corners_3d\": [[343.3, 345.59], [164.07, 344.76], [326.97, 283.8], [442.95, 284.14], [343.3, 177.13], [164.07, 177.11], [326.97, 175.6], [442.95, 175.61]]}, {\"category\": \"car\", \"corners_3d\": [[1061.08, 392.87], [857.24, 392.92], [748.8, 296.8], [863.63, 296.78], [1061.08, 213.84], [857.24, 213.85], [748.8, 195.94], [863.63, 195.94]]}, {\"category\": \"car\", \"corners_3d\": [[500.32, 238.52], [430.13, 238.65], [458.45, 227.32], [516.52, 227.23], [500.32, 170.99], [430.13, 170.98], [458.45, 171.31], [516.52, 171.31]]}, {\"category\": \"car\", \"corners_3d\": [[749.93, 227.9], [692.28, 227.86], [679.07, 218.3], [726.69, 218.33], [749.93, 177.91], [692.28, 177.91], [679.07, 177.03], [726.69, 177.03]]}, {\"category\": \"car\", \"corners_3d\": [[542.49, 217.92], [488.21, 218.02], [502.32, 211.59], [548.86, 211.52], [542.49, 173.8], [488.21, 173.8], [502.32, 173.67], [548.86, 173.67]]}, {\"category\": \"car\", \"corners_3d\": [[652.04, 197.03], [677.58, 197.05], [682.59, 199.34], [654.64, 199.32], [652.04, 172.36], [677.58, 172.36], [682.59, 172.31], [654.64, 172.31]]}, {\"category\": \"van\", \"corners_3d\": [[642.85, 189.39], [667.7, 189.4], [671.59, 191.2], [644.04, 191.17], [642.85, 159.57], [667.7, 159.56], [671.59, 158.12], [644.04, 158.13]]}]\n```" - } - ] - }, - { - "id": 2295, - "data_source": "Mono3DRefer", - "scene_name": "002380", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[657.62, 213.43], [709.55, 213.41], [740.42, 225.14], [673.47, 225.19], [657.62, 151.62], [709.55, 151.64], [740.42, 145.49], [673.47, 145.47]]}, {\"category\": \"van\", \"corners_3d\": [[633.99, 200.47], [665.89, 200.44], [678.05, 205.22], [640.63, 205.26], [633.99, 159.2], [665.89, 159.21], [678.05, 156.85], [640.63, 156.83]]}, {\"category\": \"van\", \"corners_3d\": [[564.0, 194.53], [562.4, 195.18], [494.93, 195.16], [498.51, 194.51], [564.0, 162.24], [562.4, 161.92], [494.93, 161.93], [498.51, 162.25]]}]\n```" - } - ] - }, - { - "id": 2296, - "data_source": "Mono3DRefer", - "scene_name": "002380", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1478.33, 352.05], [1702.08, 402.87], [1199.09, 387.7], [1092.33, 342.71], [1478.33, 181.2], [1702.08, 183.57], [1199.09, 182.86], [1092.33, 180.77]]}, {\"category\": \"car\", \"corners_3d\": [[843.49, 263.5], [815.05, 253.17], [1024.02, 252.66], [1078.95, 262.86], [843.49, 167.95], [815.05, 168.51], [1024.02, 168.53], [1078.95, 167.98]]}, {\"category\": \"car\", \"corners_3d\": [[311.96, 280.53], [260.11, 299.28], [-45.19, 299.28], [51.96, 280.53], [311.96, 139.06], [260.11, 133.17], [-45.19, 133.17], [51.96, 139.06]]}, {\"category\": \"car\", \"corners_3d\": [[111.54, 268.1], [163.89, 257.91], [335.32, 257.68], [303.59, 267.81], [111.54, 185.26], [163.89, 183.93], [335.32, 183.9], [303.59, 185.22]]}, {\"category\": \"car\", \"corners_3d\": [[284.14, 238.42], [309.16, 233.36], [458.26, 233.34], [445.71, 238.4], [284.14, 183.71], [309.16, 182.87], [458.26, 182.87], [445.71, 183.7]]}, {\"category\": \"car\", \"corners_3d\": [[362.95, 225.53], [373.77, 221.93], [498.86, 220.93], [497.24, 224.38], [362.95, 171.34], [373.77, 171.44], [498.86, 171.47], [497.24, 171.37]]}, {\"category\": \"car\", \"corners_3d\": [[448.04, 206.18], [454.67, 204.94], [512.35, 204.99], [507.96, 206.24], [448.04, 177.4], [454.67, 177.23], [512.35, 177.24], [507.96, 177.41]]}]\n```" - } - ] - }, - { - "id": 2297, - "data_source": "Mono3DRefer", - "scene_name": "002380", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1478.33, 352.05], [1702.08, 402.87], [1199.09, 387.7], [1092.33, 342.71], [1478.33, 181.2], [1702.08, 183.57], [1199.09, 182.86], [1092.33, 180.77]]}, {\"category\": \"car\", \"corners_3d\": [[843.49, 263.5], [815.05, 253.17], [1024.02, 252.66], [1078.95, 262.86], [843.49, 167.95], [815.05, 168.51], [1024.02, 168.53], [1078.95, 167.98]]}, {\"category\": \"car\", \"corners_3d\": [[311.96, 280.53], [260.11, 299.28], [-45.19, 299.28], [51.96, 280.53], [311.96, 139.06], [260.11, 133.17], [-45.19, 133.17], [51.96, 139.06]]}, {\"category\": \"car\", \"corners_3d\": [[111.54, 268.1], [163.89, 257.91], [335.32, 257.68], [303.59, 267.81], [111.54, 185.26], [163.89, 183.93], [335.32, 183.9], [303.59, 185.22]]}, {\"category\": \"van\", \"corners_3d\": [[657.62, 213.43], [709.55, 213.41], [740.42, 225.14], [673.47, 225.19], [657.62, 151.62], [709.55, 151.64], [740.42, 145.49], [673.47, 145.47]]}, {\"category\": \"car\", \"corners_3d\": [[284.14, 238.42], [309.16, 233.36], [458.26, 233.34], [445.71, 238.4], [284.14, 183.71], [309.16, 182.87], [458.26, 182.87], [445.71, 183.7]]}, {\"category\": \"car\", \"corners_3d\": [[362.95, 225.53], [373.77, 221.93], [498.86, 220.93], [497.24, 224.38], [362.95, 171.34], [373.77, 171.44], [498.86, 171.47], [497.24, 171.37]]}, {\"category\": \"van\", \"corners_3d\": [[633.99, 200.47], [665.89, 200.44], [678.05, 205.22], [640.63, 205.26], [633.99, 159.2], [665.89, 159.21], [678.05, 156.85], [640.63, 156.83]]}, {\"category\": \"car\", \"corners_3d\": [[448.04, 206.18], [454.67, 204.94], [512.35, 204.99], [507.96, 206.24], [448.04, 177.4], [454.67, 177.23], [512.35, 177.24], [507.96, 177.41]]}, {\"category\": \"van\", \"corners_3d\": [[564.0, 194.53], [562.4, 195.18], [494.93, 195.16], [498.51, 194.51], [564.0, 162.24], [562.4, 161.92], [494.93, 161.93], [498.51, 162.25]]}]\n```" - } - ] - }, - { - "id": 2298, - "data_source": "Mono3DRefer", - "scene_name": "002400", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[503.37, 207.29], [499.14, 208.76], [429.4, 208.8], [436.49, 207.32], [503.37, 179.88], [499.14, 180.18], [429.4, 180.19], [436.49, 179.89]]}, {\"category\": \"car\", \"corners_3d\": [[-278.97, 336.6], [-138.98, 310.55], [160.82, 309.9], [78.0, 335.67], [-278.97, 202.76], [-138.98, 198.0], [160.82, 197.88], [78.0, 202.59]]}, {\"category\": \"car\", \"corners_3d\": [[234.39, 244.05], [266.67, 238.17], [408.16, 238.45], [388.52, 244.38], [234.39, 174.67], [266.67, 174.52], [408.16, 174.52], [388.52, 174.68]]}]\n```" - } - ] - }, - { - "id": 2299, - "data_source": "Mono3DRefer", - "scene_name": "002400", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[873.13, 212.66], [888.76, 212.66], [912.5, 216.05], [895.55, 216.05], [873.13, 156.92], [888.76, 156.92], [912.5, 155.56], [895.55, 155.56]]}, {\"category\": \"cyclist\", \"corners_3d\": [[827.3, 207.45], [838.57, 207.45], [853.61, 209.73], [841.59, 209.73], [827.3, 164.98], [838.57, 164.98], [853.61, 164.46], [841.59, 164.46]]}, {\"category\": \"cyclist\", \"corners_3d\": [[802.97, 202.99], [812.68, 203.0], [823.96, 204.74], [813.69, 204.73], [802.97, 167.34], [812.68, 167.34], [823.96, 167.03], [813.69, 167.03]]}]\n```" - } - ] - }, - { - "id": 2300, - "data_source": "Mono3DRefer", - "scene_name": "002400", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[503.37, 207.29], [499.14, 208.76], [429.4, 208.8], [436.49, 207.32], [503.37, 179.88], [499.14, 180.18], [429.4, 180.19], [436.49, 179.89]]}, {\"category\": \"cyclist\", \"corners_3d\": [[873.13, 212.66], [888.76, 212.66], [912.5, 216.05], [895.55, 216.05], [873.13, 156.92], [888.76, 156.92], [912.5, 155.56], [895.55, 155.56]]}, {\"category\": \"cyclist\", \"corners_3d\": [[827.3, 207.45], [838.57, 207.45], [853.61, 209.73], [841.59, 209.73], [827.3, 164.98], [838.57, 164.98], [853.61, 164.46], [841.59, 164.46]]}, {\"category\": \"cyclist\", \"corners_3d\": [[802.97, 202.99], [812.68, 203.0], [823.96, 204.74], [813.69, 204.73], [802.97, 167.34], [812.68, 167.34], [823.96, 167.03], [813.69, 167.03]]}, {\"category\": \"car\", \"corners_3d\": [[-278.97, 336.6], [-138.98, 310.55], [160.82, 309.9], [78.0, 335.67], [-278.97, 202.76], [-138.98, 198.0], [160.82, 197.88], [78.0, 202.59]]}, {\"category\": \"car\", \"corners_3d\": [[234.39, 244.05], [266.67, 238.17], [408.16, 238.45], [388.52, 244.38], [234.39, 174.67], [266.67, 174.52], [408.16, 174.52], [388.52, 174.68]]}]\n```" - } - ] - }, - { - "id": 2301, - "data_source": "Mono3DRefer", - "scene_name": "002402", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[613.69, 356.11], [524.0, 356.35], [533.83, 332.49], [611.35, 332.31], [613.69, 146.72], [524.0, 146.67], [533.83, 151.26], [611.35, 151.3]]}]\n```" - } - ] - }, - { - "id": 2302, - "data_source": "Mono3DRefer", - "scene_name": "002408", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[814.58, 295.74], [800.06, 290.36], [871.88, 288.71], [889.62, 293.92], [814.58, 153.55], [800.06, 154.81], [871.88, 155.2], [889.62, 153.98]]}]\n```" - } - ] - }, - { - "id": 2303, - "data_source": "Mono3DRefer", - "scene_name": "002416", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[331.43, 223.63], [347.32, 223.05], [359.93, 226.23], [343.17, 226.88], [331.43, 182.08], [347.32, 181.97], [359.93, 182.55], [343.17, 182.67]]}]\n```" - } - ] - }, - { - "id": 2304, - "data_source": "Mono3DRefer", - "scene_name": "002416", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[500.63, 253.58], [479.58, 246.78], [633.94, 239.56], [665.46, 245.04], [500.63, 171.96], [479.58, 172.03], [633.94, 172.11], [665.46, 172.05]]}, {\"category\": \"car\", \"corners_3d\": [[-568.76, 385.2], [-431.28, 349.57], [56.24, 314.99], [25.92, 337.17], [-568.76, 224.85], [-431.28, 216.13], [56.24, 207.66], [25.92, 213.09]]}, {\"category\": \"car\", \"corners_3d\": [[-126.78, 277.86], [-89.18, 268.09], [127.52, 258.98], [113.76, 266.89], [-126.78, 203.14], [-89.18, 200.32], [127.52, 197.69], [113.76, 199.98]]}, {\"category\": \"car\", \"corners_3d\": [[9.11, 248.72], [29.27, 243.52], [164.19, 239.34], [154.61, 243.92], [9.11, 198.14], [29.27, 196.4], [164.19, 195.01], [154.61, 196.54]]}, {\"category\": \"car\", \"corners_3d\": [[116.49, 223.91], [123.32, 221.74], [213.15, 219.6], [210.44, 221.58], [116.49, 191.62], [123.32, 190.82], [213.15, 190.04], [210.44, 190.77]]}, {\"category\": \"car\", \"corners_3d\": [[1288.05, 357.0], [1595.29, 411.61], [1400.31, 510.53], [1061.01, 410.75], [1288.05, 195.98], [1595.29, 202.84], [1400.31, 215.26], [1061.01, 202.73]]}, {\"category\": \"car\", \"corners_3d\": [[1113.32, 310.07], [1270.16, 337.83], [935.71, 381.74], [813.83, 339.15], [1113.32, 187.89], [1270.16, 190.93], [935.71, 195.74], [813.83, 191.07]]}, {\"category\": \"car\", \"corners_3d\": [[748.09, 255.94], [798.5, 265.03], [599.41, 277.89], [564.82, 266.24], [748.09, 185.98], [798.5, 187.42], [599.41, 189.45], [564.82, 187.61]]}]\n```" - } - ] - }, - { - "id": 2305, - "data_source": "Mono3DRefer", - "scene_name": "002416", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[500.63, 253.58], [479.58, 246.78], [633.94, 239.56], [665.46, 245.04], [500.63, 171.96], [479.58, 172.03], [633.94, 172.11], [665.46, 172.05]]}, {\"category\": \"car\", \"corners_3d\": [[-568.76, 385.2], [-431.28, 349.57], [56.24, 314.99], [25.92, 337.17], [-568.76, 224.85], [-431.28, 216.13], [56.24, 207.66], [25.92, 213.09]]}, {\"category\": \"cyclist\", \"corners_3d\": [[331.43, 223.63], [347.32, 223.05], [359.93, 226.23], [343.17, 226.88], [331.43, 182.08], [347.32, 181.97], [359.93, 182.55], [343.17, 182.67]]}, {\"category\": \"car\", \"corners_3d\": [[-126.78, 277.86], [-89.18, 268.09], [127.52, 258.98], [113.76, 266.89], [-126.78, 203.14], [-89.18, 200.32], [127.52, 197.69], [113.76, 199.98]]}, {\"category\": \"car\", \"corners_3d\": [[9.11, 248.72], [29.27, 243.52], [164.19, 239.34], [154.61, 243.92], [9.11, 198.14], [29.27, 196.4], [164.19, 195.01], [154.61, 196.54]]}, {\"category\": \"car\", \"corners_3d\": [[116.49, 223.91], [123.32, 221.74], [213.15, 219.6], [210.44, 221.58], [116.49, 191.62], [123.32, 190.82], [213.15, 190.04], [210.44, 190.77]]}, {\"category\": \"car\", \"corners_3d\": [[1288.05, 357.0], [1595.29, 411.61], [1400.31, 510.53], [1061.01, 410.75], [1288.05, 195.98], [1595.29, 202.84], [1400.31, 215.26], [1061.01, 202.73]]}, {\"category\": \"car\", \"corners_3d\": [[1113.32, 310.07], [1270.16, 337.83], [935.71, 381.74], [813.83, 339.15], [1113.32, 187.89], [1270.16, 190.93], [935.71, 195.74], [813.83, 191.07]]}, {\"category\": \"car\", \"corners_3d\": [[748.09, 255.94], [798.5, 265.03], [599.41, 277.89], [564.82, 266.24], [748.09, 185.98], [798.5, 187.42], [599.41, 189.45], [564.82, 187.61]]}]\n```" - } - ] - }, - { - "id": 2306, - "data_source": "Mono3DRefer", - "scene_name": "002419", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[607.87, 351.52], [738.14, 336.78], [1050.76, 448.47], [867.79, 492.89], [607.87, 206.39], [738.14, 203.62], [1050.76, 224.58], [867.79, 232.92]]}, {\"category\": \"car\", \"corners_3d\": [[484.48, 288.27], [579.45, 282.14], [709.75, 332.49], [578.29, 345.91], [484.48, 206.33], [579.45, 204.55], [709.75, 219.15], [578.29, 223.04]]}, {\"category\": \"car\", \"corners_3d\": [[419.86, 272.01], [489.55, 267.91], [551.9, 293.93], [465.53, 300.65], [419.86, 207.73], [489.55, 206.29], [551.9, 215.44], [465.53, 217.81]]}, {\"category\": \"car\", \"corners_3d\": [[369.6, 258.19], [428.13, 255.32], [459.76, 270.99], [390.96, 275.1], [369.6, 201.99], [428.13, 201.01], [459.76, 206.37], [390.96, 207.77]]}, {\"category\": \"car\", \"corners_3d\": [[188.84, 272.93], [126.64, 275.87], [153.42, 260.29], [205.8, 258.17], [188.84, 212.88], [126.64, 214.05], [153.42, 207.82], [205.8, 206.97]]}, {\"category\": \"car\", \"corners_3d\": [[133.73, 314.05], [28.93, 322.45], [63.92, 292.75], [147.28, 287.29], [133.73, 231.82], [28.93, 235.32], [63.92, 222.92], [147.28, 220.64]]}, {\"category\": \"car\", \"corners_3d\": [[6.66, 417.14], [-242.34, 447.25], [-50.06, 338.39], [93.68, 326.93], [6.66, 252.76], [-242.34, 262.6], [-50.06, 227.0], [93.68, 223.25]]}, {\"category\": \"car\", \"corners_3d\": [[324.9, 241.2], [363.09, 239.8], [374.36, 248.81], [331.18, 250.62], [324.9, 210.29], [363.09, 209.53], [374.36, 214.46], [331.18, 215.45]]}, {\"category\": \"car\", \"corners_3d\": [[226.67, 242.07], [195.86, 243.26], [199.5, 237.74], [227.87, 236.73], [226.67, 216.0], [195.86, 216.74], [199.5, 213.3], [227.87, 212.68]]}]\n```" - } - ] - }, - { - "id": 2307, - "data_source": "Mono3DRefer", - "scene_name": "002423", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[530.17, 196.15], [529.33, 197.06], [450.74, 197.27], [454.55, 196.35], [530.17, 166.23], [529.33, 165.98], [450.74, 165.92], [454.55, 166.18]]}]\n```" - } - ] - }, - { - "id": 2308, - "data_source": "Mono3DRefer", - "scene_name": "002423", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[597.54, 192.08], [623.18, 192.04], [628.22, 194.01], [599.95, 194.05], [597.54, 167.4], [623.18, 167.41], [628.22, 166.85], [599.95, 166.84]]}]\n```" - } - ] - }, - { - "id": 2309, - "data_source": "Mono3DRefer", - "scene_name": "002423", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[250.25, 238.58], [272.25, 233.64], [424.82, 232.51], [415.4, 237.26], [250.25, 182.07], [272.25, 181.38], [424.82, 181.22], [415.4, 181.88]]}, {\"category\": \"car\", \"corners_3d\": [[650.2, 234.58], [711.06, 234.17], [755.75, 251.43], [678.0, 252.11], [650.2, 172.08], [711.06, 172.08], [755.75, 171.87], [678.0, 171.86]]}, {\"category\": \"car\", \"corners_3d\": [[470.74, 276.94], [377.1, 277.09], [429.24, 252.9], [501.12, 252.81], [470.74, 191.24], [377.1, 191.27], [429.24, 186.99], [501.12, 186.98]]}]\n```" - } - ] - }, - { - "id": 2310, - "data_source": "Mono3DRefer", - "scene_name": "002423", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[597.54, 192.08], [623.18, 192.04], [628.22, 194.01], [599.95, 194.05], [597.54, 167.4], [623.18, 167.41], [628.22, 166.85], [599.95, 166.84]]}, {\"category\": \"van\", \"corners_3d\": [[530.17, 196.15], [529.33, 197.06], [450.74, 197.27], [454.55, 196.35], [530.17, 166.23], [529.33, 165.98], [450.74, 165.92], [454.55, 166.18]]}]\n```" - } - ] - }, - { - "id": 2311, - "data_source": "Mono3DRefer", - "scene_name": "002423", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[530.17, 196.15], [529.33, 197.06], [450.74, 197.27], [454.55, 196.35], [530.17, 166.23], [529.33, 165.98], [450.74, 165.92], [454.55, 166.18]]}, {\"category\": \"car\", \"corners_3d\": [[250.25, 238.58], [272.25, 233.64], [424.82, 232.51], [415.4, 237.26], [250.25, 182.07], [272.25, 181.38], [424.82, 181.22], [415.4, 181.88]]}, {\"category\": \"car\", \"corners_3d\": [[650.2, 234.58], [711.06, 234.17], [755.75, 251.43], [678.0, 252.11], [650.2, 172.08], [711.06, 172.08], [755.75, 171.87], [678.0, 171.86]]}, {\"category\": \"car\", \"corners_3d\": [[470.74, 276.94], [377.1, 277.09], [429.24, 252.9], [501.12, 252.81], [470.74, 191.24], [377.1, 191.27], [429.24, 186.99], [501.12, 186.98]]}]\n```" - } - ] - }, - { - "id": 2312, - "data_source": "Mono3DRefer", - "scene_name": "002423", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[597.54, 192.08], [623.18, 192.04], [628.22, 194.01], [599.95, 194.05], [597.54, 167.4], [623.18, 167.41], [628.22, 166.85], [599.95, 166.84]]}, {\"category\": \"car\", \"corners_3d\": [[250.25, 238.58], [272.25, 233.64], [424.82, 232.51], [415.4, 237.26], [250.25, 182.07], [272.25, 181.38], [424.82, 181.22], [415.4, 181.88]]}, {\"category\": \"car\", \"corners_3d\": [[650.2, 234.58], [711.06, 234.17], [755.75, 251.43], [678.0, 252.11], [650.2, 172.08], [711.06, 172.08], [755.75, 171.87], [678.0, 171.86]]}, {\"category\": \"car\", \"corners_3d\": [[470.74, 276.94], [377.1, 277.09], [429.24, 252.9], [501.12, 252.81], [470.74, 191.24], [377.1, 191.27], [429.24, 186.99], [501.12, 186.98]]}]\n```" - } - ] - }, - { - "id": 2313, - "data_source": "Mono3DRefer", - "scene_name": "002423", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[597.54, 192.08], [623.18, 192.04], [628.22, 194.01], [599.95, 194.05], [597.54, 167.4], [623.18, 167.41], [628.22, 166.85], [599.95, 166.84]]}, {\"category\": \"van\", \"corners_3d\": [[530.17, 196.15], [529.33, 197.06], [450.74, 197.27], [454.55, 196.35], [530.17, 166.23], [529.33, 165.98], [450.74, 165.92], [454.55, 166.18]]}, {\"category\": \"car\", \"corners_3d\": [[250.25, 238.58], [272.25, 233.64], [424.82, 232.51], [415.4, 237.26], [250.25, 182.07], [272.25, 181.38], [424.82, 181.22], [415.4, 181.88]]}, {\"category\": \"car\", \"corners_3d\": [[650.2, 234.58], [711.06, 234.17], [755.75, 251.43], [678.0, 252.11], [650.2, 172.08], [711.06, 172.08], [755.75, 171.87], [678.0, 171.86]]}, {\"category\": \"car\", \"corners_3d\": [[470.74, 276.94], [377.1, 277.09], [429.24, 252.9], [501.12, 252.81], [470.74, 191.24], [377.1, 191.27], [429.24, 186.99], [501.12, 186.98]]}]\n```" - } - ] - }, - { - "id": 2314, - "data_source": "Mono3DRefer", - "scene_name": "002424", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[167.7, 266.34], [180.32, 268.45], [132.25, 272.9], [120.23, 270.57], [167.7, 168.51], [180.32, 168.21], [132.25, 167.59], [120.23, 167.92]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[237.32, 255.97], [248.05, 257.82], [198.29, 261.75], [188.27, 259.71], [237.32, 170.99], [248.05, 170.75], [198.29, 170.25], [188.27, 170.51]]}]\n```" - } - ] - }, - { - "id": 2315, - "data_source": "Mono3DRefer", - "scene_name": "002424", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[669.16, 308.38], [697.04, 316.61], [566.73, 330.62], [544.81, 320.67], [669.16, 176.42], [697.04, 176.15], [566.73, 175.71], [544.81, 176.02]]}]\n```" - } - ] - }, - { - "id": 2316, - "data_source": "Mono3DRefer", - "scene_name": "002424", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[669.16, 308.38], [697.04, 316.61], [566.73, 330.62], [544.81, 320.67], [669.16, 176.42], [697.04, 176.15], [566.73, 175.71], [544.81, 176.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[167.7, 266.34], [180.32, 268.45], [132.25, 272.9], [120.23, 270.57], [167.7, 168.51], [180.32, 168.21], [132.25, 167.59], [120.23, 167.92]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[237.32, 255.97], [248.05, 257.82], [198.29, 261.75], [188.27, 259.71], [237.32, 170.99], [248.05, 170.75], [198.29, 170.25], [188.27, 170.51]]}]\n```" - } - ] - }, - { - "id": 2317, - "data_source": "Mono3DRefer", - "scene_name": "002428", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[518.04, 251.14], [443.79, 251.23], [475.22, 235.63], [534.68, 235.58], [518.04, 189.35], [443.79, 189.36], [475.22, 186.08], [534.68, 186.07]]}]\n```" - } - ] - }, - { - "id": 2318, - "data_source": "Mono3DRefer", - "scene_name": "002428", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-1146.2, 1282.98], [-2150.83, 1268.43], [-2.15, 417.68], [226.38, 418.4], [-1146.2, 147.54], [-2150.83, 147.87], [-2.15, 167.27], [226.38, 167.25]]}, {\"category\": \"van\", \"corners_3d\": [[596.31, 225.66], [618.19, 225.58], [621.69, 229.76], [598.08, 229.85], [596.31, 200.95], [618.19, 200.91], [621.69, 203.13], [598.08, 203.18]]}]\n```" - } - ] - }, - { - "id": 2319, - "data_source": "Mono3DRefer", - "scene_name": "002428", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-1146.2, 1282.98], [-2150.83, 1268.43], [-2.15, 417.68], [226.38, 418.4], [-1146.2, 147.54], [-2150.83, 147.87], [-2.15, 167.27], [226.38, 167.25]]}, {\"category\": \"car\", \"corners_3d\": [[518.04, 251.14], [443.79, 251.23], [475.22, 235.63], [534.68, 235.58], [518.04, 189.35], [443.79, 189.36], [475.22, 186.08], [534.68, 186.07]]}, {\"category\": \"van\", \"corners_3d\": [[596.31, 225.66], [618.19, 225.58], [621.69, 229.76], [598.08, 229.85], [596.31, 200.95], [618.19, 200.91], [621.69, 203.13], [598.08, 203.18]]}]\n```" - } - ] - }, - { - "id": 2320, - "data_source": "Mono3DRefer", - "scene_name": "002431", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[561.55, 227.01], [510.76, 227.35], [516.2, 218.4], [558.66, 218.15], [561.55, 182.35], [510.76, 182.41], [516.2, 180.84], [558.66, 180.79]]}]\n```" - } - ] - }, - { - "id": 2321, - "data_source": "Mono3DRefer", - "scene_name": "002432", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[318.34, 222.6], [278.37, 222.55], [314.11, 217.4], [349.97, 217.45], [318.34, 187.7], [278.37, 187.69], [314.11, 186.15], [349.97, 186.16]]}, {\"category\": \"car\", \"corners_3d\": [[445.03, 197.31], [424.15, 197.29], [437.24, 195.74], [456.81, 195.76], [445.03, 180.12], [424.15, 180.12], [437.24, 179.66], [456.81, 179.66]]}]\n```" - } - ] - }, - { - "id": 2322, - "data_source": "Mono3DRefer", - "scene_name": "002432", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[663.59, 212.43], [725.93, 212.43], [804.87, 239.41], [700.03, 239.4], [663.59, 144.88], [725.93, 144.88], [804.87, 125.8], [700.03, 125.81]]}]\n```" - } - ] - }, - { - "id": 2323, - "data_source": "Mono3DRefer", - "scene_name": "002432", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[663.59, 212.43], [725.93, 212.43], [804.87, 239.41], [700.03, 239.4], [663.59, 144.88], [725.93, 144.88], [804.87, 125.8], [700.03, 125.81]]}, {\"category\": \"car\", \"corners_3d\": [[318.34, 222.6], [278.37, 222.55], [314.11, 217.4], [349.97, 217.45], [318.34, 187.7], [278.37, 187.69], [314.11, 186.15], [349.97, 186.16]]}, {\"category\": \"car\", \"corners_3d\": [[445.03, 197.31], [424.15, 197.29], [437.24, 195.74], [456.81, 195.76], [445.03, 180.12], [424.15, 180.12], [437.24, 179.66], [456.81, 179.66]]}]\n```" - } - ] - }, - { - "id": 2324, - "data_source": "Mono3DRefer", - "scene_name": "002443", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[753.03, 295.9], [877.35, 296.12], [1047.24, 376.68], [841.42, 376.05], [753.03, 189.44], [877.35, 189.47], [1047.24, 200.32], [841.42, 200.24]]}, {\"category\": \"car\", \"corners_3d\": [[524.08, 217.94], [483.02, 217.92], [498.18, 212.79], [534.57, 212.81], [524.08, 184.53], [483.02, 184.52], [498.18, 183.19], [534.57, 183.2]]}, {\"category\": \"car\", \"corners_3d\": [[318.38, 310.5], [205.38, 310.08], [321.86, 271.85], [403.56, 272.06], [318.38, 204.76], [205.38, 204.67], [321.86, 195.8], [403.56, 195.85]]}, {\"category\": \"car\", \"corners_3d\": [[418.0, 263.91], [340.07, 263.52], [400.85, 245.02], [463.03, 245.26], [418.0, 193.34], [340.07, 193.25], [400.85, 189.09], [463.03, 189.15]]}, {\"category\": \"car\", \"corners_3d\": [[-1010.26, 839.39], [-1519.33, 824.69], [-80.94, 388.59], [95.62, 390.18], [-1010.26, 377.31], [-1519.33, 372.8], [-80.94, 239.03], [95.62, 239.52]]}]\n```" - } - ] - }, - { - "id": 2325, - "data_source": "Mono3DRefer", - "scene_name": "002445", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1142.57, 262.8], [1229.74, 275.29], [978.6, 280.54], [919.54, 266.82], [1142.57, 182.37], [1229.74, 183.69], [978.6, 184.24], [919.54, 182.79]]}, {\"category\": \"car\", \"corners_3d\": [[1097.42, 247.17], [1170.96, 256.44], [933.1, 260.8], [883.76, 250.59], [1097.42, 181.32], [1170.96, 182.37], [933.1, 182.87], [883.76, 181.71]]}, {\"category\": \"car\", \"corners_3d\": [[1331.52, 343.57], [1557.95, 391.79], [1087.67, 410.22], [956.9, 354.57], [1331.52, 167.38], [1557.95, 165.83], [1087.67, 165.24], [956.9, 167.02]]}]\n```" - } - ] - }, - { - "id": 2326, - "data_source": "Mono3DRefer", - "scene_name": "002447", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[613.69, 202.83], [592.41, 202.84], [593.13, 200.54], [612.78, 200.53], [613.69, 183.96], [592.41, 183.96], [593.13, 183.11], [612.78, 183.11]]}]\n```" - } - ] - }, - { - "id": 2327, - "data_source": "Mono3DRefer", - "scene_name": "002451", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[795.51, 224.64], [853.77, 225.52], [870.7, 236.36], [800.35, 235.08], [795.51, 172.17], [853.77, 172.16], [870.7, 172.01], [800.35, 172.03]]}, {\"category\": \"car\", \"corners_3d\": [[376.59, 264.42], [302.35, 261.19], [427.71, 240.29], [487.37, 242.15], [376.59, 188.37], [302.35, 187.82], [427.71, 184.28], [487.37, 184.59]]}, {\"category\": \"car\", \"corners_3d\": [[1257.85, 197.9], [1282.43, 199.27], [1172.13, 198.38], [1153.88, 197.1], [1257.85, 161.15], [1282.43, 160.51], [1172.13, 160.93], [1153.88, 161.53]]}]\n```" - } - ] - }, - { - "id": 2328, - "data_source": "Mono3DRefer", - "scene_name": "002457", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[377.19, 232.78], [349.45, 232.74], [356.44, 231.21], [383.47, 231.25], [377.19, 178.37], [349.45, 178.37], [356.44, 178.23], [383.47, 178.23]]}]\n```" - } - ] - }, - { - "id": 2329, - "data_source": "Mono3DRefer", - "scene_name": "002460", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[713.36, 218.96], [751.3, 219.13], [764.56, 225.9], [721.03, 225.68], [713.36, 175.57], [751.3, 175.58], [764.56, 175.98], [721.03, 175.97]]}]\n```" - } - ] - }, - { - "id": 2330, - "data_source": "Mono3DRefer", - "scene_name": "002460", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[-16.06, 392.88], [72.25, 378.55], [85.4, 393.97], [-9.05, 410.62], [-16.06, 161.8], [72.25, 162.52], [85.4, 161.75], [-9.05, 160.91]]}]\n```" - } - ] - }, - { - "id": 2331, - "data_source": "Mono3DRefer", - "scene_name": "002460", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[767.33, 253.49], [830.15, 253.78], [894.61, 281.07], [810.4, 280.56], [767.33, 192.2], [830.15, 192.26], [894.61, 198.81], [810.4, 198.69]]}, {\"category\": \"car\", \"corners_3d\": [[-236.82, 367.4], [-47.12, 328.52], [243.49, 338.79], [116.55, 383.71], [-236.82, 167.79], [-47.12, 168.8], [243.49, 168.53], [116.55, 167.36]]}]\n```" - } - ] - }, - { - "id": 2332, - "data_source": "Mono3DRefer", - "scene_name": "002460", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[-16.06, 392.88], [72.25, 378.55], [85.4, 393.97], [-9.05, 410.62], [-16.06, 161.8], [72.25, 162.52], [85.4, 161.75], [-9.05, 160.91]]}, {\"category\": \"van\", \"corners_3d\": [[713.36, 218.96], [751.3, 219.13], [764.56, 225.9], [721.03, 225.68], [713.36, 175.57], [751.3, 175.58], [764.56, 175.98], [721.03, 175.97]]}]\n```" - } - ] - }, - { - "id": 2333, - "data_source": "Mono3DRefer", - "scene_name": "002460", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[767.33, 253.49], [830.15, 253.78], [894.61, 281.07], [810.4, 280.56], [767.33, 192.2], [830.15, 192.26], [894.61, 198.81], [810.4, 198.69]]}, {\"category\": \"car\", \"corners_3d\": [[-236.82, 367.4], [-47.12, 328.52], [243.49, 338.79], [116.55, 383.71], [-236.82, 167.79], [-47.12, 168.8], [243.49, 168.53], [116.55, 167.36]]}, {\"category\": \"van\", \"corners_3d\": [[713.36, 218.96], [751.3, 219.13], [764.56, 225.9], [721.03, 225.68], [713.36, 175.57], [751.3, 175.58], [764.56, 175.98], [721.03, 175.97]]}]\n```" - } - ] - }, - { - "id": 2334, - "data_source": "Mono3DRefer", - "scene_name": "002460", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[767.33, 253.49], [830.15, 253.78], [894.61, 281.07], [810.4, 280.56], [767.33, 192.2], [830.15, 192.26], [894.61, 198.81], [810.4, 198.69]]}, {\"category\": \"car\", \"corners_3d\": [[-236.82, 367.4], [-47.12, 328.52], [243.49, 338.79], [116.55, 383.71], [-236.82, 167.79], [-47.12, 168.8], [243.49, 168.53], [116.55, 167.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[-16.06, 392.88], [72.25, 378.55], [85.4, 393.97], [-9.05, 410.62], [-16.06, 161.8], [72.25, 162.52], [85.4, 161.75], [-9.05, 160.91]]}]\n```" - } - ] - }, - { - "id": 2335, - "data_source": "Mono3DRefer", - "scene_name": "002460", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[767.33, 253.49], [830.15, 253.78], [894.61, 281.07], [810.4, 280.56], [767.33, 192.2], [830.15, 192.26], [894.61, 198.81], [810.4, 198.69]]}, {\"category\": \"car\", \"corners_3d\": [[-236.82, 367.4], [-47.12, 328.52], [243.49, 338.79], [116.55, 383.71], [-236.82, 167.79], [-47.12, 168.8], [243.49, 168.53], [116.55, 167.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[-16.06, 392.88], [72.25, 378.55], [85.4, 393.97], [-9.05, 410.62], [-16.06, 161.8], [72.25, 162.52], [85.4, 161.75], [-9.05, 160.91]]}, {\"category\": \"van\", \"corners_3d\": [[713.36, 218.96], [751.3, 219.13], [764.56, 225.9], [721.03, 225.68], [713.36, 175.57], [751.3, 175.58], [764.56, 175.98], [721.03, 175.97]]}]\n```" - } - ] - }, - { - "id": 2336, - "data_source": "Mono3DRefer", - "scene_name": "002489", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[693.28, 258.92], [774.72, 258.83], [834.99, 289.0], [725.03, 289.16], [693.28, 188.76], [774.72, 188.74], [834.99, 194.32], [725.03, 194.35]]}, {\"category\": \"car\", \"corners_3d\": [[522.08, 219.33], [486.71, 219.26], [501.08, 214.63], [532.94, 214.68], [522.08, 190.43], [486.71, 190.4], [501.08, 188.65], [532.94, 188.67]]}, {\"category\": \"car\", \"corners_3d\": [[400.98, 274.15], [323.59, 274.05], [384.55, 252.97], [445.85, 253.03], [400.98, 202.27], [323.59, 202.24], [384.55, 196.12], [445.85, 196.13]]}, {\"category\": \"car\", \"corners_3d\": [[451.09, 249.96], [392.02, 249.71], [431.97, 237.22], [481.52, 237.4], [451.09, 196.63], [392.02, 196.55], [431.97, 192.7], [481.52, 192.75]]}, {\"category\": \"car\", \"corners_3d\": [[105.46, 380.41], [-60.56, 379.48], [202.35, 300.05], [305.11, 300.41], [105.46, 240.23], [-60.56, 239.93], [202.35, 214.15], [305.11, 214.26]]}]\n```" - } - ] - }, - { - "id": 2337, - "data_source": "Mono3DRefer", - "scene_name": "002491", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[517.82, 240.29], [457.48, 240.24], [484.71, 228.68], [534.71, 228.71], [517.82, 187.17], [457.48, 187.16], [484.71, 184.71], [534.71, 184.71]]}, {\"category\": \"car\", \"corners_3d\": [[566.1, 205.04], [538.63, 205.02], [544.49, 202.59], [569.89, 202.6], [566.1, 181.74], [538.63, 181.74], [544.49, 181.07], [569.89, 181.07]]}]\n```" - } - ] - }, - { - "id": 2338, - "data_source": "Mono3DRefer", - "scene_name": "002494", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-80.75, 295.87], [17.27, 278.61], [257.59, 279.03], [198.52, 296.43], [-80.75, 170.18], [17.27, 170.56], [257.59, 170.55], [198.52, 170.17]]}, {\"category\": \"car\", \"corners_3d\": [[465.18, 212.3], [457.99, 214.49], [367.3, 214.59], [379.28, 212.4], [465.18, 177.23], [457.99, 177.47], [367.3, 177.48], [379.28, 177.24]]}]\n```" - } - ] - }, - { - "id": 2339, - "data_source": "Mono3DRefer", - "scene_name": "002494", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[965.35, 226.52], [985.94, 226.52], [1029.32, 232.71], [1006.35, 232.71], [965.35, 153.06], [985.94, 153.06], [1029.32, 150.78], [1006.35, 150.78]]}, {\"category\": \"cyclist\", \"corners_3d\": [[883.98, 213.46], [898.79, 213.51], [919.21, 217.0], [903.1, 216.93], [883.98, 159.08], [898.79, 159.06], [919.21, 157.88], [903.1, 157.9]]}, {\"category\": \"cyclist\", \"corners_3d\": [[854.68, 207.51], [866.9, 207.58], [879.49, 210.05], [866.37, 209.97], [854.68, 163.98], [866.9, 163.96], [879.49, 163.33], [866.37, 163.35]]}]\n```" - } - ] - }, - { - "id": 2340, - "data_source": "Mono3DRefer", - "scene_name": "002494", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-80.75, 295.87], [17.27, 278.61], [257.59, 279.03], [198.52, 296.43], [-80.75, 170.18], [17.27, 170.56], [257.59, 170.55], [198.52, 170.17]]}, {\"category\": \"car\", \"corners_3d\": [[465.18, 212.3], [457.99, 214.49], [367.3, 214.59], [379.28, 212.4], [465.18, 177.23], [457.99, 177.47], [367.3, 177.48], [379.28, 177.24]]}, {\"category\": \"cyclist\", \"corners_3d\": [[965.35, 226.52], [985.94, 226.52], [1029.32, 232.71], [1006.35, 232.71], [965.35, 153.06], [985.94, 153.06], [1029.32, 150.78], [1006.35, 150.78]]}, {\"category\": \"cyclist\", \"corners_3d\": [[883.98, 213.46], [898.79, 213.51], [919.21, 217.0], [903.1, 216.93], [883.98, 159.08], [898.79, 159.06], [919.21, 157.88], [903.1, 157.9]]}, {\"category\": \"cyclist\", \"corners_3d\": [[854.68, 207.51], [866.9, 207.58], [879.49, 210.05], [866.37, 209.97], [854.68, 163.98], [866.9, 163.96], [879.49, 163.33], [866.37, 163.35]]}]\n```" - } - ] - }, - { - "id": 2341, - "data_source": "Mono3DRefer", - "scene_name": "002501", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[734.08, 199.41], [732.4, 198.85], [768.37, 198.93], [770.84, 199.5], [734.08, 169.05], [732.4, 169.13], [768.37, 169.12], [770.84, 169.04]]}]\n```" - } - ] - }, - { - "id": 2342, - "data_source": "Mono3DRefer", - "scene_name": "002501", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[838.2, 211.02], [823.49, 211.04], [816.38, 209.85], [830.64, 209.84], [838.2, 165.05], [823.49, 165.05], [816.38, 165.29], [830.64, 165.3]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.03, 201.89], [553.61, 202.34], [542.22, 202.4], [541.81, 201.95], [553.03, 169.75], [553.61, 169.71], [542.22, 169.7], [541.81, 169.75]]}]\n```" - } - ] - }, - { - "id": 2343, - "data_source": "Mono3DRefer", - "scene_name": "002501", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[602.01, 218.36], [652.29, 218.36], [659.11, 225.73], [600.69, 225.73], [602.01, 177.9], [652.29, 177.9], [659.11, 178.72], [600.69, 178.72]]}, {\"category\": \"car\", \"corners_3d\": [[942.55, 202.84], [925.65, 201.48], [1008.4, 201.3], [1029.12, 202.65], [942.55, 172.85], [925.65, 172.85], [1008.4, 172.85], [1029.12, 172.85]]}, {\"category\": \"car\", \"corners_3d\": [[441.01, 201.18], [414.89, 201.15], [431.2, 199.01], [455.36, 199.03], [441.01, 176.83], [414.89, 176.83], [431.2, 176.53], [455.36, 176.53]]}, {\"category\": \"car\", \"corners_3d\": [[586.14, 205.79], [621.63, 205.77], [623.85, 209.69], [584.15, 209.71], [586.14, 173.48], [621.63, 173.48], [623.85, 173.56], [584.15, 173.56]]}]\n```" - } - ] - }, - { - "id": 2344, - "data_source": "Mono3DRefer", - "scene_name": "002501", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[639.22, 197.11], [665.82, 197.14], [669.07, 199.46], [639.93, 199.43], [639.22, 166.85], [665.82, 166.85], [669.07, 166.27], [639.93, 166.28]]}]\n```" - } - ] - }, - { - "id": 2345, - "data_source": "Mono3DRefer", - "scene_name": "002501", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[838.2, 211.02], [823.49, 211.04], [816.38, 209.85], [830.64, 209.84], [838.2, 165.05], [823.49, 165.05], [816.38, 165.29], [830.64, 165.3]]}, {\"category\": \"cyclist\", \"corners_3d\": [[734.08, 199.41], [732.4, 198.85], [768.37, 198.93], [770.84, 199.5], [734.08, 169.05], [732.4, 169.13], [768.37, 169.12], [770.84, 169.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.03, 201.89], [553.61, 202.34], [542.22, 202.4], [541.81, 201.95], [553.03, 169.75], [553.61, 169.71], [542.22, 169.7], [541.81, 169.75]]}]\n```" - } - ] - }, - { - "id": 2346, - "data_source": "Mono3DRefer", - "scene_name": "002501", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[602.01, 218.36], [652.29, 218.36], [659.11, 225.73], [600.69, 225.73], [602.01, 177.9], [652.29, 177.9], [659.11, 178.72], [600.69, 178.72]]}, {\"category\": \"car\", \"corners_3d\": [[942.55, 202.84], [925.65, 201.48], [1008.4, 201.3], [1029.12, 202.65], [942.55, 172.85], [925.65, 172.85], [1008.4, 172.85], [1029.12, 172.85]]}, {\"category\": \"car\", \"corners_3d\": [[441.01, 201.18], [414.89, 201.15], [431.2, 199.01], [455.36, 199.03], [441.01, 176.83], [414.89, 176.83], [431.2, 176.53], [455.36, 176.53]]}, {\"category\": \"cyclist\", \"corners_3d\": [[734.08, 199.41], [732.4, 198.85], [768.37, 198.93], [770.84, 199.5], [734.08, 169.05], [732.4, 169.13], [768.37, 169.12], [770.84, 169.04]]}, {\"category\": \"car\", \"corners_3d\": [[586.14, 205.79], [621.63, 205.77], [623.85, 209.69], [584.15, 209.71], [586.14, 173.48], [621.63, 173.48], [623.85, 173.56], [584.15, 173.56]]}]\n```" - } - ] - }, - { - "id": 2347, - "data_source": "Mono3DRefer", - "scene_name": "002501", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[734.08, 199.41], [732.4, 198.85], [768.37, 198.93], [770.84, 199.5], [734.08, 169.05], [732.4, 169.13], [768.37, 169.12], [770.84, 169.04]]}, {\"category\": \"van\", \"corners_3d\": [[639.22, 197.11], [665.82, 197.14], [669.07, 199.46], [639.93, 199.43], [639.22, 166.85], [665.82, 166.85], [669.07, 166.27], [639.93, 166.28]]}]\n```" - } - ] - }, - { - "id": 2348, - "data_source": "Mono3DRefer", - "scene_name": "002501", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[602.01, 218.36], [652.29, 218.36], [659.11, 225.73], [600.69, 225.73], [602.01, 177.9], [652.29, 177.9], [659.11, 178.72], [600.69, 178.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[838.2, 211.02], [823.49, 211.04], [816.38, 209.85], [830.64, 209.84], [838.2, 165.05], [823.49, 165.05], [816.38, 165.29], [830.64, 165.3]]}, {\"category\": \"car\", \"corners_3d\": [[942.55, 202.84], [925.65, 201.48], [1008.4, 201.3], [1029.12, 202.65], [942.55, 172.85], [925.65, 172.85], [1008.4, 172.85], [1029.12, 172.85]]}, {\"category\": \"car\", \"corners_3d\": [[441.01, 201.18], [414.89, 201.15], [431.2, 199.01], [455.36, 199.03], [441.01, 176.83], [414.89, 176.83], [431.2, 176.53], [455.36, 176.53]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.03, 201.89], [553.61, 202.34], [542.22, 202.4], [541.81, 201.95], [553.03, 169.75], [553.61, 169.71], [542.22, 169.7], [541.81, 169.75]]}, {\"category\": \"car\", \"corners_3d\": [[586.14, 205.79], [621.63, 205.77], [623.85, 209.69], [584.15, 209.71], [586.14, 173.48], [621.63, 173.48], [623.85, 173.56], [584.15, 173.56]]}]\n```" - } - ] - }, - { - "id": 2349, - "data_source": "Mono3DRefer", - "scene_name": "002501", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[838.2, 211.02], [823.49, 211.04], [816.38, 209.85], [830.64, 209.84], [838.2, 165.05], [823.49, 165.05], [816.38, 165.29], [830.64, 165.3]]}, {\"category\": \"van\", \"corners_3d\": [[639.22, 197.11], [665.82, 197.14], [669.07, 199.46], [639.93, 199.43], [639.22, 166.85], [665.82, 166.85], [669.07, 166.27], [639.93, 166.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.03, 201.89], [553.61, 202.34], [542.22, 202.4], [541.81, 201.95], [553.03, 169.75], [553.61, 169.71], [542.22, 169.7], [541.81, 169.75]]}]\n```" - } - ] - }, - { - "id": 2350, - "data_source": "Mono3DRefer", - "scene_name": "002501", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[602.01, 218.36], [652.29, 218.36], [659.11, 225.73], [600.69, 225.73], [602.01, 177.9], [652.29, 177.9], [659.11, 178.72], [600.69, 178.72]]}, {\"category\": \"car\", \"corners_3d\": [[942.55, 202.84], [925.65, 201.48], [1008.4, 201.3], [1029.12, 202.65], [942.55, 172.85], [925.65, 172.85], [1008.4, 172.85], [1029.12, 172.85]]}, {\"category\": \"car\", \"corners_3d\": [[441.01, 201.18], [414.89, 201.15], [431.2, 199.01], [455.36, 199.03], [441.01, 176.83], [414.89, 176.83], [431.2, 176.53], [455.36, 176.53]]}, {\"category\": \"van\", \"corners_3d\": [[639.22, 197.11], [665.82, 197.14], [669.07, 199.46], [639.93, 199.43], [639.22, 166.85], [665.82, 166.85], [669.07, 166.27], [639.93, 166.28]]}, {\"category\": \"car\", \"corners_3d\": [[586.14, 205.79], [621.63, 205.77], [623.85, 209.69], [584.15, 209.71], [586.14, 173.48], [621.63, 173.48], [623.85, 173.56], [584.15, 173.56]]}]\n```" - } - ] - }, - { - "id": 2351, - "data_source": "Mono3DRefer", - "scene_name": "002501", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[602.01, 218.36], [652.29, 218.36], [659.11, 225.73], [600.69, 225.73], [602.01, 177.9], [652.29, 177.9], [659.11, 178.72], [600.69, 178.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[838.2, 211.02], [823.49, 211.04], [816.38, 209.85], [830.64, 209.84], [838.2, 165.05], [823.49, 165.05], [816.38, 165.29], [830.64, 165.3]]}, {\"category\": \"car\", \"corners_3d\": [[942.55, 202.84], [925.65, 201.48], [1008.4, 201.3], [1029.12, 202.65], [942.55, 172.85], [925.65, 172.85], [1008.4, 172.85], [1029.12, 172.85]]}, {\"category\": \"car\", \"corners_3d\": [[441.01, 201.18], [414.89, 201.15], [431.2, 199.01], [455.36, 199.03], [441.01, 176.83], [414.89, 176.83], [431.2, 176.53], [455.36, 176.53]]}, {\"category\": \"cyclist\", \"corners_3d\": [[734.08, 199.41], [732.4, 198.85], [768.37, 198.93], [770.84, 199.5], [734.08, 169.05], [732.4, 169.13], [768.37, 169.12], [770.84, 169.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.03, 201.89], [553.61, 202.34], [542.22, 202.4], [541.81, 201.95], [553.03, 169.75], [553.61, 169.71], [542.22, 169.7], [541.81, 169.75]]}, {\"category\": \"car\", \"corners_3d\": [[586.14, 205.79], [621.63, 205.77], [623.85, 209.69], [584.15, 209.71], [586.14, 173.48], [621.63, 173.48], [623.85, 173.56], [584.15, 173.56]]}]\n```" - } - ] - }, - { - "id": 2352, - "data_source": "Mono3DRefer", - "scene_name": "002501", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[838.2, 211.02], [823.49, 211.04], [816.38, 209.85], [830.64, 209.84], [838.2, 165.05], [823.49, 165.05], [816.38, 165.29], [830.64, 165.3]]}, {\"category\": \"cyclist\", \"corners_3d\": [[734.08, 199.41], [732.4, 198.85], [768.37, 198.93], [770.84, 199.5], [734.08, 169.05], [732.4, 169.13], [768.37, 169.12], [770.84, 169.04]]}, {\"category\": \"van\", \"corners_3d\": [[639.22, 197.11], [665.82, 197.14], [669.07, 199.46], [639.93, 199.43], [639.22, 166.85], [665.82, 166.85], [669.07, 166.27], [639.93, 166.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.03, 201.89], [553.61, 202.34], [542.22, 202.4], [541.81, 201.95], [553.03, 169.75], [553.61, 169.71], [542.22, 169.7], [541.81, 169.75]]}]\n```" - } - ] - }, - { - "id": 2353, - "data_source": "Mono3DRefer", - "scene_name": "002501", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[602.01, 218.36], [652.29, 218.36], [659.11, 225.73], [600.69, 225.73], [602.01, 177.9], [652.29, 177.9], [659.11, 178.72], [600.69, 178.72]]}, {\"category\": \"car\", \"corners_3d\": [[942.55, 202.84], [925.65, 201.48], [1008.4, 201.3], [1029.12, 202.65], [942.55, 172.85], [925.65, 172.85], [1008.4, 172.85], [1029.12, 172.85]]}, {\"category\": \"car\", \"corners_3d\": [[441.01, 201.18], [414.89, 201.15], [431.2, 199.01], [455.36, 199.03], [441.01, 176.83], [414.89, 176.83], [431.2, 176.53], [455.36, 176.53]]}, {\"category\": \"cyclist\", \"corners_3d\": [[734.08, 199.41], [732.4, 198.85], [768.37, 198.93], [770.84, 199.5], [734.08, 169.05], [732.4, 169.13], [768.37, 169.12], [770.84, 169.04]]}, {\"category\": \"van\", \"corners_3d\": [[639.22, 197.11], [665.82, 197.14], [669.07, 199.46], [639.93, 199.43], [639.22, 166.85], [665.82, 166.85], [669.07, 166.27], [639.93, 166.28]]}, {\"category\": \"car\", \"corners_3d\": [[586.14, 205.79], [621.63, 205.77], [623.85, 209.69], [584.15, 209.71], [586.14, 173.48], [621.63, 173.48], [623.85, 173.56], [584.15, 173.56]]}]\n```" - } - ] - }, - { - "id": 2354, - "data_source": "Mono3DRefer", - "scene_name": "002501", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[602.01, 218.36], [652.29, 218.36], [659.11, 225.73], [600.69, 225.73], [602.01, 177.9], [652.29, 177.9], [659.11, 178.72], [600.69, 178.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[838.2, 211.02], [823.49, 211.04], [816.38, 209.85], [830.64, 209.84], [838.2, 165.05], [823.49, 165.05], [816.38, 165.29], [830.64, 165.3]]}, {\"category\": \"car\", \"corners_3d\": [[942.55, 202.84], [925.65, 201.48], [1008.4, 201.3], [1029.12, 202.65], [942.55, 172.85], [925.65, 172.85], [1008.4, 172.85], [1029.12, 172.85]]}, {\"category\": \"car\", \"corners_3d\": [[441.01, 201.18], [414.89, 201.15], [431.2, 199.01], [455.36, 199.03], [441.01, 176.83], [414.89, 176.83], [431.2, 176.53], [455.36, 176.53]]}, {\"category\": \"van\", \"corners_3d\": [[639.22, 197.11], [665.82, 197.14], [669.07, 199.46], [639.93, 199.43], [639.22, 166.85], [665.82, 166.85], [669.07, 166.27], [639.93, 166.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.03, 201.89], [553.61, 202.34], [542.22, 202.4], [541.81, 201.95], [553.03, 169.75], [553.61, 169.71], [542.22, 169.7], [541.81, 169.75]]}, {\"category\": \"car\", \"corners_3d\": [[586.14, 205.79], [621.63, 205.77], [623.85, 209.69], [584.15, 209.71], [586.14, 173.48], [621.63, 173.48], [623.85, 173.56], [584.15, 173.56]]}]\n```" - } - ] - }, - { - "id": 2355, - "data_source": "Mono3DRefer", - "scene_name": "002501", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[602.01, 218.36], [652.29, 218.36], [659.11, 225.73], [600.69, 225.73], [602.01, 177.9], [652.29, 177.9], [659.11, 178.72], [600.69, 178.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[838.2, 211.02], [823.49, 211.04], [816.38, 209.85], [830.64, 209.84], [838.2, 165.05], [823.49, 165.05], [816.38, 165.29], [830.64, 165.3]]}, {\"category\": \"car\", \"corners_3d\": [[942.55, 202.84], [925.65, 201.48], [1008.4, 201.3], [1029.12, 202.65], [942.55, 172.85], [925.65, 172.85], [1008.4, 172.85], [1029.12, 172.85]]}, {\"category\": \"car\", \"corners_3d\": [[441.01, 201.18], [414.89, 201.15], [431.2, 199.01], [455.36, 199.03], [441.01, 176.83], [414.89, 176.83], [431.2, 176.53], [455.36, 176.53]]}, {\"category\": \"cyclist\", \"corners_3d\": [[734.08, 199.41], [732.4, 198.85], [768.37, 198.93], [770.84, 199.5], [734.08, 169.05], [732.4, 169.13], [768.37, 169.12], [770.84, 169.04]]}, {\"category\": \"van\", \"corners_3d\": [[639.22, 197.11], [665.82, 197.14], [669.07, 199.46], [639.93, 199.43], [639.22, 166.85], [665.82, 166.85], [669.07, 166.27], [639.93, 166.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.03, 201.89], [553.61, 202.34], [542.22, 202.4], [541.81, 201.95], [553.03, 169.75], [553.61, 169.71], [542.22, 169.7], [541.81, 169.75]]}, {\"category\": \"car\", \"corners_3d\": [[586.14, 205.79], [621.63, 205.77], [623.85, 209.69], [584.15, 209.71], [586.14, 173.48], [621.63, 173.48], [623.85, 173.56], [584.15, 173.56]]}]\n```" - } - ] - }, - { - "id": 2356, - "data_source": "Mono3DRefer", - "scene_name": "002507", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[327.92, 300.99], [304.6, 309.44], [232.97, 308.35], [260.77, 300.05], [327.92, 163.25], [304.6, 162.04], [232.97, 162.19], [260.77, 163.38]]}]\n```" - } - ] - }, - { - "id": 2357, - "data_source": "Mono3DRefer", - "scene_name": "002511", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[446.48, 273.82], [351.41, 273.96], [412.2, 249.41], [484.15, 249.33], [446.48, 170.55], [351.41, 170.55], [412.2, 171.11], [484.15, 171.11]]}]\n```" - } - ] - }, - { - "id": 2358, - "data_source": "Mono3DRefer", - "scene_name": "002511", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[310.87, 359.1], [152.1, 360.75], [303.93, 294.1], [405.83, 293.41], [310.87, 213.63], [152.1, 213.99], [303.93, 199.4], [405.83, 199.25]]}, {\"category\": \"car\", \"corners_3d\": [[548.98, 217.37], [506.63, 217.37], [519.35, 211.84], [556.44, 211.84], [548.98, 182.04], [506.63, 182.04], [519.35, 180.9], [556.44, 180.9]]}]\n```" - } - ] - }, - { - "id": 2359, - "data_source": "Mono3DRefer", - "scene_name": "002511", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[310.87, 359.1], [152.1, 360.75], [303.93, 294.1], [405.83, 293.41], [310.87, 213.63], [152.1, 213.99], [303.93, 199.4], [405.83, 199.25]]}, {\"category\": \"van\", \"corners_3d\": [[446.48, 273.82], [351.41, 273.96], [412.2, 249.41], [484.15, 249.33], [446.48, 170.55], [351.41, 170.55], [412.2, 171.11], [484.15, 171.11]]}, {\"category\": \"car\", \"corners_3d\": [[548.98, 217.37], [506.63, 217.37], [519.35, 211.84], [556.44, 211.84], [548.98, 182.04], [506.63, 182.04], [519.35, 180.9], [556.44, 180.9]]}]\n```" - } - ] - }, - { - "id": 2360, - "data_source": "Mono3DRefer", - "scene_name": "002512", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[465.56, 242.28], [380.79, 242.21], [447.45, 222.57], [508.25, 222.61], [465.56, 137.91], [380.79, 137.94], [447.45, 147.83], [508.25, 147.81]]}]\n```" - } - ] - }, - { - "id": 2361, - "data_source": "Mono3DRefer", - "scene_name": "002512", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[530.41, 215.29], [483.63, 214.84], [519.18, 208.67], [559.35, 209.0], [530.41, 177.77], [483.63, 177.72], [519.18, 177.01], [559.35, 177.04]]}]\n```" - } - ] - }, - { - "id": 2362, - "data_source": "Mono3DRefer", - "scene_name": "002512", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[648.62, 202.82], [649.83, 202.4], [682.79, 202.63], [682.04, 203.06], [648.62, 168.91], [649.83, 168.97], [682.79, 168.94], [682.04, 168.88]]}]\n```" - } - ] - }, - { - "id": 2363, - "data_source": "Mono3DRefer", - "scene_name": "002512", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[465.56, 242.28], [380.79, 242.21], [447.45, 222.57], [508.25, 222.61], [465.56, 137.91], [380.79, 137.94], [447.45, 147.83], [508.25, 147.81]]}, {\"category\": \"car\", \"corners_3d\": [[530.41, 215.29], [483.63, 214.84], [519.18, 208.67], [559.35, 209.0], [530.41, 177.77], [483.63, 177.72], [519.18, 177.01], [559.35, 177.04]]}]\n```" - } - ] - }, - { - "id": 2364, - "data_source": "Mono3DRefer", - "scene_name": "002512", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[465.56, 242.28], [380.79, 242.21], [447.45, 222.57], [508.25, 222.61], [465.56, 137.91], [380.79, 137.94], [447.45, 147.83], [508.25, 147.81]]}, {\"category\": \"cyclist\", \"corners_3d\": [[648.62, 202.82], [649.83, 202.4], [682.79, 202.63], [682.04, 203.06], [648.62, 168.91], [649.83, 168.97], [682.79, 168.94], [682.04, 168.88]]}]\n```" - } - ] - }, - { - "id": 2365, - "data_source": "Mono3DRefer", - "scene_name": "002512", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[530.41, 215.29], [483.63, 214.84], [519.18, 208.67], [559.35, 209.0], [530.41, 177.77], [483.63, 177.72], [519.18, 177.01], [559.35, 177.04]]}, {\"category\": \"cyclist\", \"corners_3d\": [[648.62, 202.82], [649.83, 202.4], [682.79, 202.63], [682.04, 203.06], [648.62, 168.91], [649.83, 168.97], [682.79, 168.94], [682.04, 168.88]]}]\n```" - } - ] - }, - { - "id": 2366, - "data_source": "Mono3DRefer", - "scene_name": "002512", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[465.56, 242.28], [380.79, 242.21], [447.45, 222.57], [508.25, 222.61], [465.56, 137.91], [380.79, 137.94], [447.45, 147.83], [508.25, 147.81]]}, {\"category\": \"car\", \"corners_3d\": [[530.41, 215.29], [483.63, 214.84], [519.18, 208.67], [559.35, 209.0], [530.41, 177.77], [483.63, 177.72], [519.18, 177.01], [559.35, 177.04]]}, {\"category\": \"cyclist\", \"corners_3d\": [[648.62, 202.82], [649.83, 202.4], [682.79, 202.63], [682.04, 203.06], [648.62, 168.91], [649.83, 168.97], [682.79, 168.94], [682.04, 168.88]]}]\n```" - } - ] - }, - { - "id": 2367, - "data_source": "Mono3DRefer", - "scene_name": "002514", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[33.78, 299.93], [101.44, 283.52], [308.72, 281.35], [272.62, 297.08], [33.78, 205.93], [101.44, 201.66], [308.72, 201.09], [272.62, 205.19]]}, {\"category\": \"car\", \"corners_3d\": [[330.78, 232.58], [346.41, 228.83], [453.35, 228.49], [444.95, 232.19], [330.78, 188.25], [346.41, 187.28], [453.35, 187.19], [444.95, 188.15]]}]\n```" - } - ] - }, - { - "id": 2368, - "data_source": "Mono3DRefer", - "scene_name": "002514", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[602.46, 218.75], [619.26, 218.72], [621.88, 222.62], [603.65, 222.66], [602.46, 173.11], [619.26, 173.11], [621.88, 173.13], [603.65, 173.13]]}]\n```" - } - ] - }, - { - "id": 2369, - "data_source": "Mono3DRefer", - "scene_name": "002514", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[602.46, 218.75], [619.26, 218.72], [621.88, 222.62], [603.65, 222.66], [602.46, 173.11], [619.26, 173.11], [621.88, 173.13], [603.65, 173.13]]}, {\"category\": \"car\", \"corners_3d\": [[33.78, 299.93], [101.44, 283.52], [308.72, 281.35], [272.62, 297.08], [33.78, 205.93], [101.44, 201.66], [308.72, 201.09], [272.62, 205.19]]}, {\"category\": \"car\", \"corners_3d\": [[330.78, 232.58], [346.41, 228.83], [453.35, 228.49], [444.95, 232.19], [330.78, 188.25], [346.41, 187.28], [453.35, 187.19], [444.95, 188.15]]}]\n```" - } - ] - }, - { - "id": 2370, - "data_source": "Mono3DRefer", - "scene_name": "002521", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[189.18, 226.96], [147.78, 226.96], [204.17, 220.35], [240.5, 220.35], [189.18, 188.99], [147.78, 188.99], [204.17, 187.02], [240.5, 187.02]]}, {\"category\": \"car\", \"corners_3d\": [[352.89, 209.76], [321.62, 209.73], [344.23, 206.97], [373.17, 206.99], [352.89, 182.29], [321.62, 182.28], [344.23, 181.57], [373.17, 181.58]]}]\n```" - } - ] - }, - { - "id": 2371, - "data_source": "Mono3DRefer", - "scene_name": "002526", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[167.28, 570.73], [-111.46, 601.81], [95.89, 441.33], [263.45, 429.11], [167.28, 153.67], [-111.46, 151.45], [95.89, 162.93], [263.45, 163.8]]}, {\"category\": \"cyclist\", \"corners_3d\": [[547.93, 235.41], [520.66, 235.48], [525.92, 230.55], [550.7, 230.5], [547.93, 165.8], [520.66, 165.78], [525.92, 167.21], [550.7, 167.23]]}]\n```" - } - ] - }, - { - "id": 2372, - "data_source": "Mono3DRefer", - "scene_name": "002526", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[834.92, 269.62], [893.71, 269.91], [927.09, 281.09], [860.78, 280.72], [834.92, 146.12], [893.71, 146.0], [927.09, 141.51], [860.78, 141.66]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[575.36, 251.0], [617.55, 253.33], [593.27, 257.85], [549.35, 255.22], [575.36, 161.13], [617.55, 160.44], [593.27, 159.12], [549.35, 159.89]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[351.75, 259.63], [354.63, 255.65], [403.39, 254.41], [403.14, 258.27], [351.75, 156.35], [354.63, 157.63], [403.39, 158.03], [403.14, 156.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[521.49, 250.38], [481.4, 250.15], [490.6, 246.24], [528.43, 246.45], [521.49, 158.12], [481.4, 158.19], [490.6, 159.52], [528.43, 159.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[784.59, 246.85], [741.89, 246.94], [730.45, 242.15], [770.04, 242.08], [784.59, 143.5], [741.89, 143.45], [730.45, 146.23], [770.04, 146.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[543.5, 210.4], [562.9, 210.35], [563.29, 211.27], [543.28, 211.32], [543.5, 162.36], [562.9, 162.4], [563.29, 161.79], [543.28, 161.75]]}]\n```" - } - ] - }, - { - "id": 2373, - "data_source": "Mono3DRefer", - "scene_name": "002526", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[167.28, 570.73], [-111.46, 601.81], [95.89, 441.33], [263.45, 429.11], [167.28, 153.67], [-111.46, 151.45], [95.89, 162.93], [263.45, 163.8]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[834.92, 269.62], [893.71, 269.91], [927.09, 281.09], [860.78, 280.72], [834.92, 146.12], [893.71, 146.0], [927.09, 141.51], [860.78, 141.66]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[575.36, 251.0], [617.55, 253.33], [593.27, 257.85], [549.35, 255.22], [575.36, 161.13], [617.55, 160.44], [593.27, 159.12], [549.35, 159.89]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[351.75, 259.63], [354.63, 255.65], [403.39, 254.41], [403.14, 258.27], [351.75, 156.35], [354.63, 157.63], [403.39, 158.03], [403.14, 156.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[521.49, 250.38], [481.4, 250.15], [490.6, 246.24], [528.43, 246.45], [521.49, 158.12], [481.4, 158.19], [490.6, 159.52], [528.43, 159.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[784.59, 246.85], [741.89, 246.94], [730.45, 242.15], [770.04, 242.08], [784.59, 143.5], [741.89, 143.45], [730.45, 146.23], [770.04, 146.27]]}, {\"category\": \"cyclist\", \"corners_3d\": [[547.93, 235.41], [520.66, 235.48], [525.92, 230.55], [550.7, 230.5], [547.93, 165.8], [520.66, 165.78], [525.92, 167.21], [550.7, 167.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[543.5, 210.4], [562.9, 210.35], [563.29, 211.27], [543.28, 211.32], [543.5, 162.36], [562.9, 162.4], [563.29, 161.79], [543.28, 161.75]]}]\n```" - } - ] - }, - { - "id": 2374, - "data_source": "Mono3DRefer", - "scene_name": "002529", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[528.48, 331.73], [465.89, 331.74], [485.15, 310.57], [538.98, 310.56], [528.48, 162.16], [465.89, 162.16], [485.15, 164.73], [538.98, 164.73]]}]\n```" - } - ] - }, - { - "id": 2375, - "data_source": "Mono3DRefer", - "scene_name": "002542", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[698.77, 193.4], [722.92, 193.56], [718.48, 195.2], [692.47, 195.01], [698.77, 170.79], [722.92, 170.78], [718.48, 170.61], [692.47, 170.63]]}, {\"category\": \"car\", \"corners_3d\": [[90.25, 330.05], [-41.31, 327.31], [194.44, 276.41], [284.92, 277.64], [90.25, 155.65], [-41.31, 155.95], [194.44, 161.52], [284.92, 161.39]]}, {\"category\": \"car\", \"corners_3d\": [[735.82, 245.31], [811.41, 245.47], [870.42, 268.4], [770.85, 268.13], [735.82, 174.19], [811.41, 174.19], [870.42, 174.61], [770.85, 174.61]]}, {\"category\": \"car\", \"corners_3d\": [[637.04, 201.45], [604.25, 201.17], [619.62, 198.55], [649.48, 198.78], [637.04, 170.63], [604.25, 170.65], [619.62, 170.86], [649.48, 170.84]]}, {\"category\": \"car\", \"corners_3d\": [[732.02, 228.06], [785.05, 228.42], [809.93, 241.01], [744.79, 240.47], [732.02, 156.87], [785.05, 156.76], [809.93, 153.11], [744.79, 153.27]]}]\n```" - } - ] - }, - { - "id": 2376, - "data_source": "Mono3DRefer", - "scene_name": "002548", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[471.4, 208.32], [483.19, 210.44], [328.63, 212.71], [328.1, 210.26], [471.4, 134.61], [483.19, 130.91], [328.63, 126.94], [328.1, 131.22]]}, {\"category\": \"tram\", \"corners_3d\": [[533.35, 203.04], [549.41, 204.67], [484.42, 205.46], [472.58, 203.72], [533.35, 132.16], [549.41, 128.42], [484.42, 126.61], [472.58, 130.6]]}, {\"category\": \"tram\", \"corners_3d\": [[746.92, 194.79], [775.3, 195.68], [551.06, 197.57], [534.93, 196.44], [746.92, 131.2], [775.3, 127.84], [551.06, 120.67], [534.93, 124.95]]}]\n```" - } - ] - }, - { - "id": 2377, - "data_source": "Mono3DRefer", - "scene_name": "002548", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[565.28, 224.67], [553.22, 223.22], [582.19, 222.33], [594.96, 223.72], [565.28, 171.97], [553.22, 172.29], [582.19, 172.49], [594.96, 172.18]]}]\n```" - } - ] - }, - { - "id": 2378, - "data_source": "Mono3DRefer", - "scene_name": "002548", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[849.39, 307.65], [872.16, 302.0], [929.08, 309.19], [907.73, 315.55], [849.39, 202.0], [872.16, 201.08], [929.08, 202.25], [907.73, 203.28]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[842.7, 299.16], [798.96, 301.45], [768.81, 292.53], [809.88, 290.57], [842.7, 208.87], [798.96, 209.41], [768.81, 207.33], [809.88, 206.88]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[882.28, 298.34], [828.88, 295.64], [830.43, 288.55], [880.47, 290.91], [882.28, 198.9], [828.88, 198.49], [830.43, 197.44], [880.47, 197.79]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1223.77, 486.77], [1259.64, 461.21], [1522.58, 525.17], [1507.72, 564.58], [1223.77, 241.49], [1259.64, 236.46], [1522.58, 249.03], [1507.72, 256.77]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1215.24, 450.07], [1079.61, 430.91], [1068.96, 396.94], [1184.21, 411.08], [1215.24, 227.23], [1079.61, 223.97], [1068.96, 218.19], [1184.21, 220.59]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[801.53, 306.89], [858.06, 306.9], [881.43, 318.29], [819.75, 318.28], [801.53, 203.83], [858.06, 203.83], [881.43, 205.86], [819.75, 205.86]]}]\n```" - } - ] - }, - { - "id": 2379, - "data_source": "Mono3DRefer", - "scene_name": "002548", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[565.28, 224.67], [553.22, 223.22], [582.19, 222.33], [594.96, 223.72], [565.28, 171.97], [553.22, 172.29], [582.19, 172.49], [594.96, 172.18]]}, {\"category\": \"tram\", \"corners_3d\": [[471.4, 208.32], [483.19, 210.44], [328.63, 212.71], [328.1, 210.26], [471.4, 134.61], [483.19, 130.91], [328.63, 126.94], [328.1, 131.22]]}, {\"category\": \"tram\", \"corners_3d\": [[533.35, 203.04], [549.41, 204.67], [484.42, 205.46], [472.58, 203.72], [533.35, 132.16], [549.41, 128.42], [484.42, 126.61], [472.58, 130.6]]}, {\"category\": \"tram\", \"corners_3d\": [[746.92, 194.79], [775.3, 195.68], [551.06, 197.57], [534.93, 196.44], [746.92, 131.2], [775.3, 127.84], [551.06, 120.67], [534.93, 124.95]]}]\n```" - } - ] - }, - { - "id": 2380, - "data_source": "Mono3DRefer", - "scene_name": "002548", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[849.39, 307.65], [872.16, 302.0], [929.08, 309.19], [907.73, 315.55], [849.39, 202.0], [872.16, 201.08], [929.08, 202.25], [907.73, 203.28]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[842.7, 299.16], [798.96, 301.45], [768.81, 292.53], [809.88, 290.57], [842.7, 208.87], [798.96, 209.41], [768.81, 207.33], [809.88, 206.88]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[882.28, 298.34], [828.88, 295.64], [830.43, 288.55], [880.47, 290.91], [882.28, 198.9], [828.88, 198.49], [830.43, 197.44], [880.47, 197.79]]}, {\"category\": \"tram\", \"corners_3d\": [[471.4, 208.32], [483.19, 210.44], [328.63, 212.71], [328.1, 210.26], [471.4, 134.61], [483.19, 130.91], [328.63, 126.94], [328.1, 131.22]]}, {\"category\": \"tram\", \"corners_3d\": [[533.35, 203.04], [549.41, 204.67], [484.42, 205.46], [472.58, 203.72], [533.35, 132.16], [549.41, 128.42], [484.42, 126.61], [472.58, 130.6]]}, {\"category\": \"tram\", \"corners_3d\": [[746.92, 194.79], [775.3, 195.68], [551.06, 197.57], [534.93, 196.44], [746.92, 131.2], [775.3, 127.84], [551.06, 120.67], [534.93, 124.95]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1223.77, 486.77], [1259.64, 461.21], [1522.58, 525.17], [1507.72, 564.58], [1223.77, 241.49], [1259.64, 236.46], [1522.58, 249.03], [1507.72, 256.77]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1215.24, 450.07], [1079.61, 430.91], [1068.96, 396.94], [1184.21, 411.08], [1215.24, 227.23], [1079.61, 223.97], [1068.96, 218.19], [1184.21, 220.59]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[801.53, 306.89], [858.06, 306.9], [881.43, 318.29], [819.75, 318.28], [801.53, 203.83], [858.06, 203.83], [881.43, 205.86], [819.75, 205.86]]}]\n```" - } - ] - }, - { - "id": 2381, - "data_source": "Mono3DRefer", - "scene_name": "002548", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[849.39, 307.65], [872.16, 302.0], [929.08, 309.19], [907.73, 315.55], [849.39, 202.0], [872.16, 201.08], [929.08, 202.25], [907.73, 203.28]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[842.7, 299.16], [798.96, 301.45], [768.81, 292.53], [809.88, 290.57], [842.7, 208.87], [798.96, 209.41], [768.81, 207.33], [809.88, 206.88]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[882.28, 298.34], [828.88, 295.64], [830.43, 288.55], [880.47, 290.91], [882.28, 198.9], [828.88, 198.49], [830.43, 197.44], [880.47, 197.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[565.28, 224.67], [553.22, 223.22], [582.19, 222.33], [594.96, 223.72], [565.28, 171.97], [553.22, 172.29], [582.19, 172.49], [594.96, 172.18]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1223.77, 486.77], [1259.64, 461.21], [1522.58, 525.17], [1507.72, 564.58], [1223.77, 241.49], [1259.64, 236.46], [1522.58, 249.03], [1507.72, 256.77]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1215.24, 450.07], [1079.61, 430.91], [1068.96, 396.94], [1184.21, 411.08], [1215.24, 227.23], [1079.61, 223.97], [1068.96, 218.19], [1184.21, 220.59]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[801.53, 306.89], [858.06, 306.9], [881.43, 318.29], [819.75, 318.28], [801.53, 203.83], [858.06, 203.83], [881.43, 205.86], [819.75, 205.86]]}]\n```" - } - ] - }, - { - "id": 2382, - "data_source": "Mono3DRefer", - "scene_name": "002548", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, pedestrian, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[849.39, 307.65], [872.16, 302.0], [929.08, 309.19], [907.73, 315.55], [849.39, 202.0], [872.16, 201.08], [929.08, 202.25], [907.73, 203.28]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[842.7, 299.16], [798.96, 301.45], [768.81, 292.53], [809.88, 290.57], [842.7, 208.87], [798.96, 209.41], [768.81, 207.33], [809.88, 206.88]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[882.28, 298.34], [828.88, 295.64], [830.43, 288.55], [880.47, 290.91], [882.28, 198.9], [828.88, 198.49], [830.43, 197.44], [880.47, 197.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[565.28, 224.67], [553.22, 223.22], [582.19, 222.33], [594.96, 223.72], [565.28, 171.97], [553.22, 172.29], [582.19, 172.49], [594.96, 172.18]]}, {\"category\": \"tram\", \"corners_3d\": [[471.4, 208.32], [483.19, 210.44], [328.63, 212.71], [328.1, 210.26], [471.4, 134.61], [483.19, 130.91], [328.63, 126.94], [328.1, 131.22]]}, {\"category\": \"tram\", \"corners_3d\": [[533.35, 203.04], [549.41, 204.67], [484.42, 205.46], [472.58, 203.72], [533.35, 132.16], [549.41, 128.42], [484.42, 126.61], [472.58, 130.6]]}, {\"category\": \"tram\", \"corners_3d\": [[746.92, 194.79], [775.3, 195.68], [551.06, 197.57], [534.93, 196.44], [746.92, 131.2], [775.3, 127.84], [551.06, 120.67], [534.93, 124.95]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1223.77, 486.77], [1259.64, 461.21], [1522.58, 525.17], [1507.72, 564.58], [1223.77, 241.49], [1259.64, 236.46], [1522.58, 249.03], [1507.72, 256.77]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1215.24, 450.07], [1079.61, 430.91], [1068.96, 396.94], [1184.21, 411.08], [1215.24, 227.23], [1079.61, 223.97], [1068.96, 218.19], [1184.21, 220.59]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[801.53, 306.89], [858.06, 306.9], [881.43, 318.29], [819.75, 318.28], [801.53, 203.83], [858.06, 203.83], [881.43, 205.86], [819.75, 205.86]]}]\n```" - } - ] - }, - { - "id": 2383, - "data_source": "Mono3DRefer", - "scene_name": "002557", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[759.13, 208.08], [774.12, 208.06], [778.03, 208.8], [762.72, 208.82], [759.13, 170.88], [774.12, 170.88], [778.03, 170.84], [762.72, 170.84]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[776.86, 207.06], [787.29, 207.06], [792.27, 207.98], [781.55, 207.99], [776.86, 170.01], [787.29, 170.01], [792.27, 169.94], [781.55, 169.94]]}]\n```" - } - ] - }, - { - "id": 2384, - "data_source": "Mono3DRefer", - "scene_name": "002557", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[832.11, 254.75], [876.28, 255.19], [911.55, 269.63], [859.46, 269.02], [832.11, 148.87], [876.28, 148.74], [911.55, 144.51], [859.46, 144.69]]}, {\"category\": \"cyclist\", \"corners_3d\": [[740.9, 213.91], [754.79, 213.99], [759.71, 216.79], [744.86, 216.7], [740.9, 166.53], [754.79, 166.51], [759.71, 166.08], [744.86, 166.1]]}, {\"category\": \"cyclist\", \"corners_3d\": [[711.39, 200.95], [718.8, 200.97], [721.25, 202.02], [713.56, 202.0], [711.39, 173.52], [718.8, 173.52], [721.25, 173.54], [713.56, 173.54]]}]\n```" - } - ] - }, - { - "id": 2385, - "data_source": "Mono3DRefer", - "scene_name": "002557", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[832.11, 254.75], [876.28, 255.19], [911.55, 269.63], [859.46, 269.02], [832.11, 148.87], [876.28, 148.74], [911.55, 144.51], [859.46, 144.69]]}, {\"category\": \"cyclist\", \"corners_3d\": [[740.9, 213.91], [754.79, 213.99], [759.71, 216.79], [744.86, 216.7], [740.9, 166.53], [754.79, 166.51], [759.71, 166.08], [744.86, 166.1]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[759.13, 208.08], [774.12, 208.06], [778.03, 208.8], [762.72, 208.82], [759.13, 170.88], [774.12, 170.88], [778.03, 170.84], [762.72, 170.84]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[776.86, 207.06], [787.29, 207.06], [792.27, 207.98], [781.55, 207.99], [776.86, 170.01], [787.29, 170.01], [792.27, 169.94], [781.55, 169.94]]}, {\"category\": \"cyclist\", \"corners_3d\": [[711.39, 200.95], [718.8, 200.97], [721.25, 202.02], [713.56, 202.0], [711.39, 173.52], [718.8, 173.52], [721.25, 173.54], [713.56, 173.54]]}]\n```" - } - ] - }, - { - "id": 2386, - "data_source": "Mono3DRefer", - "scene_name": "002560", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[499.26, 209.17], [507.84, 208.43], [528.21, 208.88], [519.96, 209.65], [499.26, 164.37], [507.84, 164.54], [528.21, 164.44], [519.96, 164.26]]}]\n```" - } - ] - }, - { - "id": 2387, - "data_source": "Mono3DRefer", - "scene_name": "002560", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-311.63, 370.37], [-139.78, 332.89], [184.36, 331.37], [89.55, 368.07], [-311.63, 215.62], [-139.78, 207.5], [184.36, 207.18], [89.55, 215.12]]}]\n```" - } - ] - }, - { - "id": 2388, - "data_source": "Mono3DRefer", - "scene_name": "002560", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[684.72, 221.52], [705.25, 221.96], [690.56, 226.97], [668.1, 226.43], [684.72, 165.28], [705.25, 165.21], [690.56, 164.43], [668.1, 164.51]]}]\n```" - } - ] - }, - { - "id": 2389, - "data_source": "Mono3DRefer", - "scene_name": "002560", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-311.63, 370.37], [-139.78, 332.89], [184.36, 331.37], [89.55, 368.07], [-311.63, 215.62], [-139.78, 207.5], [184.36, 207.18], [89.55, 215.12]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[499.26, 209.17], [507.84, 208.43], [528.21, 208.88], [519.96, 209.65], [499.26, 164.37], [507.84, 164.54], [528.21, 164.44], [519.96, 164.26]]}]\n```" - } - ] - }, - { - "id": 2390, - "data_source": "Mono3DRefer", - "scene_name": "002560", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[499.26, 209.17], [507.84, 208.43], [528.21, 208.88], [519.96, 209.65], [499.26, 164.37], [507.84, 164.54], [528.21, 164.44], [519.96, 164.26]]}, {\"category\": \"cyclist\", \"corners_3d\": [[684.72, 221.52], [705.25, 221.96], [690.56, 226.97], [668.1, 226.43], [684.72, 165.28], [705.25, 165.21], [690.56, 164.43], [668.1, 164.51]]}]\n```" - } - ] - }, - { - "id": 2391, - "data_source": "Mono3DRefer", - "scene_name": "002560", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-311.63, 370.37], [-139.78, 332.89], [184.36, 331.37], [89.55, 368.07], [-311.63, 215.62], [-139.78, 207.5], [184.36, 207.18], [89.55, 215.12]]}, {\"category\": \"cyclist\", \"corners_3d\": [[684.72, 221.52], [705.25, 221.96], [690.56, 226.97], [668.1, 226.43], [684.72, 165.28], [705.25, 165.21], [690.56, 164.43], [668.1, 164.51]]}]\n```" - } - ] - }, - { - "id": 2392, - "data_source": "Mono3DRefer", - "scene_name": "002560", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-311.63, 370.37], [-139.78, 332.89], [184.36, 331.37], [89.55, 368.07], [-311.63, 215.62], [-139.78, 207.5], [184.36, 207.18], [89.55, 215.12]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[499.26, 209.17], [507.84, 208.43], [528.21, 208.88], [519.96, 209.65], [499.26, 164.37], [507.84, 164.54], [528.21, 164.44], [519.96, 164.26]]}, {\"category\": \"cyclist\", \"corners_3d\": [[684.72, 221.52], [705.25, 221.96], [690.56, 226.97], [668.1, 226.43], [684.72, 165.28], [705.25, 165.21], [690.56, 164.43], [668.1, 164.51]]}]\n```" - } - ] - }, - { - "id": 2393, - "data_source": "Mono3DRefer", - "scene_name": "002565", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1132.66, 191.36], [1147.16, 191.63], [1099.73, 191.71], [1086.26, 191.44], [1132.66, 143.26], [1147.16, 142.36], [1099.73, 142.08], [1086.26, 143.0]]}, {\"category\": \"cyclist\", \"corners_3d\": [[951.16, 207.23], [980.01, 207.7], [979.17, 209.76], [948.19, 209.21], [951.16, 151.06], [980.01, 150.54], [979.17, 148.27], [948.19, 148.87]]}, {\"category\": \"cyclist\", \"corners_3d\": [[471.39, 259.48], [475.74, 263.47], [368.48, 267.16], [369.32, 262.81], [471.39, 164.82], [475.74, 164.02], [368.48, 163.29], [369.32, 164.16]]}]\n```" - } - ] - }, - { - "id": 2394, - "data_source": "Mono3DRefer", - "scene_name": "002565", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[749.15, 257.81], [739.89, 262.01], [679.87, 259.16], [691.81, 255.25], [749.15, 152.1], [739.89, 150.56], [679.87, 151.6], [691.81, 153.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[479.29, 268.86], [471.89, 273.57], [401.6, 273.38], [412.54, 268.69], [479.29, 159.06], [471.89, 157.92], [401.6, 157.97], [412.54, 159.1]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[481.11, 275.16], [473.05, 281.03], [398.5, 280.92], [410.91, 275.07], [481.11, 153.31], [473.05, 151.62], [398.5, 151.65], [410.91, 153.34]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[157.81, 259.35], [136.17, 262.39], [95.36, 261.54], [118.28, 258.55], [157.81, 176.06], [136.17, 175.89], [95.36, 175.94], [118.28, 176.11]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[137.29, 263.77], [118.18, 266.69], [72.7, 266.03], [93.2, 263.15], [137.29, 182.43], [118.18, 182.49], [72.7, 182.48], [93.2, 182.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[41.76, 274.05], [67.61, 270.1], [109.09, 270.5], [84.95, 274.48], [41.76, 165.36], [67.61, 166.0], [109.09, 165.93], [84.95, 165.29]]}]\n```" - } - ] - }, - { - "id": 2395, - "data_source": "Mono3DRefer", - "scene_name": "002565", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[212.46, 334.57], [382.93, 334.6], [161.18, 488.73], [-179.66, 488.61], [212.46, 190.88], [382.93, 190.88], [161.18, 201.25], [-179.66, 201.24]]}, {\"category\": \"car\", \"corners_3d\": [[1293.09, 194.3], [1348.33, 195.43], [1211.55, 195.37], [1166.82, 194.25], [1293.09, 154.01], [1348.33, 151.85], [1211.55, 151.95], [1166.82, 154.11]]}, {\"category\": \"car\", \"corners_3d\": [[482.93, 191.77], [454.62, 191.82], [460.76, 190.84], [486.61, 190.8], [482.93, 168.71], [454.62, 168.66], [460.76, 169.68], [486.61, 169.73]]}]\n```" - } - ] - }, - { - "id": 2396, - "data_source": "Mono3DRefer", - "scene_name": "002565", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[749.15, 257.81], [739.89, 262.01], [679.87, 259.16], [691.81, 255.25], [749.15, 152.1], [739.89, 150.56], [679.87, 151.6], [691.81, 153.04]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1132.66, 191.36], [1147.16, 191.63], [1099.73, 191.71], [1086.26, 191.44], [1132.66, 143.26], [1147.16, 142.36], [1099.73, 142.08], [1086.26, 143.0]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[479.29, 268.86], [471.89, 273.57], [401.6, 273.38], [412.54, 268.69], [479.29, 159.06], [471.89, 157.92], [401.6, 157.97], [412.54, 159.1]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[481.11, 275.16], [473.05, 281.03], [398.5, 280.92], [410.91, 275.07], [481.11, 153.31], [473.05, 151.62], [398.5, 151.65], [410.91, 153.34]]}, {\"category\": \"cyclist\", \"corners_3d\": [[951.16, 207.23], [980.01, 207.7], [979.17, 209.76], [948.19, 209.21], [951.16, 151.06], [980.01, 150.54], [979.17, 148.27], [948.19, 148.87]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[157.81, 259.35], [136.17, 262.39], [95.36, 261.54], [118.28, 258.55], [157.81, 176.06], [136.17, 175.89], [95.36, 175.94], [118.28, 176.11]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[137.29, 263.77], [118.18, 266.69], [72.7, 266.03], [93.2, 263.15], [137.29, 182.43], [118.18, 182.49], [72.7, 182.48], [93.2, 182.41]]}, {\"category\": \"cyclist\", \"corners_3d\": [[471.39, 259.48], [475.74, 263.47], [368.48, 267.16], [369.32, 262.81], [471.39, 164.82], [475.74, 164.02], [368.48, 163.29], [369.32, 164.16]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[41.76, 274.05], [67.61, 270.1], [109.09, 270.5], [84.95, 274.48], [41.76, 165.36], [67.61, 166.0], [109.09, 165.93], [84.95, 165.29]]}]\n```" - } - ] - }, - { - "id": 2397, - "data_source": "Mono3DRefer", - "scene_name": "002565", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[212.46, 334.57], [382.93, 334.6], [161.18, 488.73], [-179.66, 488.61], [212.46, 190.88], [382.93, 190.88], [161.18, 201.25], [-179.66, 201.24]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1132.66, 191.36], [1147.16, 191.63], [1099.73, 191.71], [1086.26, 191.44], [1132.66, 143.26], [1147.16, 142.36], [1099.73, 142.08], [1086.26, 143.0]]}, {\"category\": \"cyclist\", \"corners_3d\": [[951.16, 207.23], [980.01, 207.7], [979.17, 209.76], [948.19, 209.21], [951.16, 151.06], [980.01, 150.54], [979.17, 148.27], [948.19, 148.87]]}, {\"category\": \"cyclist\", \"corners_3d\": [[471.39, 259.48], [475.74, 263.47], [368.48, 267.16], [369.32, 262.81], [471.39, 164.82], [475.74, 164.02], [368.48, 163.29], [369.32, 164.16]]}, {\"category\": \"car\", \"corners_3d\": [[1293.09, 194.3], [1348.33, 195.43], [1211.55, 195.37], [1166.82, 194.25], [1293.09, 154.01], [1348.33, 151.85], [1211.55, 151.95], [1166.82, 154.11]]}, {\"category\": \"car\", \"corners_3d\": [[482.93, 191.77], [454.62, 191.82], [460.76, 190.84], [486.61, 190.8], [482.93, 168.71], [454.62, 168.66], [460.76, 169.68], [486.61, 169.73]]}]\n```" - } - ] - }, - { - "id": 2398, - "data_source": "Mono3DRefer", - "scene_name": "002565", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[212.46, 334.57], [382.93, 334.6], [161.18, 488.73], [-179.66, 488.61], [212.46, 190.88], [382.93, 190.88], [161.18, 201.25], [-179.66, 201.24]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[749.15, 257.81], [739.89, 262.01], [679.87, 259.16], [691.81, 255.25], [749.15, 152.1], [739.89, 150.56], [679.87, 151.6], [691.81, 153.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[479.29, 268.86], [471.89, 273.57], [401.6, 273.38], [412.54, 268.69], [479.29, 159.06], [471.89, 157.92], [401.6, 157.97], [412.54, 159.1]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[481.11, 275.16], [473.05, 281.03], [398.5, 280.92], [410.91, 275.07], [481.11, 153.31], [473.05, 151.62], [398.5, 151.65], [410.91, 153.34]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[157.81, 259.35], [136.17, 262.39], [95.36, 261.54], [118.28, 258.55], [157.81, 176.06], [136.17, 175.89], [95.36, 175.94], [118.28, 176.11]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[137.29, 263.77], [118.18, 266.69], [72.7, 266.03], [93.2, 263.15], [137.29, 182.43], [118.18, 182.49], [72.7, 182.48], [93.2, 182.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[41.76, 274.05], [67.61, 270.1], [109.09, 270.5], [84.95, 274.48], [41.76, 165.36], [67.61, 166.0], [109.09, 165.93], [84.95, 165.29]]}, {\"category\": \"car\", \"corners_3d\": [[1293.09, 194.3], [1348.33, 195.43], [1211.55, 195.37], [1166.82, 194.25], [1293.09, 154.01], [1348.33, 151.85], [1211.55, 151.95], [1166.82, 154.11]]}, {\"category\": \"car\", \"corners_3d\": [[482.93, 191.77], [454.62, 191.82], [460.76, 190.84], [486.61, 190.8], [482.93, 168.71], [454.62, 168.66], [460.76, 169.68], [486.61, 169.73]]}]\n```" - } - ] - }, - { - "id": 2399, - "data_source": "Mono3DRefer", - "scene_name": "002565", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[212.46, 334.57], [382.93, 334.6], [161.18, 488.73], [-179.66, 488.61], [212.46, 190.88], [382.93, 190.88], [161.18, 201.25], [-179.66, 201.24]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[749.15, 257.81], [739.89, 262.01], [679.87, 259.16], [691.81, 255.25], [749.15, 152.1], [739.89, 150.56], [679.87, 151.6], [691.81, 153.04]]}, {\"category\": \"cyclist\", \"corners_3d\": [[1132.66, 191.36], [1147.16, 191.63], [1099.73, 191.71], [1086.26, 191.44], [1132.66, 143.26], [1147.16, 142.36], [1099.73, 142.08], [1086.26, 143.0]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[479.29, 268.86], [471.89, 273.57], [401.6, 273.38], [412.54, 268.69], [479.29, 159.06], [471.89, 157.92], [401.6, 157.97], [412.54, 159.1]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[481.11, 275.16], [473.05, 281.03], [398.5, 280.92], [410.91, 275.07], [481.11, 153.31], [473.05, 151.62], [398.5, 151.65], [410.91, 153.34]]}, {\"category\": \"cyclist\", \"corners_3d\": [[951.16, 207.23], [980.01, 207.7], [979.17, 209.76], [948.19, 209.21], [951.16, 151.06], [980.01, 150.54], [979.17, 148.27], [948.19, 148.87]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[157.81, 259.35], [136.17, 262.39], [95.36, 261.54], [118.28, 258.55], [157.81, 176.06], [136.17, 175.89], [95.36, 175.94], [118.28, 176.11]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[137.29, 263.77], [118.18, 266.69], [72.7, 266.03], [93.2, 263.15], [137.29, 182.43], [118.18, 182.49], [72.7, 182.48], [93.2, 182.41]]}, {\"category\": \"cyclist\", \"corners_3d\": [[471.39, 259.48], [475.74, 263.47], [368.48, 267.16], [369.32, 262.81], [471.39, 164.82], [475.74, 164.02], [368.48, 163.29], [369.32, 164.16]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[41.76, 274.05], [67.61, 270.1], [109.09, 270.5], [84.95, 274.48], [41.76, 165.36], [67.61, 166.0], [109.09, 165.93], [84.95, 165.29]]}, {\"category\": \"car\", \"corners_3d\": [[1293.09, 194.3], [1348.33, 195.43], [1211.55, 195.37], [1166.82, 194.25], [1293.09, 154.01], [1348.33, 151.85], [1211.55, 151.95], [1166.82, 154.11]]}, {\"category\": \"car\", \"corners_3d\": [[482.93, 191.77], [454.62, 191.82], [460.76, 190.84], [486.61, 190.8], [482.93, 168.71], [454.62, 168.66], [460.76, 169.68], [486.61, 169.73]]}]\n```" - } - ] - }, - { - "id": 2400, - "data_source": "Mono3DRefer", - "scene_name": "002578", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[784.24, 223.73], [842.76, 224.57], [859.18, 235.34], [788.38, 234.12], [784.24, 170.81], [842.76, 170.77], [859.18, 170.34], [788.38, 170.39]]}, {\"category\": \"car\", \"corners_3d\": [[397.94, 255.17], [330.03, 252.44], [442.12, 234.95], [497.69, 236.6], [397.94, 185.99], [330.03, 185.55], [442.12, 182.76], [497.69, 183.03]]}, {\"category\": \"car\", \"corners_3d\": [[1183.8, 195.07], [1205.74, 196.23], [1106.65, 195.65], [1090.02, 194.55], [1183.8, 160.5], [1205.74, 159.85], [1106.65, 160.18], [1090.02, 160.79]]}]\n```" - } - ] - }, - { - "id": 2401, - "data_source": "Mono3DRefer", - "scene_name": "002582", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[67.19, 798.37], [-710.46, 861.74], [112.46, 407.15], [366.62, 399.34], [67.19, 176.49], [-710.46, 176.86], [112.46, 174.22], [366.62, 174.17]]}, {\"category\": \"car\", \"corners_3d\": [[465.43, 310.55], [342.07, 309.86], [431.32, 267.48], [516.71, 267.81], [465.43, 183.36], [342.07, 183.3], [431.32, 180.07], [516.71, 180.1]]}]\n```" - } - ] - }, - { - "id": 2402, - "data_source": "Mono3DRefer", - "scene_name": "002589", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[289.4, 265.3], [243.04, 277.09], [9.4, 275.0], [80.89, 263.65], [289.4, 199.12], [243.04, 202.46], [9.4, 201.87], [80.89, 198.65]]}]\n```" - } - ] - }, - { - "id": 2403, - "data_source": "Mono3DRefer", - "scene_name": "002590", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[664.7, 183.3], [682.59, 183.33], [682.14, 183.97], [663.15, 183.94], [664.7, 166.42], [682.59, 166.4], [682.14, 166.0], [663.15, 166.02]]}]\n```" - } - ] - }, - { - "id": 2404, - "data_source": "Mono3DRefer", - "scene_name": "002590", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[738.04, 277.83], [846.98, 277.69], [990.87, 339.49], [817.88, 339.86], [738.04, 151.47], [846.98, 151.5], [990.87, 138.9], [817.88, 138.83]]}]\n```" - } - ] - }, - { - "id": 2405, - "data_source": "Mono3DRefer", - "scene_name": "002590", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[738.04, 277.83], [846.98, 277.69], [990.87, 339.49], [817.88, 339.86], [738.04, 151.47], [846.98, 151.5], [990.87, 138.9], [817.88, 138.83]]}, {\"category\": \"car\", \"corners_3d\": [[664.7, 183.3], [682.59, 183.33], [682.14, 183.97], [663.15, 183.94], [664.7, 166.42], [682.59, 166.4], [682.14, 166.0], [663.15, 166.02]]}]\n```" - } - ] - }, - { - "id": 2406, - "data_source": "Mono3DRefer", - "scene_name": "002591", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[501.1, 282.6], [397.71, 284.47], [424.33, 260.56], [505.51, 259.4], [501.1, 187.78], [397.71, 188.03], [424.33, 184.78], [505.51, 184.62]]}, {\"category\": \"car\", \"corners_3d\": [[516.84, 244.63], [460.13, 244.64], [485.41, 232.44], [532.49, 232.44], [516.84, 186.83], [460.13, 186.83], [485.41, 184.46], [532.49, 184.46]]}]\n```" - } - ] - }, - { - "id": 2407, - "data_source": "Mono3DRefer", - "scene_name": "002593", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[398.52, 209.73], [407.74, 208.05], [481.9, 208.01], [476.22, 209.68], [398.52, 178.5], [407.74, 178.25], [481.9, 178.24], [476.22, 178.5]]}]\n```" - } - ] - }, - { - "id": 2408, - "data_source": "Mono3DRefer", - "scene_name": "002603", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[372.4, 206.6], [343.4, 206.57], [362.95, 204.22], [389.94, 204.24], [372.4, 181.13], [343.4, 181.12], [362.95, 180.55], [389.94, 180.55]]}, {\"category\": \"car\", \"corners_3d\": [[225.48, 220.86], [187.63, 220.86], [235.22, 215.44], [268.79, 215.44], [225.48, 186.15], [187.63, 186.15], [235.22, 184.65], [268.79, 184.65]]}]\n```" - } - ] - }, - { - "id": 2409, - "data_source": "Mono3DRefer", - "scene_name": "002605", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[877.39, 406.79], [789.04, 412.02], [747.33, 380.55], [824.73, 376.64], [877.39, 116.16], [789.04, 114.67], [747.33, 123.62], [824.73, 124.73]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[967.99, 397.66], [894.32, 405.32], [846.95, 383.77], [915.23, 377.48], [967.99, 149.81], [894.32, 148.73], [846.95, 151.78], [915.23, 152.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1117.24, 322.76], [1099.68, 329.06], [1019.53, 317.67], [1038.95, 312.28], [1117.24, 153.75], [1099.68, 152.56], [1019.53, 154.7], [1038.95, 155.72]]}]\n```" - } - ] - }, - { - "id": 2410, - "data_source": "Mono3DRefer", - "scene_name": "002624", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[540.23, 228.18], [487.53, 228.26], [503.59, 219.95], [548.37, 219.89], [540.23, 172.84], [487.53, 172.84], [503.59, 172.85], [548.37, 172.85]]}, {\"category\": \"car\", \"corners_3d\": [[568.36, 200.92], [546.37, 200.91], [551.82, 198.92], [572.26, 198.93], [568.36, 176.45], [546.37, 176.45], [551.82, 176.19], [572.26, 176.19]]}]\n```" - } - ] - }, - { - "id": 2411, - "data_source": "Mono3DRefer", - "scene_name": "002628", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[457.78, 199.73], [437.74, 199.7], [449.14, 198.17], [468.04, 198.19], [457.78, 183.84], [437.74, 183.83], [449.14, 183.2], [468.04, 183.21]]}]\n```" - } - ] - }, - { - "id": 2412, - "data_source": "Mono3DRefer", - "scene_name": "002628", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[593.76, 189.84], [617.53, 189.85], [617.56, 192.96], [589.44, 192.95], [593.76, 164.08], [617.53, 164.08], [617.56, 162.48], [589.44, 162.48]]}]\n```" - } - ] - }, - { - "id": 2413, - "data_source": "Mono3DRefer", - "scene_name": "002628", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[593.76, 189.84], [617.53, 189.85], [617.56, 192.96], [589.44, 192.95], [593.76, 164.08], [617.53, 164.08], [617.56, 162.48], [589.44, 162.48]]}, {\"category\": \"car\", \"corners_3d\": [[457.78, 199.73], [437.74, 199.7], [449.14, 198.17], [468.04, 198.19], [457.78, 183.84], [437.74, 183.83], [449.14, 183.2], [468.04, 183.21]]}]\n```" - } - ] - }, - { - "id": 2414, - "data_source": "Mono3DRefer", - "scene_name": "002630", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[324.03, 208.18], [315.23, 208.13], [327.63, 206.98], [336.17, 207.03], [324.03, 181.57], [315.23, 181.56], [327.63, 181.27], [336.17, 181.29]]}]\n```" - } - ] - }, - { - "id": 2415, - "data_source": "Mono3DRefer", - "scene_name": "002630", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[364.69, 210.45], [337.33, 210.3], [361.97, 207.62], [387.45, 207.75], [364.69, 186.64], [337.33, 186.59], [361.97, 185.6], [387.45, 185.65]]}, {\"category\": \"car\", \"corners_3d\": [[53.53, 249.07], [-11.85, 249.07], [79.06, 237.91], [134.86, 237.91], [53.53, 198.98], [-11.85, 198.98], [79.06, 195.15], [134.86, 195.15]]}]\n```" - } - ] - }, - { - "id": 2416, - "data_source": "Mono3DRefer", - "scene_name": "002630", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[366.99, 202.73], [342.54, 202.7], [366.07, 200.26], [388.54, 200.29], [366.99, 174.65], [342.54, 174.65], [366.07, 174.5], [388.54, 174.51]]}]\n```" - } - ] - }, - { - "id": 2417, - "data_source": "Mono3DRefer", - "scene_name": "002630", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[364.69, 210.45], [337.33, 210.3], [361.97, 207.62], [387.45, 207.75], [364.69, 186.64], [337.33, 186.59], [361.97, 185.6], [387.45, 185.65]]}, {\"category\": \"car\", \"corners_3d\": [[53.53, 249.07], [-11.85, 249.07], [79.06, 237.91], [134.86, 237.91], [53.53, 198.98], [-11.85, 198.98], [79.06, 195.15], [134.86, 195.15]]}, {\"category\": \"cyclist\", \"corners_3d\": [[324.03, 208.18], [315.23, 208.13], [327.63, 206.98], [336.17, 207.03], [324.03, 181.57], [315.23, 181.56], [327.63, 181.27], [336.17, 181.29]]}]\n```" - } - ] - }, - { - "id": 2418, - "data_source": "Mono3DRefer", - "scene_name": "002630", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[366.99, 202.73], [342.54, 202.7], [366.07, 200.26], [388.54, 200.29], [366.99, 174.65], [342.54, 174.65], [366.07, 174.5], [388.54, 174.51]]}, {\"category\": \"cyclist\", \"corners_3d\": [[324.03, 208.18], [315.23, 208.13], [327.63, 206.98], [336.17, 207.03], [324.03, 181.57], [315.23, 181.56], [327.63, 181.27], [336.17, 181.29]]}]\n```" - } - ] - }, - { - "id": 2419, - "data_source": "Mono3DRefer", - "scene_name": "002630", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[364.69, 210.45], [337.33, 210.3], [361.97, 207.62], [387.45, 207.75], [364.69, 186.64], [337.33, 186.59], [361.97, 185.6], [387.45, 185.65]]}, {\"category\": \"van\", \"corners_3d\": [[366.99, 202.73], [342.54, 202.7], [366.07, 200.26], [388.54, 200.29], [366.99, 174.65], [342.54, 174.65], [366.07, 174.5], [388.54, 174.51]]}, {\"category\": \"car\", \"corners_3d\": [[53.53, 249.07], [-11.85, 249.07], [79.06, 237.91], [134.86, 237.91], [53.53, 198.98], [-11.85, 198.98], [79.06, 195.15], [134.86, 195.15]]}]\n```" - } - ] - }, - { - "id": 2420, - "data_source": "Mono3DRefer", - "scene_name": "002630", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[364.69, 210.45], [337.33, 210.3], [361.97, 207.62], [387.45, 207.75], [364.69, 186.64], [337.33, 186.59], [361.97, 185.6], [387.45, 185.65]]}, {\"category\": \"van\", \"corners_3d\": [[366.99, 202.73], [342.54, 202.7], [366.07, 200.26], [388.54, 200.29], [366.99, 174.65], [342.54, 174.65], [366.07, 174.5], [388.54, 174.51]]}, {\"category\": \"car\", \"corners_3d\": [[53.53, 249.07], [-11.85, 249.07], [79.06, 237.91], [134.86, 237.91], [53.53, 198.98], [-11.85, 198.98], [79.06, 195.15], [134.86, 195.15]]}, {\"category\": \"cyclist\", \"corners_3d\": [[324.03, 208.18], [315.23, 208.13], [327.63, 206.98], [336.17, 207.03], [324.03, 181.57], [315.23, 181.56], [327.63, 181.27], [336.17, 181.29]]}]\n```" - } - ] - }, - { - "id": 2421, - "data_source": "Mono3DRefer", - "scene_name": "002631", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-551.26, 396.9], [-472.36, 358.38], [83.33, 301.85], [107.17, 319.37], [-551.26, 143.85], [-472.36, 148.84], [83.33, 156.16], [107.17, 153.89]]}]\n```" - } - ] - }, - { - "id": 2422, - "data_source": "Mono3DRefer", - "scene_name": "002631", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[885.46, 230.09], [839.25, 227.34], [897.17, 221.36], [940.66, 223.52], [885.46, 183.08], [839.25, 182.59], [897.17, 181.52], [940.66, 181.9]]}, {\"category\": \"car\", \"corners_3d\": [[1013.03, 217.34], [971.18, 215.63], [1006.07, 211.85], [1045.36, 213.27], [1013.03, 184.2], [971.18, 183.77], [1006.07, 182.81], [1045.36, 183.17]]}, {\"category\": \"car\", \"corners_3d\": [[1077.03, 212.74], [1040.03, 211.47], [1061.68, 208.91], [1096.81, 210.02], [1077.03, 183.23], [1040.03, 182.9], [1061.68, 182.23], [1096.81, 182.52]]}, {\"category\": \"car\", \"corners_3d\": [[1236.4, 217.15], [1286.48, 218.84], [1278.18, 223.48], [1223.59, 221.44], [1236.4, 183.07], [1286.48, 183.46], [1278.18, 184.53], [1223.59, 184.06]]}, {\"category\": \"car\", \"corners_3d\": [[1119.24, 244.15], [1204.86, 249.26], [1176.31, 262.57], [1079.15, 255.61], [1119.24, 178.42], [1204.86, 178.82], [1176.31, 179.85], [1079.15, 179.31]]}, {\"category\": \"car\", \"corners_3d\": [[615.23, 262.31], [565.76, 255.8], [685.72, 243.61], [735.36, 248.29], [615.23, 182.83], [565.76, 182.11], [685.72, 180.75], [735.36, 181.27]]}]\n```" - } - ] - }, - { - "id": 2423, - "data_source": "Mono3DRefer", - "scene_name": "002631", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[885.46, 230.09], [839.25, 227.34], [897.17, 221.36], [940.66, 223.52], [885.46, 183.08], [839.25, 182.59], [897.17, 181.52], [940.66, 181.9]]}, {\"category\": \"car\", \"corners_3d\": [[1013.03, 217.34], [971.18, 215.63], [1006.07, 211.85], [1045.36, 213.27], [1013.03, 184.2], [971.18, 183.77], [1006.07, 182.81], [1045.36, 183.17]]}, {\"category\": \"car\", \"corners_3d\": [[1077.03, 212.74], [1040.03, 211.47], [1061.68, 208.91], [1096.81, 210.02], [1077.03, 183.23], [1040.03, 182.9], [1061.68, 182.23], [1096.81, 182.52]]}, {\"category\": \"car\", \"corners_3d\": [[1236.4, 217.15], [1286.48, 218.84], [1278.18, 223.48], [1223.59, 221.44], [1236.4, 183.07], [1286.48, 183.46], [1278.18, 184.53], [1223.59, 184.06]]}, {\"category\": \"car\", \"corners_3d\": [[1119.24, 244.15], [1204.86, 249.26], [1176.31, 262.57], [1079.15, 255.61], [1119.24, 178.42], [1204.86, 178.82], [1176.31, 179.85], [1079.15, 179.31]]}, {\"category\": \"van\", \"corners_3d\": [[-551.26, 396.9], [-472.36, 358.38], [83.33, 301.85], [107.17, 319.37], [-551.26, 143.85], [-472.36, 148.84], [83.33, 156.16], [107.17, 153.89]]}, {\"category\": \"car\", \"corners_3d\": [[615.23, 262.31], [565.76, 255.8], [685.72, 243.61], [735.36, 248.29], [615.23, 182.83], [565.76, 182.11], [685.72, 180.75], [735.36, 181.27]]}]\n```" - } - ] - }, - { - "id": 2424, - "data_source": "Mono3DRefer", - "scene_name": "002641", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[448.85, 297.54], [329.94, 297.56], [411.42, 261.14], [495.61, 261.13], [448.85, 173.56], [329.94, 173.56], [411.42, 173.35], [495.61, 173.35]]}, {\"category\": \"car\", \"corners_3d\": [[809.76, 366.65], [997.27, 368.19], [1396.15, 581.36], [1000.81, 574.69], [809.76, 207.63], [997.27, 207.91], [1396.15, 246.17], [1000.81, 244.97]]}, {\"category\": \"car\", \"corners_3d\": [[486.88, 230.24], [542.11, 230.38], [523.56, 242.27], [457.0, 242.07], [486.88, 183.79], [542.11, 183.82], [523.56, 186.08], [457.0, 186.05]]}, {\"category\": \"car\", \"corners_3d\": [[671.68, 207.4], [638.42, 207.45], [633.47, 204.1], [663.51, 204.06], [671.68, 176.57], [638.42, 176.58], [633.47, 176.22], [663.51, 176.21]]}]\n```" - } - ] - }, - { - "id": 2425, - "data_source": "Mono3DRefer", - "scene_name": "002641", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[241.03, 640.81], [89.72, 639.9], [254.57, 493.66], [358.66, 494.09], [241.03, 139.52], [89.72, 139.58], [254.57, 150.0], [358.66, 149.97]]}]\n```" - } - ] - }, - { - "id": 2426, - "data_source": "Mono3DRefer", - "scene_name": "002641", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[577.79, 193.31], [553.14, 193.24], [561.65, 191.81], [584.59, 191.86], [577.79, 168.24], [553.14, 168.26], [561.65, 168.58], [584.59, 168.57]]}]\n```" - } - ] - }, - { - "id": 2427, - "data_source": "Mono3DRefer", - "scene_name": "002641", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[448.85, 297.54], [329.94, 297.56], [411.42, 261.14], [495.61, 261.13], [448.85, 173.56], [329.94, 173.56], [411.42, 173.35], [495.61, 173.35]]}, {\"category\": \"cyclist\", \"corners_3d\": [[241.03, 640.81], [89.72, 639.9], [254.57, 493.66], [358.66, 494.09], [241.03, 139.52], [89.72, 139.58], [254.57, 150.0], [358.66, 149.97]]}, {\"category\": \"car\", \"corners_3d\": [[809.76, 366.65], [997.27, 368.19], [1396.15, 581.36], [1000.81, 574.69], [809.76, 207.63], [997.27, 207.91], [1396.15, 246.17], [1000.81, 244.97]]}, {\"category\": \"car\", \"corners_3d\": [[486.88, 230.24], [542.11, 230.38], [523.56, 242.27], [457.0, 242.07], [486.88, 183.79], [542.11, 183.82], [523.56, 186.08], [457.0, 186.05]]}, {\"category\": \"car\", \"corners_3d\": [[671.68, 207.4], [638.42, 207.45], [633.47, 204.1], [663.51, 204.06], [671.68, 176.57], [638.42, 176.58], [633.47, 176.22], [663.51, 176.21]]}]\n```" - } - ] - }, - { - "id": 2428, - "data_source": "Mono3DRefer", - "scene_name": "002641", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[448.85, 297.54], [329.94, 297.56], [411.42, 261.14], [495.61, 261.13], [448.85, 173.56], [329.94, 173.56], [411.42, 173.35], [495.61, 173.35]]}, {\"category\": \"car\", \"corners_3d\": [[809.76, 366.65], [997.27, 368.19], [1396.15, 581.36], [1000.81, 574.69], [809.76, 207.63], [997.27, 207.91], [1396.15, 246.17], [1000.81, 244.97]]}, {\"category\": \"car\", \"corners_3d\": [[486.88, 230.24], [542.11, 230.38], [523.56, 242.27], [457.0, 242.07], [486.88, 183.79], [542.11, 183.82], [523.56, 186.08], [457.0, 186.05]]}, {\"category\": \"car\", \"corners_3d\": [[671.68, 207.4], [638.42, 207.45], [633.47, 204.1], [663.51, 204.06], [671.68, 176.57], [638.42, 176.58], [633.47, 176.22], [663.51, 176.21]]}, {\"category\": \"van\", \"corners_3d\": [[577.79, 193.31], [553.14, 193.24], [561.65, 191.81], [584.59, 191.86], [577.79, 168.24], [553.14, 168.26], [561.65, 168.58], [584.59, 168.57]]}]\n```" - } - ] - }, - { - "id": 2429, - "data_source": "Mono3DRefer", - "scene_name": "002641", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[241.03, 640.81], [89.72, 639.9], [254.57, 493.66], [358.66, 494.09], [241.03, 139.52], [89.72, 139.58], [254.57, 150.0], [358.66, 149.97]]}, {\"category\": \"van\", \"corners_3d\": [[577.79, 193.31], [553.14, 193.24], [561.65, 191.81], [584.59, 191.86], [577.79, 168.24], [553.14, 168.26], [561.65, 168.58], [584.59, 168.57]]}]\n```" - } - ] - }, - { - "id": 2430, - "data_source": "Mono3DRefer", - "scene_name": "002641", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[448.85, 297.54], [329.94, 297.56], [411.42, 261.14], [495.61, 261.13], [448.85, 173.56], [329.94, 173.56], [411.42, 173.35], [495.61, 173.35]]}, {\"category\": \"cyclist\", \"corners_3d\": [[241.03, 640.81], [89.72, 639.9], [254.57, 493.66], [358.66, 494.09], [241.03, 139.52], [89.72, 139.58], [254.57, 150.0], [358.66, 149.97]]}, {\"category\": \"car\", \"corners_3d\": [[809.76, 366.65], [997.27, 368.19], [1396.15, 581.36], [1000.81, 574.69], [809.76, 207.63], [997.27, 207.91], [1396.15, 246.17], [1000.81, 244.97]]}, {\"category\": \"car\", \"corners_3d\": [[486.88, 230.24], [542.11, 230.38], [523.56, 242.27], [457.0, 242.07], [486.88, 183.79], [542.11, 183.82], [523.56, 186.08], [457.0, 186.05]]}, {\"category\": \"car\", \"corners_3d\": [[671.68, 207.4], [638.42, 207.45], [633.47, 204.1], [663.51, 204.06], [671.68, 176.57], [638.42, 176.58], [633.47, 176.22], [663.51, 176.21]]}, {\"category\": \"van\", \"corners_3d\": [[577.79, 193.31], [553.14, 193.24], [561.65, 191.81], [584.59, 191.86], [577.79, 168.24], [553.14, 168.26], [561.65, 168.58], [584.59, 168.57]]}]\n```" - } - ] - }, - { - "id": 2431, - "data_source": "Mono3DRefer", - "scene_name": "002647", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[529.19, 262.86], [602.21, 268.17], [492.12, 294.26], [408.2, 285.78], [529.19, 189.68], [602.21, 190.68], [492.12, 195.55], [408.2, 193.97]]}, {\"category\": \"car\", \"corners_3d\": [[469.97, 256.72], [527.14, 261.12], [361.35, 290.05], [297.47, 282.41], [469.97, 189.8], [527.14, 190.69], [361.35, 196.53], [297.47, 194.99]]}, {\"category\": \"car\", \"corners_3d\": [[249.07, 250.4], [290.98, 253.68], [116.49, 273.71], [73.48, 268.64], [249.07, 194.57], [290.98, 195.49], [116.49, 201.1], [73.48, 199.68]]}, {\"category\": \"car\", \"corners_3d\": [[-22.68, 268.38], [-58.16, 264.28], [129.4, 247.54], [164.96, 250.25], [-22.68, 188.04], [-58.16, 187.39], [129.4, 184.73], [164.96, 185.16]]}]\n```" - } - ] - }, - { - "id": 2432, - "data_source": "Mono3DRefer", - "scene_name": "002651", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[512.66, 324.36], [519.39, 313.62], [618.43, 313.6], [619.67, 324.32], [512.66, 164.33], [519.39, 165.53], [618.43, 165.54], [619.67, 164.33]]}]\n```" - } - ] - }, - { - "id": 2433, - "data_source": "Mono3DRefer", - "scene_name": "002654", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[752.98, 197.5], [764.19, 197.47], [775.57, 198.84], [763.74, 198.87], [752.98, 158.24], [764.19, 158.25], [775.57, 157.44], [763.74, 157.42]]}]\n```" - } - ] - }, - { - "id": 2434, - "data_source": "Mono3DRefer", - "scene_name": "002654", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[254.72, 316.06], [112.97, 316.08], [264.9, 272.21], [363.22, 272.2], [254.72, 197.87], [112.97, 197.88], [264.9, 190.21], [363.22, 190.21]]}]\n```" - } - ] - }, - { - "id": 2435, - "data_source": "Mono3DRefer", - "scene_name": "002654", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[254.72, 316.06], [112.97, 316.08], [264.9, 272.21], [363.22, 272.2], [254.72, 197.87], [112.97, 197.88], [264.9, 190.21], [363.22, 190.21]]}, {\"category\": \"cyclist\", \"corners_3d\": [[752.98, 197.5], [764.19, 197.47], [775.57, 198.84], [763.74, 198.87], [752.98, 158.24], [764.19, 158.25], [775.57, 157.44], [763.74, 157.42]]}]\n```" - } - ] - }, - { - "id": 2436, - "data_source": "Mono3DRefer", - "scene_name": "002655", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[555.46, 212.33], [516.05, 212.39], [524.37, 207.78], [559.17, 207.73], [555.46, 170.96], [516.05, 170.95], [524.37, 171.17], [559.17, 171.18]]}, {\"category\": \"car\", \"corners_3d\": [[571.95, 196.93], [553.18, 196.91], [557.47, 195.45], [575.09, 195.46], [571.95, 176.04], [553.18, 176.04], [557.47, 175.85], [575.09, 175.85]]}]\n```" - } - ] - }, - { - "id": 2437, - "data_source": "Mono3DRefer", - "scene_name": "002656", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-9.94, 276.98], [-107.85, 276.99], [63.54, 252.09], [138.02, 252.08], [-9.94, 191.77], [-107.85, 191.77], [63.54, 187.25], [138.02, 187.25]]}, {\"category\": \"car\", \"corners_3d\": [[281.7, 212.62], [249.87, 212.57], [287.16, 208.68], [315.92, 208.72], [281.7, 182.97], [249.87, 182.96], [287.16, 181.97], [315.92, 181.98]]}, {\"category\": \"car\", \"corners_3d\": [[403.79, 201.22], [378.54, 201.19], [394.03, 199.45], [417.75, 199.47], [403.79, 178.99], [378.54, 178.98], [394.03, 178.6], [417.75, 178.61]]}]\n```" - } - ] - }, - { - "id": 2438, - "data_source": "Mono3DRefer", - "scene_name": "002657", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[199.72, 406.52], [-58.57, 409.86], [244.82, 297.27], [379.08, 296.34], [199.72, 175.87], [-58.57, 175.92], [244.82, 174.46], [379.08, 174.45]]}, {\"category\": \"car\", \"corners_3d\": [[644.45, 208.84], [666.14, 208.83], [671.86, 212.09], [648.21, 212.1], [644.45, 189.47], [666.14, 189.47], [671.86, 190.97], [648.21, 190.97]]}]\n```" - } - ] - }, - { - "id": 2439, - "data_source": "Mono3DRefer", - "scene_name": "002662", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[182.07, 228.74], [132.68, 229.0], [199.18, 220.23], [240.63, 220.05], [182.07, 177.76], [132.68, 177.78], [199.18, 177.01], [240.63, 176.99]]}, {\"category\": \"car\", \"corners_3d\": [[584.78, 220.72], [633.17, 220.63], [640.08, 228.02], [584.21, 228.14], [584.78, 181.73], [633.17, 181.71], [640.08, 183.08], [584.21, 183.11]]}, {\"category\": \"car\", \"corners_3d\": [[91.68, 298.22], [214.53, 283.8], [350.73, 311.23], [215.23, 334.4], [91.68, 186.26], [214.53, 184.72], [350.73, 187.65], [215.23, 190.13]]}, {\"category\": \"car\", \"corners_3d\": [[238.5, 268.37], [325.67, 259.88], [446.26, 278.3], [352.98, 291.02], [238.5, 166.62], [325.67, 167.17], [446.26, 165.97], [352.98, 165.14]]}, {\"category\": \"car\", \"corners_3d\": [[362.58, 236.92], [413.92, 233.3], [505.51, 242.65], [452.07, 247.52], [362.58, 185.87], [413.92, 185.14], [505.51, 187.04], [452.07, 188.03]]}, {\"category\": \"car\", \"corners_3d\": [[385.92, 229.53], [428.49, 226.5], [510.58, 232.57], [468.09, 236.34], [385.92, 182.86], [428.49, 182.33], [510.58, 183.4], [468.09, 184.07]]}, {\"category\": \"car\", \"corners_3d\": [[443.31, 216.26], [474.31, 214.66], [529.04, 218.25], [497.55, 220.15], [443.31, 183.82], [474.31, 183.41], [529.04, 184.32], [497.55, 184.8]]}, {\"category\": \"car\", \"corners_3d\": [[474.27, 212.09], [498.36, 210.73], [548.15, 213.05], [524.43, 214.58], [474.27, 182.71], [498.36, 182.37], [548.15, 182.95], [524.43, 183.33]]}, {\"category\": \"car\", \"corners_3d\": [[694.22, 229.08], [746.15, 229.04], [776.69, 241.02], [713.71, 241.07], [694.22, 184.87], [746.15, 184.86], [776.69, 187.42], [713.71, 187.44]]}, {\"category\": \"car\", \"corners_3d\": [[658.33, 208.44], [687.07, 208.47], [693.84, 212.33], [661.98, 212.29], [658.33, 184.25], [687.07, 184.26], [693.84, 185.5], [661.98, 185.49]]}, {\"category\": \"car\", \"corners_3d\": [[590.46, 207.91], [621.88, 207.84], [626.85, 211.47], [592.17, 211.56], [590.46, 179.26], [621.88, 179.24], [626.85, 179.91], [592.17, 179.92]]}, {\"category\": \"car\", \"corners_3d\": [[-197.03, 351.56], [-383.49, 353.78], [-52.49, 291.16], [67.29, 290.21], [-197.03, 203.67], [-383.49, 204.05], [-52.49, 193.26], [67.29, 193.09]]}, {\"category\": \"car\", \"corners_3d\": [[11.8, 251.26], [-63.75, 251.19], [85.89, 233.95], [144.91, 233.99], [11.8, 186.57], [-63.75, 186.56], [85.89, 183.54], [144.91, 183.55]]}]\n```" - } - ] - }, - { - "id": 2440, - "data_source": "Mono3DRefer", - "scene_name": "002662", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[705.64, 245.01], [787.04, 244.77], [877.55, 277.55], [759.3, 278.06], [705.64, 151.15], [787.04, 151.22], [877.55, 141.36], [759.3, 141.21]]}, {\"category\": \"van\", \"corners_3d\": [[637.73, 194.13], [660.92, 194.12], [667.13, 196.14], [641.74, 196.16], [637.73, 166.44], [660.92, 166.45], [667.13, 165.84], [641.74, 165.83]]}, {\"category\": \"van\", \"corners_3d\": [[279.53, 239.11], [207.9, 239.24], [287.67, 225.55], [344.41, 225.46], [279.53, 150.47], [207.9, 150.43], [287.67, 155.05], [344.41, 155.08]]}]\n```" - } - ] - }, - { - "id": 2441, - "data_source": "Mono3DRefer", - "scene_name": "002662", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[182.07, 228.74], [132.68, 229.0], [199.18, 220.23], [240.63, 220.05], [182.07, 177.76], [132.68, 177.78], [199.18, 177.01], [240.63, 176.99]]}, {\"category\": \"car\", \"corners_3d\": [[584.78, 220.72], [633.17, 220.63], [640.08, 228.02], [584.21, 228.14], [584.78, 181.73], [633.17, 181.71], [640.08, 183.08], [584.21, 183.11]]}, {\"category\": \"car\", \"corners_3d\": [[91.68, 298.22], [214.53, 283.8], [350.73, 311.23], [215.23, 334.4], [91.68, 186.26], [214.53, 184.72], [350.73, 187.65], [215.23, 190.13]]}, {\"category\": \"car\", \"corners_3d\": [[238.5, 268.37], [325.67, 259.88], [446.26, 278.3], [352.98, 291.02], [238.5, 166.62], [325.67, 167.17], [446.26, 165.97], [352.98, 165.14]]}, {\"category\": \"car\", \"corners_3d\": [[362.58, 236.92], [413.92, 233.3], [505.51, 242.65], [452.07, 247.52], [362.58, 185.87], [413.92, 185.14], [505.51, 187.04], [452.07, 188.03]]}, {\"category\": \"car\", \"corners_3d\": [[385.92, 229.53], [428.49, 226.5], [510.58, 232.57], [468.09, 236.34], [385.92, 182.86], [428.49, 182.33], [510.58, 183.4], [468.09, 184.07]]}, {\"category\": \"car\", \"corners_3d\": [[443.31, 216.26], [474.31, 214.66], [529.04, 218.25], [497.55, 220.15], [443.31, 183.82], [474.31, 183.41], [529.04, 184.32], [497.55, 184.8]]}, {\"category\": \"car\", \"corners_3d\": [[474.27, 212.09], [498.36, 210.73], [548.15, 213.05], [524.43, 214.58], [474.27, 182.71], [498.36, 182.37], [548.15, 182.95], [524.43, 183.33]]}, {\"category\": \"van\", \"corners_3d\": [[705.64, 245.01], [787.04, 244.77], [877.55, 277.55], [759.3, 278.06], [705.64, 151.15], [787.04, 151.22], [877.55, 141.36], [759.3, 141.21]]}, {\"category\": \"car\", \"corners_3d\": [[694.22, 229.08], [746.15, 229.04], [776.69, 241.02], [713.71, 241.07], [694.22, 184.87], [746.15, 184.86], [776.69, 187.42], [713.71, 187.44]]}, {\"category\": \"car\", \"corners_3d\": [[658.33, 208.44], [687.07, 208.47], [693.84, 212.33], [661.98, 212.29], [658.33, 184.25], [687.07, 184.26], [693.84, 185.5], [661.98, 185.49]]}, {\"category\": \"van\", \"corners_3d\": [[637.73, 194.13], [660.92, 194.12], [667.13, 196.14], [641.74, 196.16], [637.73, 166.44], [660.92, 166.45], [667.13, 165.84], [641.74, 165.83]]}, {\"category\": \"car\", \"corners_3d\": [[590.46, 207.91], [621.88, 207.84], [626.85, 211.47], [592.17, 211.56], [590.46, 179.26], [621.88, 179.24], [626.85, 179.91], [592.17, 179.92]]}, {\"category\": \"car\", \"corners_3d\": [[-197.03, 351.56], [-383.49, 353.78], [-52.49, 291.16], [67.29, 290.21], [-197.03, 203.67], [-383.49, 204.05], [-52.49, 193.26], [67.29, 193.09]]}, {\"category\": \"car\", \"corners_3d\": [[11.8, 251.26], [-63.75, 251.19], [85.89, 233.95], [144.91, 233.99], [11.8, 186.57], [-63.75, 186.56], [85.89, 183.54], [144.91, 183.55]]}, {\"category\": \"van\", \"corners_3d\": [[279.53, 239.11], [207.9, 239.24], [287.67, 225.55], [344.41, 225.46], [279.53, 150.47], [207.9, 150.43], [287.67, 155.05], [344.41, 155.08]]}]\n```" - } - ] - }, - { - "id": 2442, - "data_source": "Mono3DRefer", - "scene_name": "002671", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[989.82, 316.13], [962.98, 306.49], [1052.72, 306.67], [1086.48, 316.34], [989.82, 143.31], [962.98, 145.95], [1052.72, 145.9], [1086.48, 143.25]]}]\n```" - } - ] - }, - { - "id": 2443, - "data_source": "Mono3DRefer", - "scene_name": "002684", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[489.59, 279.22], [394.08, 279.51], [443.43, 253.25], [515.38, 253.08], [489.59, 198.0], [394.08, 198.07], [443.43, 191.86], [515.38, 191.82]]}, {\"category\": \"car\", \"corners_3d\": [[582.47, 191.87], [559.19, 191.89], [561.75, 190.51], [583.35, 190.5], [582.47, 167.38], [559.19, 167.37], [561.75, 167.77], [583.35, 167.77]]}]\n```" - } - ] - }, - { - "id": 2444, - "data_source": "Mono3DRefer", - "scene_name": "002686", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (motorcyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"motorcyclist\", \"corners_3d\": [[661.01, 274.58], [608.39, 272.1], [642.49, 258.43], [688.45, 260.27], [661.01, 141.03], [608.39, 141.8], [642.49, 146.08], [688.45, 145.5]]}]\n```" - } - ] - }, - { - "id": 2445, - "data_source": "Mono3DRefer", - "scene_name": "002686", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[905.35, 137.14], [877.12, 137.83], [890.83, 140.44], [917.18, 139.85], [905.35, 112.1], [877.12, 113.29], [890.83, 117.73], [917.18, 116.72]]}]\n```" - } - ] - }, - { - "id": 2446, - "data_source": "Mono3DRefer", - "scene_name": "002686", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (motorcyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"motorcyclist\", \"corners_3d\": [[661.01, 274.58], [608.39, 272.1], [642.49, 258.43], [688.45, 260.27], [661.01, 141.03], [608.39, 141.8], [642.49, 146.08], [688.45, 145.5]]}, {\"category\": \"car\", \"corners_3d\": [[905.35, 137.14], [877.12, 137.83], [890.83, 140.44], [917.18, 139.85], [905.35, 112.1], [877.12, 113.29], [890.83, 117.73], [917.18, 116.72]]}]\n```" - } - ] - }, - { - "id": 2447, - "data_source": "Mono3DRefer", - "scene_name": "002694", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[380.44, 204.47], [381.62, 203.13], [466.09, 202.23], [468.54, 203.48], [380.44, 170.81], [381.62, 170.9], [466.09, 170.96], [468.54, 170.88]]}, {\"category\": \"car\", \"corners_3d\": [[728.5, 193.11], [703.38, 193.28], [682.85, 191.69], [706.18, 191.54], [728.5, 159.69], [703.38, 159.58], [682.85, 160.62], [706.18, 160.71]]}, {\"category\": \"car\", \"corners_3d\": [[711.85, 190.33], [731.55, 190.19], [757.08, 191.57], [736.02, 191.73], [711.85, 166.43], [731.55, 166.48], [757.08, 165.98], [736.02, 165.92]]}, {\"category\": \"car\", \"corners_3d\": [[750.89, 187.95], [771.95, 187.83], [806.34, 189.32], [783.46, 189.46], [750.89, 167.19], [771.95, 167.23], [806.34, 166.67], [783.46, 166.62]]}, {\"category\": \"car\", \"corners_3d\": [[923.31, 184.65], [940.05, 184.59], [970.53, 185.35], [952.84, 185.42], [923.31, 164.37], [940.05, 164.41], [970.53, 163.86], [952.84, 163.81]]}, {\"category\": \"car\", \"corners_3d\": [[963.38, 183.73], [980.35, 183.63], [1016.77, 184.3], [999.08, 184.4], [963.38, 161.58], [980.35, 161.67], [1016.77, 160.99], [999.08, 160.88]]}]\n```" - } - ] - }, - { - "id": 2448, - "data_source": "Mono3DRefer", - "scene_name": "002694", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[947.33, 189.11], [924.54, 189.24], [886.56, 187.91], [907.78, 187.81], [947.33, 158.04], [924.54, 157.93], [886.56, 159.13], [907.78, 159.23]]}]\n```" - } - ] - }, - { - "id": 2449, - "data_source": "Mono3DRefer", - "scene_name": "002694", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[736.7, 193.46], [734.04, 193.2], [750.55, 193.14], [753.41, 193.4], [736.7, 163.74], [734.04, 163.85], [750.55, 163.88], [753.41, 163.77]]}]\n```" - } - ] - }, - { - "id": 2450, - "data_source": "Mono3DRefer", - "scene_name": "002694", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[380.44, 204.47], [381.62, 203.13], [466.09, 202.23], [468.54, 203.48], [380.44, 170.81], [381.62, 170.9], [466.09, 170.96], [468.54, 170.88]]}, {\"category\": \"car\", \"corners_3d\": [[728.5, 193.11], [703.38, 193.28], [682.85, 191.69], [706.18, 191.54], [728.5, 159.69], [703.38, 159.58], [682.85, 160.62], [706.18, 160.71]]}, {\"category\": \"car\", \"corners_3d\": [[711.85, 190.33], [731.55, 190.19], [757.08, 191.57], [736.02, 191.73], [711.85, 166.43], [731.55, 166.48], [757.08, 165.98], [736.02, 165.92]]}, {\"category\": \"car\", \"corners_3d\": [[750.89, 187.95], [771.95, 187.83], [806.34, 189.32], [783.46, 189.46], [750.89, 167.19], [771.95, 167.23], [806.34, 166.67], [783.46, 166.62]]}, {\"category\": \"van\", \"corners_3d\": [[947.33, 189.11], [924.54, 189.24], [886.56, 187.91], [907.78, 187.81], [947.33, 158.04], [924.54, 157.93], [886.56, 159.13], [907.78, 159.23]]}, {\"category\": \"car\", \"corners_3d\": [[923.31, 184.65], [940.05, 184.59], [970.53, 185.35], [952.84, 185.42], [923.31, 164.37], [940.05, 164.41], [970.53, 163.86], [952.84, 163.81]]}, {\"category\": \"car\", \"corners_3d\": [[963.38, 183.73], [980.35, 183.63], [1016.77, 184.3], [999.08, 184.4], [963.38, 161.58], [980.35, 161.67], [1016.77, 160.99], [999.08, 160.88]]}]\n```" - } - ] - }, - { - "id": 2451, - "data_source": "Mono3DRefer", - "scene_name": "002694", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[380.44, 204.47], [381.62, 203.13], [466.09, 202.23], [468.54, 203.48], [380.44, 170.81], [381.62, 170.9], [466.09, 170.96], [468.54, 170.88]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[736.7, 193.46], [734.04, 193.2], [750.55, 193.14], [753.41, 193.4], [736.7, 163.74], [734.04, 163.85], [750.55, 163.88], [753.41, 163.77]]}, {\"category\": \"car\", \"corners_3d\": [[728.5, 193.11], [703.38, 193.28], [682.85, 191.69], [706.18, 191.54], [728.5, 159.69], [703.38, 159.58], [682.85, 160.62], [706.18, 160.71]]}, {\"category\": \"car\", \"corners_3d\": [[711.85, 190.33], [731.55, 190.19], [757.08, 191.57], [736.02, 191.73], [711.85, 166.43], [731.55, 166.48], [757.08, 165.98], [736.02, 165.92]]}, {\"category\": \"car\", \"corners_3d\": [[750.89, 187.95], [771.95, 187.83], [806.34, 189.32], [783.46, 189.46], [750.89, 167.19], [771.95, 167.23], [806.34, 166.67], [783.46, 166.62]]}, {\"category\": \"car\", \"corners_3d\": [[923.31, 184.65], [940.05, 184.59], [970.53, 185.35], [952.84, 185.42], [923.31, 164.37], [940.05, 164.41], [970.53, 163.86], [952.84, 163.81]]}, {\"category\": \"car\", \"corners_3d\": [[963.38, 183.73], [980.35, 183.63], [1016.77, 184.3], [999.08, 184.4], [963.38, 161.58], [980.35, 161.67], [1016.77, 160.99], [999.08, 160.88]]}]\n```" - } - ] - }, - { - "id": 2452, - "data_source": "Mono3DRefer", - "scene_name": "002694", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[736.7, 193.46], [734.04, 193.2], [750.55, 193.14], [753.41, 193.4], [736.7, 163.74], [734.04, 163.85], [750.55, 163.88], [753.41, 163.77]]}, {\"category\": \"van\", \"corners_3d\": [[947.33, 189.11], [924.54, 189.24], [886.56, 187.91], [907.78, 187.81], [947.33, 158.04], [924.54, 157.93], [886.56, 159.13], [907.78, 159.23]]}]\n```" - } - ] - }, - { - "id": 2453, - "data_source": "Mono3DRefer", - "scene_name": "002694", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[380.44, 204.47], [381.62, 203.13], [466.09, 202.23], [468.54, 203.48], [380.44, 170.81], [381.62, 170.9], [466.09, 170.96], [468.54, 170.88]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[736.7, 193.46], [734.04, 193.2], [750.55, 193.14], [753.41, 193.4], [736.7, 163.74], [734.04, 163.85], [750.55, 163.88], [753.41, 163.77]]}, {\"category\": \"car\", \"corners_3d\": [[728.5, 193.11], [703.38, 193.28], [682.85, 191.69], [706.18, 191.54], [728.5, 159.69], [703.38, 159.58], [682.85, 160.62], [706.18, 160.71]]}, {\"category\": \"car\", \"corners_3d\": [[711.85, 190.33], [731.55, 190.19], [757.08, 191.57], [736.02, 191.73], [711.85, 166.43], [731.55, 166.48], [757.08, 165.98], [736.02, 165.92]]}, {\"category\": \"car\", \"corners_3d\": [[750.89, 187.95], [771.95, 187.83], [806.34, 189.32], [783.46, 189.46], [750.89, 167.19], [771.95, 167.23], [806.34, 166.67], [783.46, 166.62]]}, {\"category\": \"van\", \"corners_3d\": [[947.33, 189.11], [924.54, 189.24], [886.56, 187.91], [907.78, 187.81], [947.33, 158.04], [924.54, 157.93], [886.56, 159.13], [907.78, 159.23]]}, {\"category\": \"car\", \"corners_3d\": [[923.31, 184.65], [940.05, 184.59], [970.53, 185.35], [952.84, 185.42], [923.31, 164.37], [940.05, 164.41], [970.53, 163.86], [952.84, 163.81]]}, {\"category\": \"car\", \"corners_3d\": [[963.38, 183.73], [980.35, 183.63], [1016.77, 184.3], [999.08, 184.4], [963.38, 161.58], [980.35, 161.67], [1016.77, 160.99], [999.08, 160.88]]}]\n```" - } - ] - }, - { - "id": 2454, - "data_source": "Mono3DRefer", - "scene_name": "002696", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[605.41, 226.8], [617.45, 226.12], [625.31, 226.87], [613.2, 227.57], [605.41, 175.59], [617.45, 175.55], [625.31, 175.59], [613.2, 175.63]]}]\n```" - } - ] - }, - { - "id": 2455, - "data_source": "Mono3DRefer", - "scene_name": "002696", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[276.7, 390.9], [82.76, 391.53], [284.1, 306.7], [402.58, 306.47], [276.7, 200.8], [82.76, 200.88], [284.1, 190.01], [402.58, 189.98]]}, {\"category\": \"car\", \"corners_3d\": [[617.28, 222.24], [660.93, 222.21], [669.25, 229.29], [619.34, 229.32], [617.28, 180.55], [660.93, 180.55], [669.25, 181.65], [619.34, 181.66]]}]\n```" - } - ] - }, - { - "id": 2456, - "data_source": "Mono3DRefer", - "scene_name": "002696", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[510.0, 235.79], [442.36, 235.44], [492.04, 220.63], [543.82, 220.84], [510.0, 131.94], [442.36, 132.17], [492.04, 141.8], [543.82, 141.66]]}]\n```" - } - ] - }, - { - "id": 2457, - "data_source": "Mono3DRefer", - "scene_name": "002696", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[276.7, 390.9], [82.76, 391.53], [284.1, 306.7], [402.58, 306.47], [276.7, 200.8], [82.76, 200.88], [284.1, 190.01], [402.58, 189.98]]}, {\"category\": \"car\", \"corners_3d\": [[617.28, 222.24], [660.93, 222.21], [669.25, 229.29], [619.34, 229.32], [617.28, 180.55], [660.93, 180.55], [669.25, 181.65], [619.34, 181.66]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[605.41, 226.8], [617.45, 226.12], [625.31, 226.87], [613.2, 227.57], [605.41, 175.59], [617.45, 175.55], [625.31, 175.59], [613.2, 175.63]]}]\n```" - } - ] - }, - { - "id": 2458, - "data_source": "Mono3DRefer", - "scene_name": "002696", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[510.0, 235.79], [442.36, 235.44], [492.04, 220.63], [543.82, 220.84], [510.0, 131.94], [442.36, 132.17], [492.04, 141.8], [543.82, 141.66]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[605.41, 226.8], [617.45, 226.12], [625.31, 226.87], [613.2, 227.57], [605.41, 175.59], [617.45, 175.55], [625.31, 175.59], [613.2, 175.63]]}]\n```" - } - ] - }, - { - "id": 2459, - "data_source": "Mono3DRefer", - "scene_name": "002696", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[276.7, 390.9], [82.76, 391.53], [284.1, 306.7], [402.58, 306.47], [276.7, 200.8], [82.76, 200.88], [284.1, 190.01], [402.58, 189.98]]}, {\"category\": \"truck\", \"corners_3d\": [[510.0, 235.79], [442.36, 235.44], [492.04, 220.63], [543.82, 220.84], [510.0, 131.94], [442.36, 132.17], [492.04, 141.8], [543.82, 141.66]]}, {\"category\": \"car\", \"corners_3d\": [[617.28, 222.24], [660.93, 222.21], [669.25, 229.29], [619.34, 229.32], [617.28, 180.55], [660.93, 180.55], [669.25, 181.65], [619.34, 181.66]]}]\n```" - } - ] - }, - { - "id": 2460, - "data_source": "Mono3DRefer", - "scene_name": "002696", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[276.7, 390.9], [82.76, 391.53], [284.1, 306.7], [402.58, 306.47], [276.7, 200.8], [82.76, 200.88], [284.1, 190.01], [402.58, 189.98]]}, {\"category\": \"truck\", \"corners_3d\": [[510.0, 235.79], [442.36, 235.44], [492.04, 220.63], [543.82, 220.84], [510.0, 131.94], [442.36, 132.17], [492.04, 141.8], [543.82, 141.66]]}, {\"category\": \"car\", \"corners_3d\": [[617.28, 222.24], [660.93, 222.21], [669.25, 229.29], [619.34, 229.32], [617.28, 180.55], [660.93, 180.55], [669.25, 181.65], [619.34, 181.66]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[605.41, 226.8], [617.45, 226.12], [625.31, 226.87], [613.2, 227.57], [605.41, 175.59], [617.45, 175.55], [625.31, 175.59], [613.2, 175.63]]}]\n```" - } - ] - }, - { - "id": 2461, - "data_source": "Mono3DRefer", - "scene_name": "002699", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[349.14, 391.01], [81.56, 387.1], [337.55, 292.14], [482.88, 293.26], [349.14, 103.69], [81.56, 105.14], [337.55, 140.42], [482.88, 140.0]]}]\n```" - } - ] - }, - { - "id": 2462, - "data_source": "Mono3DRefer", - "scene_name": "002699", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[516.56, 265.34], [408.61, 264.97], [471.58, 240.27], [547.68, 240.45], [516.56, 138.25], [408.61, 138.44], [471.58, 151.19], [547.68, 151.09]]}]\n```" - } - ] - }, - { - "id": 2463, - "data_source": "Mono3DRefer", - "scene_name": "002699", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[963.08, 535.08], [1140.39, 517.68], [1460.94, 678.49], [1217.32, 717.48], [963.08, 117.88], [1140.39, 121.01], [1460.94, 92.07], [1217.32, 85.05]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[298.93, 470.46], [179.23, 449.0], [308.94, 398.78], [413.05, 412.73], [298.93, 140.71], [179.23, 143.74], [308.94, 150.83], [413.05, 148.86]]}]\n```" - } - ] - }, - { - "id": 2464, - "data_source": "Mono3DRefer", - "scene_name": "002699", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[349.14, 391.01], [81.56, 387.1], [337.55, 292.14], [482.88, 293.26], [349.14, 103.69], [81.56, 105.14], [337.55, 140.42], [482.88, 140.0]]}, {\"category\": \"van\", \"corners_3d\": [[516.56, 265.34], [408.61, 264.97], [471.58, 240.27], [547.68, 240.45], [516.56, 138.25], [408.61, 138.44], [471.58, 151.19], [547.68, 151.09]]}]\n```" - } - ] - }, - { - "id": 2465, - "data_source": "Mono3DRefer", - "scene_name": "002699", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[963.08, 535.08], [1140.39, 517.68], [1460.94, 678.49], [1217.32, 717.48], [963.08, 117.88], [1140.39, 121.01], [1460.94, 92.07], [1217.32, 85.05]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[298.93, 470.46], [179.23, 449.0], [308.94, 398.78], [413.05, 412.73], [298.93, 140.71], [179.23, 143.74], [308.94, 150.83], [413.05, 148.86]]}, {\"category\": \"car\", \"corners_3d\": [[349.14, 391.01], [81.56, 387.1], [337.55, 292.14], [482.88, 293.26], [349.14, 103.69], [81.56, 105.14], [337.55, 140.42], [482.88, 140.0]]}]\n```" - } - ] - }, - { - "id": 2466, - "data_source": "Mono3DRefer", - "scene_name": "002699", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[963.08, 535.08], [1140.39, 517.68], [1460.94, 678.49], [1217.32, 717.48], [963.08, 117.88], [1140.39, 121.01], [1460.94, 92.07], [1217.32, 85.05]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[298.93, 470.46], [179.23, 449.0], [308.94, 398.78], [413.05, 412.73], [298.93, 140.71], [179.23, 143.74], [308.94, 150.83], [413.05, 148.86]]}, {\"category\": \"van\", \"corners_3d\": [[516.56, 265.34], [408.61, 264.97], [471.58, 240.27], [547.68, 240.45], [516.56, 138.25], [408.61, 138.44], [471.58, 151.19], [547.68, 151.09]]}]\n```" - } - ] - }, - { - "id": 2467, - "data_source": "Mono3DRefer", - "scene_name": "002699", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[963.08, 535.08], [1140.39, 517.68], [1460.94, 678.49], [1217.32, 717.48], [963.08, 117.88], [1140.39, 121.01], [1460.94, 92.07], [1217.32, 85.05]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[298.93, 470.46], [179.23, 449.0], [308.94, 398.78], [413.05, 412.73], [298.93, 140.71], [179.23, 143.74], [308.94, 150.83], [413.05, 148.86]]}, {\"category\": \"car\", \"corners_3d\": [[349.14, 391.01], [81.56, 387.1], [337.55, 292.14], [482.88, 293.26], [349.14, 103.69], [81.56, 105.14], [337.55, 140.42], [482.88, 140.0]]}, {\"category\": \"van\", \"corners_3d\": [[516.56, 265.34], [408.61, 264.97], [471.58, 240.27], [547.68, 240.45], [516.56, 138.25], [408.61, 138.44], [471.58, 151.19], [547.68, 151.09]]}]\n```" - } - ] - }, - { - "id": 2468, - "data_source": "Mono3DRefer", - "scene_name": "002709", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-33.45, 749.39], [-672.4, 734.25], [199.64, 361.69], [411.79, 363.27], [-33.45, 52.15], [-672.4, 55.6], [199.64, 140.47], [411.79, 140.11]]}]\n```" - } - ] - }, - { - "id": 2469, - "data_source": "Mono3DRefer", - "scene_name": "002709", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[469.47, 328.22], [321.52, 327.94], [425.76, 274.47], [519.75, 274.59], [469.47, 154.84], [321.52, 154.89], [425.76, 164.61], [519.75, 164.59]]}]\n```" - } - ] - }, - { - "id": 2470, - "data_source": "Mono3DRefer", - "scene_name": "002709", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[755.05, 227.86], [778.51, 227.89], [784.78, 229.67], [760.42, 229.64], [755.05, 176.74], [778.51, 176.74], [784.78, 176.55], [760.42, 176.56]]}]\n```" - } - ] - }, - { - "id": 2471, - "data_source": "Mono3DRefer", - "scene_name": "002709", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-33.45, 749.39], [-672.4, 734.25], [199.64, 361.69], [411.79, 363.27], [-33.45, 52.15], [-672.4, 55.6], [199.64, 140.47], [411.79, 140.11]]}, {\"category\": \"van\", \"corners_3d\": [[469.47, 328.22], [321.52, 327.94], [425.76, 274.47], [519.75, 274.59], [469.47, 154.84], [321.52, 154.89], [425.76, 164.61], [519.75, 164.59]]}]\n```" - } - ] - }, - { - "id": 2472, - "data_source": "Mono3DRefer", - "scene_name": "002709", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-33.45, 749.39], [-672.4, 734.25], [199.64, 361.69], [411.79, 363.27], [-33.45, 52.15], [-672.4, 55.6], [199.64, 140.47], [411.79, 140.11]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[755.05, 227.86], [778.51, 227.89], [784.78, 229.67], [760.42, 229.64], [755.05, 176.74], [778.51, 176.74], [784.78, 176.55], [760.42, 176.56]]}]\n```" - } - ] - }, - { - "id": 2473, - "data_source": "Mono3DRefer", - "scene_name": "002709", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[469.47, 328.22], [321.52, 327.94], [425.76, 274.47], [519.75, 274.59], [469.47, 154.84], [321.52, 154.89], [425.76, 164.61], [519.75, 164.59]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[755.05, 227.86], [778.51, 227.89], [784.78, 229.67], [760.42, 229.64], [755.05, 176.74], [778.51, 176.74], [784.78, 176.55], [760.42, 176.56]]}]\n```" - } - ] - }, - { - "id": 2474, - "data_source": "Mono3DRefer", - "scene_name": "002709", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-33.45, 749.39], [-672.4, 734.25], [199.64, 361.69], [411.79, 363.27], [-33.45, 52.15], [-672.4, 55.6], [199.64, 140.47], [411.79, 140.11]]}, {\"category\": \"van\", \"corners_3d\": [[469.47, 328.22], [321.52, 327.94], [425.76, 274.47], [519.75, 274.59], [469.47, 154.84], [321.52, 154.89], [425.76, 164.61], [519.75, 164.59]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[755.05, 227.86], [778.51, 227.89], [784.78, 229.67], [760.42, 229.64], [755.05, 176.74], [778.51, 176.74], [784.78, 176.55], [760.42, 176.56]]}]\n```" - } - ] - }, - { - "id": 2475, - "data_source": "Mono3DRefer", - "scene_name": "002712", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[736.37, 257.7], [826.88, 257.71], [916.99, 292.99], [788.85, 292.97], [736.37, 179.02], [826.88, 179.02], [916.99, 181.58], [788.85, 181.58]]}, {\"category\": \"car\", \"corners_3d\": [[698.66, 232.28], [760.0, 232.63], [780.28, 244.85], [706.36, 244.33], [698.66, 177.86], [760.0, 177.89], [780.28, 178.92], [706.36, 178.87]]}, {\"category\": \"car\", \"corners_3d\": [[417.25, 251.31], [360.44, 251.19], [400.86, 239.15], [448.98, 239.23], [417.25, 195.21], [360.44, 195.17], [400.86, 191.74], [448.98, 191.77]]}, {\"category\": \"car\", \"corners_3d\": [[516.76, 213.21], [490.6, 213.23], [499.05, 210.16], [523.22, 210.15], [516.76, 189.13], [490.6, 189.14], [499.05, 187.9], [523.22, 187.89]]}, {\"category\": \"car\", \"corners_3d\": [[638.42, 200.89], [663.84, 200.89], [668.27, 203.2], [640.75, 203.2], [638.42, 178.95], [663.84, 178.95], [668.27, 179.45], [640.75, 179.45]]}, {\"category\": \"car\", \"corners_3d\": [[526.71, 209.8], [503.97, 209.79], [512.67, 206.92], [533.65, 206.93], [526.71, 189.29], [503.97, 189.28], [512.67, 188.01], [533.65, 188.01]]}, {\"category\": \"car\", \"corners_3d\": [[830.19, 337.18], [1031.66, 337.22], [2256.24, 814.73], [1468.91, 814.17], [830.19, 172.82], [1031.66, 172.82], [2256.24, 172.7], [1468.91, 172.7]]}]\n```" - } - ] - }, - { - "id": 2476, - "data_source": "Mono3DRefer", - "scene_name": "002713", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[595.7, 201.68], [624.03, 201.71], [623.98, 204.41], [593.0, 204.38], [595.7, 173.7], [624.03, 173.7], [623.98, 173.78], [593.0, 173.77]]}]\n```" - } - ] - }, - { - "id": 2477, - "data_source": "Mono3DRefer", - "scene_name": "002713", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[272.76, 227.38], [272.6, 227.97], [224.27, 229.34], [224.96, 228.72], [272.76, 182.7], [272.6, 182.8], [224.27, 183.05], [224.96, 182.94]]}]\n```" - } - ] - }, - { - "id": 2478, - "data_source": "Mono3DRefer", - "scene_name": "002713", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[206.48, 231.18], [199.57, 232.2], [188.02, 232.21], [195.13, 231.19], [206.48, 186.33], [199.57, 186.56], [188.02, 186.57], [195.13, 186.33]]}]\n```" - } - ] - }, - { - "id": 2479, - "data_source": "Mono3DRefer", - "scene_name": "002713", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[595.7, 201.68], [624.03, 201.71], [623.98, 204.41], [593.0, 204.38], [595.7, 173.7], [624.03, 173.7], [623.98, 173.78], [593.0, 173.77]]}, {\"category\": \"cyclist\", \"corners_3d\": [[272.76, 227.38], [272.6, 227.97], [224.27, 229.34], [224.96, 228.72], [272.76, 182.7], [272.6, 182.8], [224.27, 183.05], [224.96, 182.94]]}]\n```" - } - ] - }, - { - "id": 2480, - "data_source": "Mono3DRefer", - "scene_name": "002713", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[595.7, 201.68], [624.03, 201.71], [623.98, 204.41], [593.0, 204.38], [595.7, 173.7], [624.03, 173.7], [623.98, 173.78], [593.0, 173.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[206.48, 231.18], [199.57, 232.2], [188.02, 232.21], [195.13, 231.19], [206.48, 186.33], [199.57, 186.56], [188.02, 186.57], [195.13, 186.33]]}]\n```" - } - ] - }, - { - "id": 2481, - "data_source": "Mono3DRefer", - "scene_name": "002713", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[206.48, 231.18], [199.57, 232.2], [188.02, 232.21], [195.13, 231.19], [206.48, 186.33], [199.57, 186.56], [188.02, 186.57], [195.13, 186.33]]}, {\"category\": \"cyclist\", \"corners_3d\": [[272.76, 227.38], [272.6, 227.97], [224.27, 229.34], [224.96, 228.72], [272.76, 182.7], [272.6, 182.8], [224.27, 183.05], [224.96, 182.94]]}]\n```" - } - ] - }, - { - "id": 2482, - "data_source": "Mono3DRefer", - "scene_name": "002713", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[595.7, 201.68], [624.03, 201.71], [623.98, 204.41], [593.0, 204.38], [595.7, 173.7], [624.03, 173.7], [623.98, 173.78], [593.0, 173.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[206.48, 231.18], [199.57, 232.2], [188.02, 232.21], [195.13, 231.19], [206.48, 186.33], [199.57, 186.56], [188.02, 186.57], [195.13, 186.33]]}, {\"category\": \"cyclist\", \"corners_3d\": [[272.76, 227.38], [272.6, 227.97], [224.27, 229.34], [224.96, 228.72], [272.76, 182.7], [272.6, 182.8], [224.27, 183.05], [224.96, 182.94]]}]\n```" - } - ] - }, - { - "id": 2483, - "data_source": "Mono3DRefer", - "scene_name": "002715", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[749.13, 211.94], [732.82, 211.95], [730.19, 211.21], [746.2, 211.19], [749.13, 173.94], [732.82, 173.94], [730.19, 173.92], [746.2, 173.92]]}]\n```" - } - ] - }, - { - "id": 2484, - "data_source": "Mono3DRefer", - "scene_name": "002715", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[351.62, 212.24], [317.06, 212.17], [347.1, 208.49], [378.47, 208.55], [351.62, 181.73], [317.06, 181.71], [347.1, 180.88], [378.47, 180.9]]}, {\"category\": \"car\", \"corners_3d\": [[-9.46, 269.54], [-96.39, 269.78], [44.74, 249.98], [113.67, 249.83], [-9.46, 192.94], [-96.39, 192.99], [44.74, 188.88], [113.67, 188.85]]}]\n```" - } - ] - }, - { - "id": 2485, - "data_source": "Mono3DRefer", - "scene_name": "002715", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[351.62, 212.24], [317.06, 212.17], [347.1, 208.49], [378.47, 208.55], [351.62, 181.73], [317.06, 181.71], [347.1, 180.88], [378.47, 180.9]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[749.13, 211.94], [732.82, 211.95], [730.19, 211.21], [746.2, 211.19], [749.13, 173.94], [732.82, 173.94], [730.19, 173.92], [746.2, 173.92]]}, {\"category\": \"car\", \"corners_3d\": [[-9.46, 269.54], [-96.39, 269.78], [44.74, 249.98], [113.67, 249.83], [-9.46, 192.94], [-96.39, 192.99], [44.74, 188.88], [113.67, 188.85]]}]\n```" - } - ] - }, - { - "id": 2486, - "data_source": "Mono3DRefer", - "scene_name": "002718", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[272.57, 292.48], [250.07, 299.63], [192.3, 299.33], [218.21, 292.22], [272.57, 159.16], [250.07, 157.8], [192.3, 157.86], [218.21, 159.21]]}]\n```" - } - ] - }, - { - "id": 2487, - "data_source": "Mono3DRefer", - "scene_name": "002727", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[868.93, 270.23], [890.5, 270.2], [924.22, 281.62], [900.14, 281.65], [868.93, 183.65], [890.5, 183.65], [924.22, 184.92], [900.14, 184.92]]}]\n```" - } - ] - }, - { - "id": 2488, - "data_source": "Mono3DRefer", - "scene_name": "002727", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[665.78, 227.17], [719.2, 227.05], [749.68, 239.69], [683.85, 239.87], [665.78, 170.9], [719.2, 170.91], [749.68, 170.45], [683.85, 170.45]]}, {\"category\": \"van\", \"corners_3d\": [[648.21, 205.98], [682.39, 205.93], [695.72, 210.63], [656.7, 210.69], [648.21, 167.0], [682.39, 167.01], [695.72, 166.18], [656.7, 166.17]]}]\n```" - } - ] - }, - { - "id": 2489, - "data_source": "Mono3DRefer", - "scene_name": "002727", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[453.85, 316.74], [316.78, 317.58], [393.08, 276.76], [491.34, 276.33], [453.85, 193.14], [316.78, 193.26], [393.08, 187.5], [491.34, 187.44]]}, {\"category\": \"car\", \"corners_3d\": [[715.26, 273.58], [815.74, 273.45], [902.83, 314.62], [761.32, 314.88], [715.26, 176.39], [815.74, 176.38], [902.83, 177.83], [761.32, 177.84]]}, {\"category\": \"car\", \"corners_3d\": [[686.65, 241.61], [751.7, 241.49], [791.39, 258.95], [709.84, 259.14], [686.65, 182.35], [751.7, 182.34], [791.39, 184.75], [709.84, 184.78]]}, {\"category\": \"car\", \"corners_3d\": [[564.53, 210.43], [527.09, 210.48], [535.18, 205.96], [568.13, 205.93], [564.53, 175.96], [527.09, 175.96], [535.18, 175.59], [568.13, 175.59]]}]\n```" - } - ] - }, - { - "id": 2490, - "data_source": "Mono3DRefer", - "scene_name": "002727", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[868.93, 270.23], [890.5, 270.2], [924.22, 281.62], [900.14, 281.65], [868.93, 183.65], [890.5, 183.65], [924.22, 184.92], [900.14, 184.92]]}, {\"category\": \"van\", \"corners_3d\": [[665.78, 227.17], [719.2, 227.05], [749.68, 239.69], [683.85, 239.87], [665.78, 170.9], [719.2, 170.91], [749.68, 170.45], [683.85, 170.45]]}, {\"category\": \"van\", \"corners_3d\": [[648.21, 205.98], [682.39, 205.93], [695.72, 210.63], [656.7, 210.69], [648.21, 167.0], [682.39, 167.01], [695.72, 166.18], [656.7, 166.17]]}]\n```" - } - ] - }, - { - "id": 2491, - "data_source": "Mono3DRefer", - "scene_name": "002727", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[868.93, 270.23], [890.5, 270.2], [924.22, 281.62], [900.14, 281.65], [868.93, 183.65], [890.5, 183.65], [924.22, 184.92], [900.14, 184.92]]}, {\"category\": \"car\", \"corners_3d\": [[453.85, 316.74], [316.78, 317.58], [393.08, 276.76], [491.34, 276.33], [453.85, 193.14], [316.78, 193.26], [393.08, 187.5], [491.34, 187.44]]}, {\"category\": \"car\", \"corners_3d\": [[715.26, 273.58], [815.74, 273.45], [902.83, 314.62], [761.32, 314.88], [715.26, 176.39], [815.74, 176.38], [902.83, 177.83], [761.32, 177.84]]}, {\"category\": \"car\", \"corners_3d\": [[686.65, 241.61], [751.7, 241.49], [791.39, 258.95], [709.84, 259.14], [686.65, 182.35], [751.7, 182.34], [791.39, 184.75], [709.84, 184.78]]}, {\"category\": \"car\", \"corners_3d\": [[564.53, 210.43], [527.09, 210.48], [535.18, 205.96], [568.13, 205.93], [564.53, 175.96], [527.09, 175.96], [535.18, 175.59], [568.13, 175.59]]}]\n```" - } - ] - }, - { - "id": 2492, - "data_source": "Mono3DRefer", - "scene_name": "002727", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[453.85, 316.74], [316.78, 317.58], [393.08, 276.76], [491.34, 276.33], [453.85, 193.14], [316.78, 193.26], [393.08, 187.5], [491.34, 187.44]]}, {\"category\": \"car\", \"corners_3d\": [[715.26, 273.58], [815.74, 273.45], [902.83, 314.62], [761.32, 314.88], [715.26, 176.39], [815.74, 176.38], [902.83, 177.83], [761.32, 177.84]]}, {\"category\": \"car\", \"corners_3d\": [[686.65, 241.61], [751.7, 241.49], [791.39, 258.95], [709.84, 259.14], [686.65, 182.35], [751.7, 182.34], [791.39, 184.75], [709.84, 184.78]]}, {\"category\": \"van\", \"corners_3d\": [[665.78, 227.17], [719.2, 227.05], [749.68, 239.69], [683.85, 239.87], [665.78, 170.9], [719.2, 170.91], [749.68, 170.45], [683.85, 170.45]]}, {\"category\": \"car\", \"corners_3d\": [[564.53, 210.43], [527.09, 210.48], [535.18, 205.96], [568.13, 205.93], [564.53, 175.96], [527.09, 175.96], [535.18, 175.59], [568.13, 175.59]]}, {\"category\": \"van\", \"corners_3d\": [[648.21, 205.98], [682.39, 205.93], [695.72, 210.63], [656.7, 210.69], [648.21, 167.0], [682.39, 167.01], [695.72, 166.18], [656.7, 166.17]]}]\n```" - } - ] - }, - { - "id": 2493, - "data_source": "Mono3DRefer", - "scene_name": "002727", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[868.93, 270.23], [890.5, 270.2], [924.22, 281.62], [900.14, 281.65], [868.93, 183.65], [890.5, 183.65], [924.22, 184.92], [900.14, 184.92]]}, {\"category\": \"car\", \"corners_3d\": [[453.85, 316.74], [316.78, 317.58], [393.08, 276.76], [491.34, 276.33], [453.85, 193.14], [316.78, 193.26], [393.08, 187.5], [491.34, 187.44]]}, {\"category\": \"car\", \"corners_3d\": [[715.26, 273.58], [815.74, 273.45], [902.83, 314.62], [761.32, 314.88], [715.26, 176.39], [815.74, 176.38], [902.83, 177.83], [761.32, 177.84]]}, {\"category\": \"car\", \"corners_3d\": [[686.65, 241.61], [751.7, 241.49], [791.39, 258.95], [709.84, 259.14], [686.65, 182.35], [751.7, 182.34], [791.39, 184.75], [709.84, 184.78]]}, {\"category\": \"van\", \"corners_3d\": [[665.78, 227.17], [719.2, 227.05], [749.68, 239.69], [683.85, 239.87], [665.78, 170.9], [719.2, 170.91], [749.68, 170.45], [683.85, 170.45]]}, {\"category\": \"car\", \"corners_3d\": [[564.53, 210.43], [527.09, 210.48], [535.18, 205.96], [568.13, 205.93], [564.53, 175.96], [527.09, 175.96], [535.18, 175.59], [568.13, 175.59]]}, {\"category\": \"van\", \"corners_3d\": [[648.21, 205.98], [682.39, 205.93], [695.72, 210.63], [656.7, 210.69], [648.21, 167.0], [682.39, 167.01], [695.72, 166.18], [656.7, 166.17]]}]\n```" - } - ] - }, - { - "id": 2494, - "data_source": "Mono3DRefer", - "scene_name": "002730", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-323.49, 370.42], [-132.72, 329.34], [216.52, 327.74], [118.7, 367.87], [-323.49, 206.52], [-132.72, 199.52], [216.52, 199.25], [118.7, 206.09]]}, {\"category\": \"car\", \"corners_3d\": [[663.28, 209.79], [697.57, 209.81], [708.84, 215.0], [669.73, 214.98], [663.28, 180.53], [697.57, 180.53], [708.84, 181.61], [669.73, 181.6]]}, {\"category\": \"car\", \"corners_3d\": [[526.4, 216.37], [520.65, 219.37], [397.69, 219.37], [411.38, 216.37], [526.4, 179.06], [520.65, 179.49], [397.69, 179.49], [411.38, 179.06]]}, {\"category\": \"car\", \"corners_3d\": [[493.9, 227.61], [481.22, 232.18], [333.43, 231.57], [357.27, 227.09], [493.9, 181.2], [481.22, 181.89], [333.43, 181.8], [357.27, 181.12]]}, {\"category\": \"car\", \"corners_3d\": [[487.41, 235.03], [469.16, 241.08], [310.49, 239.71], [342.23, 233.88], [487.41, 180.35], [469.16, 181.08], [310.49, 180.91], [342.23, 180.21]]}, {\"category\": \"car\", \"corners_3d\": [[468.7, 244.45], [449.55, 250.88], [282.25, 249.03], [314.46, 242.89], [468.7, 183.07], [449.55, 183.99], [282.25, 183.72], [314.46, 182.85]]}, {\"category\": \"car\", \"corners_3d\": [[426.1, 252.62], [403.48, 262.08], [216.63, 261.85], [258.96, 252.44], [426.1, 181.24], [403.48, 182.23], [216.63, 182.21], [258.96, 181.22]]}, {\"category\": \"car\", \"corners_3d\": [[-68.93, 313.81], [31.58, 292.96], [270.25, 293.03], [211.13, 313.89], [-68.93, 186.92], [31.58, 184.84], [270.25, 184.85], [211.13, 186.93]]}]\n```" - } - ] - }, - { - "id": 2495, - "data_source": "Mono3DRefer", - "scene_name": "002730", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[536.01, 212.86], [532.73, 215.38], [406.14, 215.6], [416.99, 213.05], [536.01, 165.46], [532.73, 165.0], [406.14, 164.95], [416.99, 165.43]]}]\n```" - } - ] - }, - { - "id": 2496, - "data_source": "Mono3DRefer", - "scene_name": "002730", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[643.63, 198.44], [678.22, 198.44], [687.98, 202.11], [648.43, 202.11], [643.63, 165.24], [678.22, 165.24], [687.98, 164.15], [648.43, 164.15]]}]\n```" - } - ] - }, - { - "id": 2497, - "data_source": "Mono3DRefer", - "scene_name": "002730", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-323.49, 370.42], [-132.72, 329.34], [216.52, 327.74], [118.7, 367.87], [-323.49, 206.52], [-132.72, 199.52], [216.52, 199.25], [118.7, 206.09]]}, {\"category\": \"car\", \"corners_3d\": [[663.28, 209.79], [697.57, 209.81], [708.84, 215.0], [669.73, 214.98], [663.28, 180.53], [697.57, 180.53], [708.84, 181.61], [669.73, 181.6]]}, {\"category\": \"van\", \"corners_3d\": [[536.01, 212.86], [532.73, 215.38], [406.14, 215.6], [416.99, 213.05], [536.01, 165.46], [532.73, 165.0], [406.14, 164.95], [416.99, 165.43]]}, {\"category\": \"car\", \"corners_3d\": [[526.4, 216.37], [520.65, 219.37], [397.69, 219.37], [411.38, 216.37], [526.4, 179.06], [520.65, 179.49], [397.69, 179.49], [411.38, 179.06]]}, {\"category\": \"car\", \"corners_3d\": [[493.9, 227.61], [481.22, 232.18], [333.43, 231.57], [357.27, 227.09], [493.9, 181.2], [481.22, 181.89], [333.43, 181.8], [357.27, 181.12]]}, {\"category\": \"car\", \"corners_3d\": [[487.41, 235.03], [469.16, 241.08], [310.49, 239.71], [342.23, 233.88], [487.41, 180.35], [469.16, 181.08], [310.49, 180.91], [342.23, 180.21]]}, {\"category\": \"car\", \"corners_3d\": [[468.7, 244.45], [449.55, 250.88], [282.25, 249.03], [314.46, 242.89], [468.7, 183.07], [449.55, 183.99], [282.25, 183.72], [314.46, 182.85]]}, {\"category\": \"car\", \"corners_3d\": [[426.1, 252.62], [403.48, 262.08], [216.63, 261.85], [258.96, 252.44], [426.1, 181.24], [403.48, 182.23], [216.63, 182.21], [258.96, 181.22]]}, {\"category\": \"car\", \"corners_3d\": [[-68.93, 313.81], [31.58, 292.96], [270.25, 293.03], [211.13, 313.89], [-68.93, 186.92], [31.58, 184.84], [270.25, 184.85], [211.13, 186.93]]}]\n```" - } - ] - }, - { - "id": 2498, - "data_source": "Mono3DRefer", - "scene_name": "002730", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-323.49, 370.42], [-132.72, 329.34], [216.52, 327.74], [118.7, 367.87], [-323.49, 206.52], [-132.72, 199.52], [216.52, 199.25], [118.7, 206.09]]}, {\"category\": \"truck\", \"corners_3d\": [[643.63, 198.44], [678.22, 198.44], [687.98, 202.11], [648.43, 202.11], [643.63, 165.24], [678.22, 165.24], [687.98, 164.15], [648.43, 164.15]]}, {\"category\": \"car\", \"corners_3d\": [[663.28, 209.79], [697.57, 209.81], [708.84, 215.0], [669.73, 214.98], [663.28, 180.53], [697.57, 180.53], [708.84, 181.61], [669.73, 181.6]]}, {\"category\": \"car\", \"corners_3d\": [[526.4, 216.37], [520.65, 219.37], [397.69, 219.37], [411.38, 216.37], [526.4, 179.06], [520.65, 179.49], [397.69, 179.49], [411.38, 179.06]]}, {\"category\": \"car\", \"corners_3d\": [[493.9, 227.61], [481.22, 232.18], [333.43, 231.57], [357.27, 227.09], [493.9, 181.2], [481.22, 181.89], [333.43, 181.8], [357.27, 181.12]]}, {\"category\": \"car\", \"corners_3d\": [[487.41, 235.03], [469.16, 241.08], [310.49, 239.71], [342.23, 233.88], [487.41, 180.35], [469.16, 181.08], [310.49, 180.91], [342.23, 180.21]]}, {\"category\": \"car\", \"corners_3d\": [[468.7, 244.45], [449.55, 250.88], [282.25, 249.03], [314.46, 242.89], [468.7, 183.07], [449.55, 183.99], [282.25, 183.72], [314.46, 182.85]]}, {\"category\": \"car\", \"corners_3d\": [[426.1, 252.62], [403.48, 262.08], [216.63, 261.85], [258.96, 252.44], [426.1, 181.24], [403.48, 182.23], [216.63, 182.21], [258.96, 181.22]]}, {\"category\": \"car\", \"corners_3d\": [[-68.93, 313.81], [31.58, 292.96], [270.25, 293.03], [211.13, 313.89], [-68.93, 186.92], [31.58, 184.84], [270.25, 184.85], [211.13, 186.93]]}]\n```" - } - ] - }, - { - "id": 2499, - "data_source": "Mono3DRefer", - "scene_name": "002730", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[643.63, 198.44], [678.22, 198.44], [687.98, 202.11], [648.43, 202.11], [643.63, 165.24], [678.22, 165.24], [687.98, 164.15], [648.43, 164.15]]}, {\"category\": \"van\", \"corners_3d\": [[536.01, 212.86], [532.73, 215.38], [406.14, 215.6], [416.99, 213.05], [536.01, 165.46], [532.73, 165.0], [406.14, 164.95], [416.99, 165.43]]}]\n```" - } - ] - }, - { - "id": 2500, - "data_source": "Mono3DRefer", - "scene_name": "002730", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-323.49, 370.42], [-132.72, 329.34], [216.52, 327.74], [118.7, 367.87], [-323.49, 206.52], [-132.72, 199.52], [216.52, 199.25], [118.7, 206.09]]}, {\"category\": \"truck\", \"corners_3d\": [[643.63, 198.44], [678.22, 198.44], [687.98, 202.11], [648.43, 202.11], [643.63, 165.24], [678.22, 165.24], [687.98, 164.15], [648.43, 164.15]]}, {\"category\": \"car\", \"corners_3d\": [[663.28, 209.79], [697.57, 209.81], [708.84, 215.0], [669.73, 214.98], [663.28, 180.53], [697.57, 180.53], [708.84, 181.61], [669.73, 181.6]]}, {\"category\": \"van\", \"corners_3d\": [[536.01, 212.86], [532.73, 215.38], [406.14, 215.6], [416.99, 213.05], [536.01, 165.46], [532.73, 165.0], [406.14, 164.95], [416.99, 165.43]]}, {\"category\": \"car\", \"corners_3d\": [[526.4, 216.37], [520.65, 219.37], [397.69, 219.37], [411.38, 216.37], [526.4, 179.06], [520.65, 179.49], [397.69, 179.49], [411.38, 179.06]]}, {\"category\": \"car\", \"corners_3d\": [[493.9, 227.61], [481.22, 232.18], [333.43, 231.57], [357.27, 227.09], [493.9, 181.2], [481.22, 181.89], [333.43, 181.8], [357.27, 181.12]]}, {\"category\": \"car\", \"corners_3d\": [[487.41, 235.03], [469.16, 241.08], [310.49, 239.71], [342.23, 233.88], [487.41, 180.35], [469.16, 181.08], [310.49, 180.91], [342.23, 180.21]]}, {\"category\": \"car\", \"corners_3d\": [[468.7, 244.45], [449.55, 250.88], [282.25, 249.03], [314.46, 242.89], [468.7, 183.07], [449.55, 183.99], [282.25, 183.72], [314.46, 182.85]]}, {\"category\": \"car\", \"corners_3d\": [[426.1, 252.62], [403.48, 262.08], [216.63, 261.85], [258.96, 252.44], [426.1, 181.24], [403.48, 182.23], [216.63, 182.21], [258.96, 181.22]]}, {\"category\": \"car\", \"corners_3d\": [[-68.93, 313.81], [31.58, 292.96], [270.25, 293.03], [211.13, 313.89], [-68.93, 186.92], [31.58, 184.84], [270.25, 184.85], [211.13, 186.93]]}]\n```" - } - ] - }, - { - "id": 2501, - "data_source": "Mono3DRefer", - "scene_name": "002732", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[140.81, 232.01], [182.02, 231.73], [127.09, 240.64], [79.3, 241.02], [140.81, 199.19], [182.02, 199.07], [127.09, 203.03], [79.3, 203.2]]}, {\"category\": \"car\", \"corners_3d\": [[78.29, 233.34], [119.35, 233.17], [50.2, 242.35], [2.66, 242.56], [78.29, 197.97], [119.35, 197.9], [50.2, 201.71], [2.66, 201.8]]}]\n```" - } - ] - }, - { - "id": 2502, - "data_source": "Mono3DRefer", - "scene_name": "002736", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[553.92, 192.13], [537.62, 192.14], [540.32, 191.09], [555.73, 191.08], [553.92, 174.93], [537.62, 174.94], [540.32, 174.82], [555.73, 174.82]]}]\n```" - } - ] - }, - { - "id": 2503, - "data_source": "Mono3DRefer", - "scene_name": "002739", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[550.71, 268.75], [591.74, 269.33], [582.79, 278.06], [537.82, 277.37], [550.71, 158.46], [591.74, 158.32], [582.79, 156.14], [537.82, 156.31]]}]\n```" - } - ] - }, - { - "id": 2504, - "data_source": "Mono3DRefer", - "scene_name": "002741", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[344.13, 307.86], [463.89, 308.55], [327.63, 418.68], [112.86, 416.4], [344.13, 185.92], [463.89, 185.99], [327.63, 196.64], [112.86, 196.42]]}, {\"category\": \"car\", \"corners_3d\": [[527.67, 258.63], [456.8, 258.38], [489.43, 242.32], [547.05, 242.49], [527.67, 190.68], [456.8, 190.63], [489.43, 187.29], [547.05, 187.32]]}, {\"category\": \"car\", \"corners_3d\": [[590.21, 209.66], [564.13, 209.64], [569.0, 206.61], [592.94, 206.64], [590.21, 187.32], [564.13, 187.31], [569.0, 186.12], [592.94, 186.13]]}]\n```" - } - ] - }, - { - "id": 2505, - "data_source": "Mono3DRefer", - "scene_name": "002751", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[570.29, 220.98], [600.37, 220.98], [599.22, 226.64], [565.6, 226.63], [570.29, 193.22], [600.37, 193.23], [599.22, 195.62], [565.6, 195.62]]}, {\"category\": \"car\", \"corners_3d\": [[1172.97, 449.03], [996.79, 368.02], [1502.27, 360.76], [1868.23, 434.71], [1172.97, 174.61], [996.79, 174.09], [1502.27, 174.05], [1868.23, 174.52]]}, {\"category\": \"car\", \"corners_3d\": [[1274.79, 317.97], [1417.41, 350.87], [949.97, 345.41], [895.15, 314.32], [1274.79, 185.37], [1417.41, 188.2], [949.97, 187.73], [895.15, 185.05]]}, {\"category\": \"car\", \"corners_3d\": [[959.56, 256.36], [991.17, 266.39], [758.69, 262.98], [752.24, 253.62], [959.56, 186.68], [991.17, 188.34], [758.69, 187.78], [752.24, 186.23]]}, {\"category\": \"car\", \"corners_3d\": [[874.74, 303.19], [840.92, 285.0], [1101.52, 286.27], [1178.62, 304.9], [874.74, 181.41], [840.92, 180.22], [1101.52, 180.3], [1178.62, 181.53]]}]\n```" - } - ] - }, - { - "id": 2506, - "data_source": "Mono3DRefer", - "scene_name": "002776", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[532.2, 230.03], [482.06, 230.0], [501.75, 221.61], [544.54, 221.63], [532.2, 185.91], [482.06, 185.9], [501.75, 183.98], [544.54, 183.99]]}, {\"category\": \"car\", \"corners_3d\": [[569.51, 202.77], [544.47, 202.76], [549.45, 200.69], [572.75, 200.7], [569.51, 181.55], [544.47, 181.54], [549.45, 180.94], [572.75, 180.94]]}]\n```" - } - ] - }, - { - "id": 2507, - "data_source": "Mono3DRefer", - "scene_name": "002797", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[705.25, 256.09], [789.06, 256.1], [874.47, 295.83], [750.65, 295.81], [705.25, 176.06], [789.06, 176.06], [874.47, 177.59], [750.65, 177.59]]}, {\"category\": \"car\", \"corners_3d\": [[281.41, 311.03], [415.7, 310.03], [255.11, 443.09], [-13.6, 446.97], [281.41, 187.6], [415.7, 187.49], [255.11, 201.69], [-13.6, 202.11]]}, {\"category\": \"car\", \"corners_3d\": [[647.41, 209.3], [686.92, 209.42], [691.35, 214.26], [646.61, 214.1], [647.41, 176.2], [686.92, 176.21], [691.35, 176.66], [646.61, 176.64]]}]\n```" - } - ] - }, - { - "id": 2508, - "data_source": "Mono3DRefer", - "scene_name": "002797", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[707.41, 187.95], [728.9, 188.14], [711.57, 189.39], [688.59, 189.17], [707.41, 159.19], [728.9, 159.02], [711.57, 157.89], [688.59, 158.09]]}]\n```" - } - ] - }, - { - "id": 2509, - "data_source": "Mono3DRefer", - "scene_name": "002797", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[705.25, 256.09], [789.06, 256.1], [874.47, 295.83], [750.65, 295.81], [705.25, 176.06], [789.06, 176.06], [874.47, 177.59], [750.65, 177.59]]}, {\"category\": \"car\", \"corners_3d\": [[281.41, 311.03], [415.7, 310.03], [255.11, 443.09], [-13.6, 446.97], [281.41, 187.6], [415.7, 187.49], [255.11, 201.69], [-13.6, 202.11]]}, {\"category\": \"car\", \"corners_3d\": [[647.41, 209.3], [686.92, 209.42], [691.35, 214.26], [646.61, 214.1], [647.41, 176.2], [686.92, 176.21], [691.35, 176.66], [646.61, 176.64]]}, {\"category\": \"van\", \"corners_3d\": [[707.41, 187.95], [728.9, 188.14], [711.57, 189.39], [688.59, 189.17], [707.41, 159.19], [728.9, 159.02], [711.57, 157.89], [688.59, 158.09]]}]\n```" - } - ] - }, - { - "id": 2510, - "data_source": "Mono3DRefer", - "scene_name": "002798", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[537.62, 215.41], [552.73, 215.32], [553.85, 218.5], [537.61, 218.6], [537.62, 174.52], [552.73, 174.52], [553.85, 174.64], [537.61, 174.64]]}]\n```" - } - ] - }, - { - "id": 2511, - "data_source": "Mono3DRefer", - "scene_name": "002798", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1182.97, 306.67], [1340.45, 338.23], [939.51, 352.07], [853.14, 315.59], [1182.97, 182.03], [1340.45, 184.19], [939.51, 185.14], [853.14, 182.64]]}, {\"category\": \"car\", \"corners_3d\": [[731.83, 280.5], [699.8, 266.56], [923.05, 261.42], [984.23, 273.78], [731.83, 160.51], [699.8, 162.11], [923.05, 162.7], [984.23, 161.28]]}, {\"category\": \"car\", \"corners_3d\": [[-827.07, 470.67], [-495.05, 396.12], [89.42, 373.73], [-29.74, 432.12], [-827.07, 230.14], [-495.05, 215.8], [89.42, 211.49], [-29.74, 222.73]]}, {\"category\": \"car\", \"corners_3d\": [[18.64, 283.04], [75.0, 270.5], [289.48, 267.0], [261.82, 278.6], [18.64, 197.39], [75.0, 194.6], [289.48, 193.82], [261.82, 196.4]]}, {\"category\": \"car\", \"corners_3d\": [[205.56, 245.26], [226.54, 239.67], [354.79, 237.68], [344.88, 242.93], [205.56, 191.42], [226.54, 189.98], [354.79, 189.48], [344.88, 190.82]]}, {\"category\": \"car\", \"corners_3d\": [[331.97, 218.98], [340.49, 216.82], [422.07, 216.2], [417.62, 218.29], [331.97, 186.29], [340.49, 185.66], [422.07, 185.48], [417.62, 186.09]]}]\n```" - } - ] - }, - { - "id": 2512, - "data_source": "Mono3DRefer", - "scene_name": "002798", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1182.97, 306.67], [1340.45, 338.23], [939.51, 352.07], [853.14, 315.59], [1182.97, 182.03], [1340.45, 184.19], [939.51, 185.14], [853.14, 182.64]]}, {\"category\": \"car\", \"corners_3d\": [[731.83, 280.5], [699.8, 266.56], [923.05, 261.42], [984.23, 273.78], [731.83, 160.51], [699.8, 162.11], [923.05, 162.7], [984.23, 161.28]]}, {\"category\": \"car\", \"corners_3d\": [[-827.07, 470.67], [-495.05, 396.12], [89.42, 373.73], [-29.74, 432.12], [-827.07, 230.14], [-495.05, 215.8], [89.42, 211.49], [-29.74, 222.73]]}, {\"category\": \"cyclist\", \"corners_3d\": [[537.62, 215.41], [552.73, 215.32], [553.85, 218.5], [537.61, 218.6], [537.62, 174.52], [552.73, 174.52], [553.85, 174.64], [537.61, 174.64]]}, {\"category\": \"car\", \"corners_3d\": [[18.64, 283.04], [75.0, 270.5], [289.48, 267.0], [261.82, 278.6], [18.64, 197.39], [75.0, 194.6], [289.48, 193.82], [261.82, 196.4]]}, {\"category\": \"car\", \"corners_3d\": [[205.56, 245.26], [226.54, 239.67], [354.79, 237.68], [344.88, 242.93], [205.56, 191.42], [226.54, 189.98], [354.79, 189.48], [344.88, 190.82]]}, {\"category\": \"car\", \"corners_3d\": [[331.97, 218.98], [340.49, 216.82], [422.07, 216.2], [417.62, 218.29], [331.97, 186.29], [340.49, 185.66], [422.07, 185.48], [417.62, 186.09]]}]\n```" - } - ] - }, - { - "id": 2513, - "data_source": "Mono3DRefer", - "scene_name": "002800", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[181.53, 392.57], [-41.59, 397.64], [205.79, 306.06], [332.28, 304.36], [181.53, 196.33], [-41.59, 196.68], [205.79, 190.25], [332.28, 190.13]]}]\n```" - } - ] - }, - { - "id": 2514, - "data_source": "Mono3DRefer", - "scene_name": "002800", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[132.6, 537.74], [-47.91, 542.15], [174.86, 410.83], [288.41, 409.04], [132.6, 144.0], [-47.91, 143.54], [174.86, 157.36], [288.41, 157.55]]}]\n```" - } - ] - }, - { - "id": 2515, - "data_source": "Mono3DRefer", - "scene_name": "002800", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[491.26, 217.67], [514.45, 217.5], [515.19, 218.7], [491.22, 218.88], [491.26, 170.23], [514.45, 170.28], [515.19, 169.91], [491.22, 169.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[510.32, 217.43], [534.94, 217.33], [534.62, 218.66], [509.09, 218.77], [510.32, 168.12], [534.94, 168.16], [534.62, 167.66], [509.09, 167.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[367.05, 222.58], [395.04, 222.44], [388.92, 224.21], [359.68, 224.36], [367.05, 169.55], [395.04, 169.59], [388.92, 169.08], [359.68, 169.03]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[737.07, 245.22], [683.92, 246.43], [670.73, 243.14], [721.47, 242.04], [737.07, 149.01], [683.92, 148.39], [670.73, 150.07], [721.47, 150.63]]}]\n```" - } - ] - }, - { - "id": 2516, - "data_source": "Mono3DRefer", - "scene_name": "002800", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[181.53, 392.57], [-41.59, 397.64], [205.79, 306.06], [332.28, 304.36], [181.53, 196.33], [-41.59, 196.68], [205.79, 190.25], [332.28, 190.13]]}, {\"category\": \"cyclist\", \"corners_3d\": [[132.6, 537.74], [-47.91, 542.15], [174.86, 410.83], [288.41, 409.04], [132.6, 144.0], [-47.91, 143.54], [174.86, 157.36], [288.41, 157.55]]}]\n```" - } - ] - }, - { - "id": 2517, - "data_source": "Mono3DRefer", - "scene_name": "002800", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[491.26, 217.67], [514.45, 217.5], [515.19, 218.7], [491.22, 218.88], [491.26, 170.23], [514.45, 170.28], [515.19, 169.91], [491.22, 169.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[510.32, 217.43], [534.94, 217.33], [534.62, 218.66], [509.09, 218.77], [510.32, 168.12], [534.94, 168.16], [534.62, 167.66], [509.09, 167.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[367.05, 222.58], [395.04, 222.44], [388.92, 224.21], [359.68, 224.36], [367.05, 169.55], [395.04, 169.59], [388.92, 169.08], [359.68, 169.03]]}, {\"category\": \"car\", \"corners_3d\": [[181.53, 392.57], [-41.59, 397.64], [205.79, 306.06], [332.28, 304.36], [181.53, 196.33], [-41.59, 196.68], [205.79, 190.25], [332.28, 190.13]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[737.07, 245.22], [683.92, 246.43], [670.73, 243.14], [721.47, 242.04], [737.07, 149.01], [683.92, 148.39], [670.73, 150.07], [721.47, 150.63]]}]\n```" - } - ] - }, - { - "id": 2518, - "data_source": "Mono3DRefer", - "scene_name": "002800", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[491.26, 217.67], [514.45, 217.5], [515.19, 218.7], [491.22, 218.88], [491.26, 170.23], [514.45, 170.28], [515.19, 169.91], [491.22, 169.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[510.32, 217.43], [534.94, 217.33], [534.62, 218.66], [509.09, 218.77], [510.32, 168.12], [534.94, 168.16], [534.62, 167.66], [509.09, 167.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[367.05, 222.58], [395.04, 222.44], [388.92, 224.21], [359.68, 224.36], [367.05, 169.55], [395.04, 169.59], [388.92, 169.08], [359.68, 169.03]]}, {\"category\": \"cyclist\", \"corners_3d\": [[132.6, 537.74], [-47.91, 542.15], [174.86, 410.83], [288.41, 409.04], [132.6, 144.0], [-47.91, 143.54], [174.86, 157.36], [288.41, 157.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[737.07, 245.22], [683.92, 246.43], [670.73, 243.14], [721.47, 242.04], [737.07, 149.01], [683.92, 148.39], [670.73, 150.07], [721.47, 150.63]]}]\n```" - } - ] - }, - { - "id": 2519, - "data_source": "Mono3DRefer", - "scene_name": "002800", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[491.26, 217.67], [514.45, 217.5], [515.19, 218.7], [491.22, 218.88], [491.26, 170.23], [514.45, 170.28], [515.19, 169.91], [491.22, 169.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[510.32, 217.43], [534.94, 217.33], [534.62, 218.66], [509.09, 218.77], [510.32, 168.12], [534.94, 168.16], [534.62, 167.66], [509.09, 167.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[367.05, 222.58], [395.04, 222.44], [388.92, 224.21], [359.68, 224.36], [367.05, 169.55], [395.04, 169.59], [388.92, 169.08], [359.68, 169.03]]}, {\"category\": \"car\", \"corners_3d\": [[181.53, 392.57], [-41.59, 397.64], [205.79, 306.06], [332.28, 304.36], [181.53, 196.33], [-41.59, 196.68], [205.79, 190.25], [332.28, 190.13]]}, {\"category\": \"cyclist\", \"corners_3d\": [[132.6, 537.74], [-47.91, 542.15], [174.86, 410.83], [288.41, 409.04], [132.6, 144.0], [-47.91, 143.54], [174.86, 157.36], [288.41, 157.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[737.07, 245.22], [683.92, 246.43], [670.73, 243.14], [721.47, 242.04], [737.07, 149.01], [683.92, 148.39], [670.73, 150.07], [721.47, 150.63]]}]\n```" - } - ] - }, - { - "id": 2520, - "data_source": "Mono3DRefer", - "scene_name": "002818", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[660.88, 214.52], [704.55, 214.52], [718.75, 220.79], [668.51, 220.79], [660.88, 176.84], [704.55, 176.84], [718.75, 177.44], [668.51, 177.44]]}, {\"category\": \"car\", \"corners_3d\": [[465.62, 231.14], [424.96, 231.11], [449.95, 223.5], [485.3, 223.52], [465.62, 194.43], [424.96, 194.42], [449.95, 191.61], [485.3, 191.61]]}, {\"category\": \"car\", \"corners_3d\": [[502.66, 218.89], [468.89, 218.89], [483.52, 214.09], [513.76, 214.09], [502.66, 183.28], [468.89, 183.28], [483.52, 182.19], [513.76, 182.19]]}, {\"category\": \"car\", \"corners_3d\": [[515.21, 211.35], [487.07, 211.34], [498.3, 207.99], [524.0, 208.0], [515.21, 183.68], [487.07, 183.67], [498.3, 182.73], [524.0, 182.74]]}, {\"category\": \"car\", \"corners_3d\": [[27.11, 242.12], [66.29, 242.12], [-7.56, 251.53], [-52.06, 251.52], [27.11, 210.97], [66.29, 210.97], [-7.56, 216.14], [-52.06, 216.14]]}, {\"category\": \"car\", \"corners_3d\": [[82.24, 239.91], [117.06, 239.91], [49.94, 249.04], [10.39, 249.04], [82.24, 209.43], [117.06, 209.43], [49.94, 214.41], [10.39, 214.4]]}, {\"category\": \"car\", \"corners_3d\": [[135.1, 237.69], [172.22, 237.72], [101.28, 248.06], [58.3, 248.01], [135.1, 204.94], [172.22, 204.96], [101.28, 210.07], [58.3, 210.05]]}, {\"category\": \"car\", \"corners_3d\": [[430.97, 245.46], [379.5, 245.51], [410.44, 235.4], [454.73, 235.36], [430.97, 198.16], [379.5, 198.18], [410.44, 194.65], [454.73, 194.64]]}]\n```" - } - ] - }, - { - "id": 2521, - "data_source": "Mono3DRefer", - "scene_name": "002818", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1151.73, 181.03], [1111.54, 181.06], [1044.57, 180.07], [1080.15, 180.04], [1151.73, 132.78], [1111.54, 132.61], [1044.57, 137.48], [1080.15, 137.61]]}]\n```" - } - ] - }, - { - "id": 2522, - "data_source": "Mono3DRefer", - "scene_name": "002818", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[660.88, 214.52], [704.55, 214.52], [718.75, 220.79], [668.51, 220.79], [660.88, 176.84], [704.55, 176.84], [718.75, 177.44], [668.51, 177.44]]}, {\"category\": \"car\", \"corners_3d\": [[465.62, 231.14], [424.96, 231.11], [449.95, 223.5], [485.3, 223.52], [465.62, 194.43], [424.96, 194.42], [449.95, 191.61], [485.3, 191.61]]}, {\"category\": \"car\", \"corners_3d\": [[502.66, 218.89], [468.89, 218.89], [483.52, 214.09], [513.76, 214.09], [502.66, 183.28], [468.89, 183.28], [483.52, 182.19], [513.76, 182.19]]}, {\"category\": \"car\", \"corners_3d\": [[515.21, 211.35], [487.07, 211.34], [498.3, 207.99], [524.0, 208.0], [515.21, 183.68], [487.07, 183.67], [498.3, 182.73], [524.0, 182.74]]}, {\"category\": \"van\", \"corners_3d\": [[1151.73, 181.03], [1111.54, 181.06], [1044.57, 180.07], [1080.15, 180.04], [1151.73, 132.78], [1111.54, 132.61], [1044.57, 137.48], [1080.15, 137.61]]}, {\"category\": \"car\", \"corners_3d\": [[27.11, 242.12], [66.29, 242.12], [-7.56, 251.53], [-52.06, 251.52], [27.11, 210.97], [66.29, 210.97], [-7.56, 216.14], [-52.06, 216.14]]}, {\"category\": \"car\", \"corners_3d\": [[82.24, 239.91], [117.06, 239.91], [49.94, 249.04], [10.39, 249.04], [82.24, 209.43], [117.06, 209.43], [49.94, 214.41], [10.39, 214.4]]}, {\"category\": \"car\", \"corners_3d\": [[135.1, 237.69], [172.22, 237.72], [101.28, 248.06], [58.3, 248.01], [135.1, 204.94], [172.22, 204.96], [101.28, 210.07], [58.3, 210.05]]}, {\"category\": \"car\", \"corners_3d\": [[430.97, 245.46], [379.5, 245.51], [410.44, 235.4], [454.73, 235.36], [430.97, 198.16], [379.5, 198.18], [410.44, 194.65], [454.73, 194.64]]}]\n```" - } - ] - }, - { - "id": 2523, - "data_source": "Mono3DRefer", - "scene_name": "002828", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[291.04, 215.47], [241.09, 215.47], [303.44, 208.25], [344.92, 208.25], [291.04, 144.32], [241.09, 144.32], [303.44, 149.15], [344.92, 149.15]]}, {\"category\": \"truck\", \"corners_3d\": [[436.78, 197.57], [412.92, 197.57], [430.04, 195.41], [451.82, 195.41], [436.78, 165.7], [412.92, 165.7], [430.04, 166.33], [451.82, 166.33]]}, {\"category\": \"truck\", \"corners_3d\": [[193.25, 230.0], [126.47, 230.06], [228.81, 217.74], [281.14, 217.7], [193.25, 137.57], [126.47, 137.54], [228.81, 145.14], [281.14, 145.16]]}]\n```" - } - ] - }, - { - "id": 2524, - "data_source": "Mono3DRefer", - "scene_name": "002828", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[358.54, 216.82], [321.62, 216.82], [353.52, 211.94], [386.34, 211.94], [358.54, 186.39], [321.62, 186.39], [353.52, 184.89], [386.34, 184.89]]}]\n```" - } - ] - }, - { - "id": 2525, - "data_source": "Mono3DRefer", - "scene_name": "002828", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[291.04, 215.47], [241.09, 215.47], [303.44, 208.25], [344.92, 208.25], [291.04, 144.32], [241.09, 144.32], [303.44, 149.15], [344.92, 149.15]]}, {\"category\": \"truck\", \"corners_3d\": [[436.78, 197.57], [412.92, 197.57], [430.04, 195.41], [451.82, 195.41], [436.78, 165.7], [412.92, 165.7], [430.04, 166.33], [451.82, 166.33]]}, {\"category\": \"car\", \"corners_3d\": [[358.54, 216.82], [321.62, 216.82], [353.52, 211.94], [386.34, 211.94], [358.54, 186.39], [321.62, 186.39], [353.52, 184.89], [386.34, 184.89]]}, {\"category\": \"truck\", \"corners_3d\": [[193.25, 230.0], [126.47, 230.06], [228.81, 217.74], [281.14, 217.7], [193.25, 137.57], [126.47, 137.54], [228.81, 145.14], [281.14, 145.16]]}]\n```" - } - ] - }, - { - "id": 2526, - "data_source": "Mono3DRefer", - "scene_name": "002842", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[522.45, 235.22], [472.41, 235.13], [495.53, 225.6], [537.93, 225.66], [522.45, 187.62], [472.41, 187.6], [495.53, 185.34], [537.93, 185.35]]}]\n```" - } - ] - }, - { - "id": 2527, - "data_source": "Mono3DRefer", - "scene_name": "002842", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[338.53, 230.3], [324.26, 230.22], [342.16, 227.16], [355.69, 227.23], [338.53, 186.73], [324.26, 186.71], [342.16, 185.97], [355.69, 185.99]]}]\n```" - } - ] - }, - { - "id": 2528, - "data_source": "Mono3DRefer", - "scene_name": "002842", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[522.45, 235.22], [472.41, 235.13], [495.53, 225.6], [537.93, 225.66], [522.45, 187.62], [472.41, 187.6], [495.53, 185.34], [537.93, 185.35]]}, {\"category\": \"cyclist\", \"corners_3d\": [[338.53, 230.3], [324.26, 230.22], [342.16, 227.16], [355.69, 227.23], [338.53, 186.73], [324.26, 186.71], [342.16, 185.97], [355.69, 185.99]]}]\n```" - } - ] - }, - { - "id": 2529, - "data_source": "Mono3DRefer", - "scene_name": "002844", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[89.01, 249.33], [38.39, 249.28], [122.98, 238.09], [166.23, 238.13], [89.01, 202.54], [38.39, 202.52], [122.98, 198.18], [166.23, 198.19]]}, {\"category\": \"car\", \"corners_3d\": [[303.93, 226.72], [266.65, 226.69], [297.34, 221.96], [331.36, 221.99], [303.93, 194.07], [266.65, 194.06], [297.34, 192.2], [331.36, 192.21]]}]\n```" - } - ] - }, - { - "id": 2530, - "data_source": "Mono3DRefer", - "scene_name": "002852", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[182.19, 499.84], [-64.79, 502.14], [219.21, 360.31], [359.1, 359.56], [182.19, 211.77], [-64.79, 212.04], [219.21, 195.16], [359.1, 195.07]]}, {\"category\": \"car\", \"corners_3d\": [[539.64, 209.2], [516.14, 209.4], [513.93, 206.8], [535.78, 206.62], [539.64, 185.16], [516.14, 185.22], [513.93, 184.34], [535.78, 184.28]]}]\n```" - } - ] - }, - { - "id": 2531, - "data_source": "Mono3DRefer", - "scene_name": "002862", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[560.11, 214.98], [526.38, 215.22], [525.94, 210.99], [556.32, 210.8], [560.11, 180.45], [526.38, 180.49], [525.94, 179.73], [556.32, 179.7]]}, {\"category\": \"car\", \"corners_3d\": [[526.75, 201.78], [503.89, 202.02], [497.5, 199.87], [518.74, 199.67], [526.75, 182.09], [503.89, 182.17], [497.5, 181.48], [518.74, 181.42]]}]\n```" - } - ] - }, - { - "id": 2532, - "data_source": "Mono3DRefer", - "scene_name": "002863", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[443.66, 217.34], [424.26, 217.31], [427.84, 216.51], [446.89, 216.54], [443.66, 179.35], [424.26, 179.34], [427.84, 179.23], [446.89, 179.23]]}]\n```" - } - ] - }, - { - "id": 2533, - "data_source": "Mono3DRefer", - "scene_name": "002867", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[143.87, 217.22], [100.1, 217.24], [172.94, 210.8], [210.32, 210.78], [143.87, 172.38], [100.1, 172.38], [172.94, 172.45], [210.32, 172.45]]}]\n```" - } - ] - }, - { - "id": 2534, - "data_source": "Mono3DRefer", - "scene_name": "002867", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[702.83, 200.69], [691.99, 200.7], [690.75, 200.35], [701.46, 200.34], [702.83, 175.46], [691.99, 175.46], [690.75, 175.43], [701.46, 175.43]]}]\n```" - } - ] - }, - { - "id": 2535, - "data_source": "Mono3DRefer", - "scene_name": "002867", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[303.9, 216.86], [265.81, 216.88], [299.28, 212.5], [333.55, 212.48], [303.9, 182.02], [265.81, 182.03], [299.28, 181.12], [333.55, 181.11]]}, {\"category\": \"car\", \"corners_3d\": [[253.9, 216.7], [213.85, 216.68], [261.57, 211.48], [296.89, 211.5], [253.9, 184.32], [213.85, 184.31], [261.57, 182.95], [296.89, 182.96]]}, {\"category\": \"car\", \"corners_3d\": [[404.72, 202.12], [377.13, 202.09], [396.69, 199.83], [422.17, 199.86], [404.72, 177.07], [377.13, 177.07], [396.69, 176.74], [422.17, 176.75]]}, {\"category\": \"car\", \"corners_3d\": [[372.66, 201.76], [346.34, 201.74], [366.83, 199.61], [391.22, 199.62], [372.66, 177.04], [346.34, 177.04], [366.83, 176.73], [391.22, 176.73]]}, {\"category\": \"car\", \"corners_3d\": [[471.24, 193.37], [452.22, 193.35], [461.74, 192.26], [479.75, 192.27], [471.24, 176.74], [452.22, 176.74], [461.74, 176.53], [479.75, 176.54]]}]\n```" - } - ] - }, - { - "id": 2536, - "data_source": "Mono3DRefer", - "scene_name": "002867", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[702.83, 200.69], [691.99, 200.7], [690.75, 200.35], [701.46, 200.34], [702.83, 175.46], [691.99, 175.46], [690.75, 175.43], [701.46, 175.43]]}, {\"category\": \"van\", \"corners_3d\": [[143.87, 217.22], [100.1, 217.24], [172.94, 210.8], [210.32, 210.78], [143.87, 172.38], [100.1, 172.38], [172.94, 172.45], [210.32, 172.45]]}]\n```" - } - ] - }, - { - "id": 2537, - "data_source": "Mono3DRefer", - "scene_name": "002867", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[143.87, 217.22], [100.1, 217.24], [172.94, 210.8], [210.32, 210.78], [143.87, 172.38], [100.1, 172.38], [172.94, 172.45], [210.32, 172.45]]}, {\"category\": \"car\", \"corners_3d\": [[303.9, 216.86], [265.81, 216.88], [299.28, 212.5], [333.55, 212.48], [303.9, 182.02], [265.81, 182.03], [299.28, 181.12], [333.55, 181.11]]}, {\"category\": \"car\", \"corners_3d\": [[253.9, 216.7], [213.85, 216.68], [261.57, 211.48], [296.89, 211.5], [253.9, 184.32], [213.85, 184.31], [261.57, 182.95], [296.89, 182.96]]}, {\"category\": \"car\", \"corners_3d\": [[404.72, 202.12], [377.13, 202.09], [396.69, 199.83], [422.17, 199.86], [404.72, 177.07], [377.13, 177.07], [396.69, 176.74], [422.17, 176.75]]}, {\"category\": \"car\", \"corners_3d\": [[372.66, 201.76], [346.34, 201.74], [366.83, 199.61], [391.22, 199.62], [372.66, 177.04], [346.34, 177.04], [366.83, 176.73], [391.22, 176.73]]}, {\"category\": \"car\", \"corners_3d\": [[471.24, 193.37], [452.22, 193.35], [461.74, 192.26], [479.75, 192.27], [471.24, 176.74], [452.22, 176.74], [461.74, 176.53], [479.75, 176.54]]}]\n```" - } - ] - }, - { - "id": 2538, - "data_source": "Mono3DRefer", - "scene_name": "002867", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[702.83, 200.69], [691.99, 200.7], [690.75, 200.35], [701.46, 200.34], [702.83, 175.46], [691.99, 175.46], [690.75, 175.43], [701.46, 175.43]]}, {\"category\": \"car\", \"corners_3d\": [[303.9, 216.86], [265.81, 216.88], [299.28, 212.5], [333.55, 212.48], [303.9, 182.02], [265.81, 182.03], [299.28, 181.12], [333.55, 181.11]]}, {\"category\": \"car\", \"corners_3d\": [[253.9, 216.7], [213.85, 216.68], [261.57, 211.48], [296.89, 211.5], [253.9, 184.32], [213.85, 184.31], [261.57, 182.95], [296.89, 182.96]]}, {\"category\": \"car\", \"corners_3d\": [[404.72, 202.12], [377.13, 202.09], [396.69, 199.83], [422.17, 199.86], [404.72, 177.07], [377.13, 177.07], [396.69, 176.74], [422.17, 176.75]]}, {\"category\": \"car\", \"corners_3d\": [[372.66, 201.76], [346.34, 201.74], [366.83, 199.61], [391.22, 199.62], [372.66, 177.04], [346.34, 177.04], [366.83, 176.73], [391.22, 176.73]]}, {\"category\": \"car\", \"corners_3d\": [[471.24, 193.37], [452.22, 193.35], [461.74, 192.26], [479.75, 192.27], [471.24, 176.74], [452.22, 176.74], [461.74, 176.53], [479.75, 176.54]]}]\n```" - } - ] - }, - { - "id": 2539, - "data_source": "Mono3DRefer", - "scene_name": "002867", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[702.83, 200.69], [691.99, 200.7], [690.75, 200.35], [701.46, 200.34], [702.83, 175.46], [691.99, 175.46], [690.75, 175.43], [701.46, 175.43]]}, {\"category\": \"van\", \"corners_3d\": [[143.87, 217.22], [100.1, 217.24], [172.94, 210.8], [210.32, 210.78], [143.87, 172.38], [100.1, 172.38], [172.94, 172.45], [210.32, 172.45]]}, {\"category\": \"car\", \"corners_3d\": [[303.9, 216.86], [265.81, 216.88], [299.28, 212.5], [333.55, 212.48], [303.9, 182.02], [265.81, 182.03], [299.28, 181.12], [333.55, 181.11]]}, {\"category\": \"car\", \"corners_3d\": [[253.9, 216.7], [213.85, 216.68], [261.57, 211.48], [296.89, 211.5], [253.9, 184.32], [213.85, 184.31], [261.57, 182.95], [296.89, 182.96]]}, {\"category\": \"car\", \"corners_3d\": [[404.72, 202.12], [377.13, 202.09], [396.69, 199.83], [422.17, 199.86], [404.72, 177.07], [377.13, 177.07], [396.69, 176.74], [422.17, 176.75]]}, {\"category\": \"car\", \"corners_3d\": [[372.66, 201.76], [346.34, 201.74], [366.83, 199.61], [391.22, 199.62], [372.66, 177.04], [346.34, 177.04], [366.83, 176.73], [391.22, 176.73]]}, {\"category\": \"car\", \"corners_3d\": [[471.24, 193.37], [452.22, 193.35], [461.74, 192.26], [479.75, 192.27], [471.24, 176.74], [452.22, 176.74], [461.74, 176.53], [479.75, 176.54]]}]\n```" - } - ] - }, - { - "id": 2540, - "data_source": "Mono3DRefer", - "scene_name": "002875", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[464.33, 389.73], [412.19, 373.78], [519.82, 347.73], [572.04, 359.54], [464.33, 136.17], [412.19, 139.55], [519.82, 145.07], [572.04, 142.56]]}]\n```" - } - ] - }, - { - "id": 2541, - "data_source": "Mono3DRefer", - "scene_name": "002876", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[758.99, 207.08], [767.92, 208.79], [672.71, 208.98], [668.29, 207.25], [758.99, 175.86], [767.92, 176.01], [672.71, 176.03], [668.29, 175.88]]}, {\"category\": \"car\", \"corners_3d\": [[649.31, 195.18], [649.23, 194.49], [701.2, 194.57], [702.94, 195.26], [649.31, 175.39], [649.23, 175.31], [701.2, 175.32], [702.94, 175.4]]}]\n```" - } - ] - }, - { - "id": 2542, - "data_source": "Mono3DRefer", - "scene_name": "002882", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[214.7, 463.51], [-50.89, 461.43], [272.27, 323.13], [411.31, 323.69], [214.7, 225.94], [-50.89, 225.56], [272.27, 200.3], [411.31, 200.4]]}, {\"category\": \"car\", \"corners_3d\": [[544.23, 224.45], [499.91, 224.36], [515.18, 218.34], [554.34, 218.42], [544.23, 186.73], [499.91, 186.7], [515.18, 185.09], [554.34, 185.11]]}]\n```" - } - ] - }, - { - "id": 2543, - "data_source": "Mono3DRefer", - "scene_name": "002884", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[469.03, 247.13], [394.76, 248.11], [417.97, 233.54], [477.77, 232.9], [469.03, 172.84], [394.76, 172.84], [417.97, 172.84], [477.77, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[479.47, 211.06], [438.3, 211.66], [433.78, 206.89], [470.01, 206.43], [479.47, 176.38], [438.3, 176.44], [433.78, 176.0], [470.01, 175.96]]}]\n```" - } - ] - }, - { - "id": 2544, - "data_source": "Mono3DRefer", - "scene_name": "002885", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[803.12, 223.86], [829.18, 223.86], [838.43, 226.02], [811.28, 226.02], [803.12, 162.85], [829.18, 162.85], [838.43, 162.43], [811.28, 162.43]]}]\n```" - } - ] - }, - { - "id": 2545, - "data_source": "Mono3DRefer", - "scene_name": "002885", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[692.71, 205.69], [709.14, 205.74], [712.05, 207.72], [694.62, 207.66], [692.71, 165.43], [709.14, 165.42], [712.05, 164.97], [694.62, 164.99]]}, {\"category\": \"cyclist\", \"corners_3d\": [[678.75, 196.36], [686.43, 196.39], [686.31, 197.25], [678.34, 197.22], [678.75, 169.91], [686.43, 169.91], [686.31, 169.8], [678.34, 169.8]]}]\n```" - } - ] - }, - { - "id": 2546, - "data_source": "Mono3DRefer", - "scene_name": "002885", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[803.12, 223.86], [829.18, 223.86], [838.43, 226.02], [811.28, 226.02], [803.12, 162.85], [829.18, 162.85], [838.43, 162.43], [811.28, 162.43]]}, {\"category\": \"cyclist\", \"corners_3d\": [[692.71, 205.69], [709.14, 205.74], [712.05, 207.72], [694.62, 207.66], [692.71, 165.43], [709.14, 165.42], [712.05, 164.97], [694.62, 164.99]]}, {\"category\": \"cyclist\", \"corners_3d\": [[678.75, 196.36], [686.43, 196.39], [686.31, 197.25], [678.34, 197.22], [678.75, 169.91], [686.43, 169.91], [686.31, 169.8], [678.34, 169.8]]}]\n```" - } - ] - }, - { - "id": 2547, - "data_source": "Mono3DRefer", - "scene_name": "002891", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[340.46, 209.99], [297.84, 210.01], [343.22, 204.47], [379.46, 204.45], [340.46, 150.85], [297.84, 150.84], [343.22, 154.12], [379.46, 154.13]]}, {\"category\": \"truck\", \"corners_3d\": [[385.26, 204.88], [350.37, 204.88], [382.63, 200.89], [413.17, 200.89], [385.26, 155.17], [350.37, 155.17], [382.63, 157.38], [413.17, 157.38]]}]\n```" - } - ] - }, - { - "id": 2548, - "data_source": "Mono3DRefer", - "scene_name": "002891", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[445.24, 202.55], [421.51, 202.56], [434.88, 200.36], [456.84, 200.35], [445.24, 183.04], [421.51, 183.05], [434.88, 182.29], [456.84, 182.29]]}]\n```" - } - ] - }, - { - "id": 2549, - "data_source": "Mono3DRefer", - "scene_name": "002891", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[445.24, 202.55], [421.51, 202.56], [434.88, 200.36], [456.84, 200.35], [445.24, 183.04], [421.51, 183.05], [434.88, 182.29], [456.84, 182.29]]}, {\"category\": \"truck\", \"corners_3d\": [[340.46, 209.99], [297.84, 210.01], [343.22, 204.47], [379.46, 204.45], [340.46, 150.85], [297.84, 150.84], [343.22, 154.12], [379.46, 154.13]]}, {\"category\": \"truck\", \"corners_3d\": [[385.26, 204.88], [350.37, 204.88], [382.63, 200.89], [413.17, 200.89], [385.26, 155.17], [350.37, 155.17], [382.63, 157.38], [413.17, 157.38]]}]\n```" - } - ] - }, - { - "id": 2550, - "data_source": "Mono3DRefer", - "scene_name": "002893", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[140.23, 349.36], [204.99, 340.52], [219.62, 351.31], [151.28, 361.36], [140.23, 162.82], [204.99, 163.32], [219.62, 162.71], [151.28, 162.14]]}]\n```" - } - ] - }, - { - "id": 2551, - "data_source": "Mono3DRefer", - "scene_name": "002893", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[792.31, 244.66], [852.55, 245.14], [910.53, 267.84], [831.04, 267.02], [792.31, 186.59], [852.55, 186.68], [910.53, 191.03], [831.04, 190.87]]}, {\"category\": \"car\", \"corners_3d\": [[-33.06, 323.33], [96.26, 298.46], [336.24, 307.86], [246.99, 337.02], [-33.06, 163.79], [96.26, 165.29], [336.24, 164.72], [246.99, 162.96]]}]\n```" - } - ] - }, - { - "id": 2552, - "data_source": "Mono3DRefer", - "scene_name": "002893", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[745.8, 212.56], [783.24, 212.8], [795.47, 218.48], [752.67, 218.16], [745.8, 170.19], [783.24, 170.17], [795.47, 169.79], [752.67, 169.81]]}]\n```" - } - ] - }, - { - "id": 2553, - "data_source": "Mono3DRefer", - "scene_name": "002893", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[792.31, 244.66], [852.55, 245.14], [910.53, 267.84], [831.04, 267.02], [792.31, 186.59], [852.55, 186.68], [910.53, 191.03], [831.04, 190.87]]}, {\"category\": \"car\", \"corners_3d\": [[-33.06, 323.33], [96.26, 298.46], [336.24, 307.86], [246.99, 337.02], [-33.06, 163.79], [96.26, 165.29], [336.24, 164.72], [246.99, 162.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[140.23, 349.36], [204.99, 340.52], [219.62, 351.31], [151.28, 361.36], [140.23, 162.82], [204.99, 163.32], [219.62, 162.71], [151.28, 162.14]]}]\n```" - } - ] - }, - { - "id": 2554, - "data_source": "Mono3DRefer", - "scene_name": "002893", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[140.23, 349.36], [204.99, 340.52], [219.62, 351.31], [151.28, 361.36], [140.23, 162.82], [204.99, 163.32], [219.62, 162.71], [151.28, 162.14]]}, {\"category\": \"van\", \"corners_3d\": [[745.8, 212.56], [783.24, 212.8], [795.47, 218.48], [752.67, 218.16], [745.8, 170.19], [783.24, 170.17], [795.47, 169.79], [752.67, 169.81]]}]\n```" - } - ] - }, - { - "id": 2555, - "data_source": "Mono3DRefer", - "scene_name": "002893", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[792.31, 244.66], [852.55, 245.14], [910.53, 267.84], [831.04, 267.02], [792.31, 186.59], [852.55, 186.68], [910.53, 191.03], [831.04, 190.87]]}, {\"category\": \"car\", \"corners_3d\": [[-33.06, 323.33], [96.26, 298.46], [336.24, 307.86], [246.99, 337.02], [-33.06, 163.79], [96.26, 165.29], [336.24, 164.72], [246.99, 162.96]]}, {\"category\": \"van\", \"corners_3d\": [[745.8, 212.56], [783.24, 212.8], [795.47, 218.48], [752.67, 218.16], [745.8, 170.19], [783.24, 170.17], [795.47, 169.79], [752.67, 169.81]]}]\n```" - } - ] - }, - { - "id": 2556, - "data_source": "Mono3DRefer", - "scene_name": "002893", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[792.31, 244.66], [852.55, 245.14], [910.53, 267.84], [831.04, 267.02], [792.31, 186.59], [852.55, 186.68], [910.53, 191.03], [831.04, 190.87]]}, {\"category\": \"car\", \"corners_3d\": [[-33.06, 323.33], [96.26, 298.46], [336.24, 307.86], [246.99, 337.02], [-33.06, 163.79], [96.26, 165.29], [336.24, 164.72], [246.99, 162.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[140.23, 349.36], [204.99, 340.52], [219.62, 351.31], [151.28, 361.36], [140.23, 162.82], [204.99, 163.32], [219.62, 162.71], [151.28, 162.14]]}, {\"category\": \"van\", \"corners_3d\": [[745.8, 212.56], [783.24, 212.8], [795.47, 218.48], [752.67, 218.16], [745.8, 170.19], [783.24, 170.17], [795.47, 169.79], [752.67, 169.81]]}]\n```" - } - ] - }, - { - "id": 2557, - "data_source": "Mono3DRefer", - "scene_name": "002898", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[235.18, 298.01], [255.81, 292.14], [338.69, 292.97], [322.3, 298.94], [235.18, 156.33], [255.81, 157.54], [338.69, 157.37], [322.3, 156.14]]}]\n```" - } - ] - }, - { - "id": 2558, - "data_source": "Mono3DRefer", - "scene_name": "002901", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[251.44, 211.14], [240.92, 211.06], [259.12, 209.56], [269.25, 209.63], [251.44, 178.71], [240.92, 178.7], [259.12, 178.47], [269.25, 178.48]]}]\n```" - } - ] - }, - { - "id": 2559, - "data_source": "Mono3DRefer", - "scene_name": "002901", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[257.77, 221.26], [218.4, 221.12], [260.64, 216.34], [296.21, 216.45], [257.77, 187.51], [218.4, 187.47], [260.64, 186.02], [296.21, 186.05]]}]\n```" - } - ] - }, - { - "id": 2560, - "data_source": "Mono3DRefer", - "scene_name": "002901", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[328.31, 204.17], [299.78, 204.14], [330.15, 201.21], [356.02, 201.23], [328.31, 171.5], [299.78, 171.5], [330.15, 171.63], [356.02, 171.63]]}]\n```" - } - ] - }, - { - "id": 2561, - "data_source": "Mono3DRefer", - "scene_name": "002901", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[251.44, 211.14], [240.92, 211.06], [259.12, 209.56], [269.25, 209.63], [251.44, 178.71], [240.92, 178.7], [259.12, 178.47], [269.25, 178.48]]}, {\"category\": \"car\", \"corners_3d\": [[257.77, 221.26], [218.4, 221.12], [260.64, 216.34], [296.21, 216.45], [257.77, 187.51], [218.4, 187.47], [260.64, 186.02], [296.21, 186.05]]}]\n```" - } - ] - }, - { - "id": 2562, - "data_source": "Mono3DRefer", - "scene_name": "002901", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[328.31, 204.17], [299.78, 204.14], [330.15, 201.21], [356.02, 201.23], [328.31, 171.5], [299.78, 171.5], [330.15, 171.63], [356.02, 171.63]]}, {\"category\": \"cyclist\", \"corners_3d\": [[251.44, 211.14], [240.92, 211.06], [259.12, 209.56], [269.25, 209.63], [251.44, 178.71], [240.92, 178.7], [259.12, 178.47], [269.25, 178.48]]}]\n```" - } - ] - }, - { - "id": 2563, - "data_source": "Mono3DRefer", - "scene_name": "002901", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[328.31, 204.17], [299.78, 204.14], [330.15, 201.21], [356.02, 201.23], [328.31, 171.5], [299.78, 171.5], [330.15, 171.63], [356.02, 171.63]]}, {\"category\": \"car\", \"corners_3d\": [[257.77, 221.26], [218.4, 221.12], [260.64, 216.34], [296.21, 216.45], [257.77, 187.51], [218.4, 187.47], [260.64, 186.02], [296.21, 186.05]]}]\n```" - } - ] - }, - { - "id": 2564, - "data_source": "Mono3DRefer", - "scene_name": "002901", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[328.31, 204.17], [299.78, 204.14], [330.15, 201.21], [356.02, 201.23], [328.31, 171.5], [299.78, 171.5], [330.15, 171.63], [356.02, 171.63]]}, {\"category\": \"cyclist\", \"corners_3d\": [[251.44, 211.14], [240.92, 211.06], [259.12, 209.56], [269.25, 209.63], [251.44, 178.71], [240.92, 178.7], [259.12, 178.47], [269.25, 178.48]]}, {\"category\": \"car\", \"corners_3d\": [[257.77, 221.26], [218.4, 221.12], [260.64, 216.34], [296.21, 216.45], [257.77, 187.51], [218.4, 187.47], [260.64, 186.02], [296.21, 186.05]]}]\n```" - } - ] - }, - { - "id": 2565, - "data_source": "Mono3DRefer", - "scene_name": "002908", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[292.0, 202.28], [254.65, 202.35], [288.73, 199.26], [322.11, 199.21], [292.0, 173.91], [254.65, 173.91], [288.73, 173.8], [322.11, 173.8]]}]\n```" - } - ] - }, - { - "id": 2566, - "data_source": "Mono3DRefer", - "scene_name": "002913", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-104.64, 253.32], [-173.46, 253.01], [-2.96, 236.16], [51.88, 236.36], [-104.64, 173.29], [-173.46, 173.29], [-2.96, 173.2], [51.88, 173.2]]}, {\"category\": \"car\", \"corners_3d\": [[91.59, 233.05], [45.87, 232.73], [134.04, 224.23], [173.65, 224.47], [91.59, 183.15], [45.87, 183.1], [134.04, 181.64], [173.65, 181.69]]}]\n```" - } - ] - }, - { - "id": 2567, - "data_source": "Mono3DRefer", - "scene_name": "002914", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[788.78, 221.7], [812.72, 221.7], [820.56, 223.6], [795.69, 223.59], [788.78, 165.63], [812.72, 165.63], [820.56, 165.35], [795.69, 165.35]]}]\n```" - } - ] - }, - { - "id": 2568, - "data_source": "Mono3DRefer", - "scene_name": "002914", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[678.19, 195.55], [685.65, 195.58], [685.24, 196.39], [677.51, 196.36], [678.19, 169.85], [685.65, 169.85], [685.24, 169.74], [677.51, 169.74]]}, {\"category\": \"cyclist\", \"corners_3d\": [[691.04, 204.59], [706.83, 204.65], [709.06, 206.48], [692.36, 206.41], [691.04, 165.93], [706.83, 165.92], [709.06, 165.52], [692.36, 165.53]]}]\n```" - } - ] - }, - { - "id": 2569, - "data_source": "Mono3DRefer", - "scene_name": "002914", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[678.19, 195.55], [685.65, 195.58], [685.24, 196.39], [677.51, 196.36], [678.19, 169.85], [685.65, 169.85], [685.24, 169.74], [677.51, 169.74]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[788.78, 221.7], [812.72, 221.7], [820.56, 223.6], [795.69, 223.59], [788.78, 165.63], [812.72, 165.63], [820.56, 165.35], [795.69, 165.35]]}, {\"category\": \"cyclist\", \"corners_3d\": [[691.04, 204.59], [706.83, 204.65], [709.06, 206.48], [692.36, 206.41], [691.04, 165.93], [706.83, 165.92], [709.06, 165.52], [692.36, 165.53]]}]\n```" - } - ] - }, - { - "id": 2570, - "data_source": "Mono3DRefer", - "scene_name": "002916", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[548.17, 189.19], [569.66, 189.17], [568.38, 190.28], [545.43, 190.3], [548.17, 167.45], [569.66, 167.45], [568.38, 167.09], [545.43, 167.08]]}, {\"category\": \"car\", \"corners_3d\": [[519.05, 241.2], [451.9, 240.89], [485.12, 229.19], [540.81, 229.4], [519.05, 181.24], [451.9, 181.2], [485.12, 179.76], [540.81, 179.79]]}]\n```" - } - ] - }, - { - "id": 2571, - "data_source": "Mono3DRefer", - "scene_name": "002928", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[697.17, 181.1], [716.57, 181.21], [698.17, 181.82], [677.61, 181.7], [697.17, 154.71], [716.57, 154.48], [698.17, 153.14], [677.61, 153.4]]}]\n```" - } - ] - }, - { - "id": 2572, - "data_source": "Mono3DRefer", - "scene_name": "002928", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[800.42, 346.2], [971.34, 344.57], [1522.62, 587.26], [1117.33, 596.86], [800.42, 171.72], [971.34, 171.73], [1522.62, 170.14], [1117.33, 170.07]]}, {\"category\": \"car\", \"corners_3d\": [[638.05, 200.41], [671.16, 200.5], [672.3, 203.5], [635.6, 203.39], [638.05, 172.65], [671.16, 172.64], [672.3, 172.62], [635.6, 172.62]]}, {\"category\": \"car\", \"corners_3d\": [[88.63, 395.79], [326.6, 393.69], [-562.93, 1143.82], [-1682.72, 1185.84], [88.63, 212.72], [326.6, 212.35], [-562.93, 346.49], [-1682.72, 354.01]]}, {\"category\": \"car\", \"corners_3d\": [[669.63, 228.14], [728.02, 228.01], [768.68, 244.09], [693.36, 244.31], [669.63, 172.09], [728.02, 172.09], [768.68, 171.87], [693.36, 171.86]]}, {\"category\": \"car\", \"corners_3d\": [[408.56, 253.74], [487.37, 253.4], [448.85, 286.44], [337.29, 287.12], [408.56, 181.03], [487.37, 180.99], [448.85, 184.33], [337.29, 184.4]]}]\n```" - } - ] - }, - { - "id": 2573, - "data_source": "Mono3DRefer", - "scene_name": "002928", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[800.42, 346.2], [971.34, 344.57], [1522.62, 587.26], [1117.33, 596.86], [800.42, 171.72], [971.34, 171.73], [1522.62, 170.14], [1117.33, 170.07]]}, {\"category\": \"car\", \"corners_3d\": [[638.05, 200.41], [671.16, 200.5], [672.3, 203.5], [635.6, 203.39], [638.05, 172.65], [671.16, 172.64], [672.3, 172.62], [635.6, 172.62]]}, {\"category\": \"van\", \"corners_3d\": [[697.17, 181.1], [716.57, 181.21], [698.17, 181.82], [677.61, 181.7], [697.17, 154.71], [716.57, 154.48], [698.17, 153.14], [677.61, 153.4]]}, {\"category\": \"car\", \"corners_3d\": [[88.63, 395.79], [326.6, 393.69], [-562.93, 1143.82], [-1682.72, 1185.84], [88.63, 212.72], [326.6, 212.35], [-562.93, 346.49], [-1682.72, 354.01]]}, {\"category\": \"car\", \"corners_3d\": [[669.63, 228.14], [728.02, 228.01], [768.68, 244.09], [693.36, 244.31], [669.63, 172.09], [728.02, 172.09], [768.68, 171.87], [693.36, 171.86]]}, {\"category\": \"car\", \"corners_3d\": [[408.56, 253.74], [487.37, 253.4], [448.85, 286.44], [337.29, 287.12], [408.56, 181.03], [487.37, 180.99], [448.85, 184.33], [337.29, 184.4]]}]\n```" - } - ] - }, - { - "id": 2574, - "data_source": "Mono3DRefer", - "scene_name": "002932", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[311.62, 384.91], [108.84, 384.96], [310.61, 299.39], [431.57, 299.37], [311.62, 215.23], [108.84, 215.24], [310.61, 198.14], [431.57, 198.14]]}]\n```" - } - ] - }, - { - "id": 2575, - "data_source": "Mono3DRefer", - "scene_name": "002942", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[306.87, 366.34], [110.35, 369.97], [291.54, 290.25], [407.46, 288.96], [306.87, 200.3], [110.35, 200.81], [291.54, 189.5], [407.46, 189.32]]}, {\"category\": \"car\", \"corners_3d\": [[744.31, 245.38], [677.79, 245.79], [656.85, 232.17], [711.1, 231.9], [744.31, 176.66], [677.79, 176.68], [656.85, 175.96], [711.1, 175.95]]}, {\"category\": \"car\", \"corners_3d\": [[517.61, 229.05], [465.64, 229.09], [487.45, 220.08], [531.08, 220.05], [517.61, 184.02], [465.64, 184.03], [487.45, 182.24], [531.08, 182.23]]}, {\"category\": \"car\", \"corners_3d\": [[637.88, 221.5], [679.17, 221.25], [699.35, 228.53], [651.93, 228.86], [637.88, 182.84], [679.17, 182.79], [699.35, 184.28], [651.93, 184.35]]}]\n```" - } - ] - }, - { - "id": 2576, - "data_source": "Mono3DRefer", - "scene_name": "002944", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[141.95, 425.07], [206.77, 400.8], [319.48, 411.6], [264.25, 438.45], [141.95, 240.71], [206.77, 234.81], [319.48, 237.43], [264.25, 243.96]]}]\n```" - } - ] - }, - { - "id": 2577, - "data_source": "Mono3DRefer", - "scene_name": "002944", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[500.21, 320.53], [438.96, 320.54], [455.71, 306.07], [510.58, 306.06], [500.21, 156.33], [438.96, 156.32], [455.71, 158.95], [510.58, 158.95]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[450.16, 329.36], [378.57, 329.08], [401.06, 314.99], [465.84, 315.22], [450.16, 164.59], [378.57, 164.62], [401.06, 166.24], [465.84, 166.21]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[628.94, 235.14], [662.22, 235.44], [660.55, 239.0], [625.09, 238.66], [628.94, 158.31], [662.22, 158.18], [660.55, 156.64], [625.09, 156.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[623.07, 213.09], [651.24, 213.16], [651.82, 214.43], [622.51, 214.36], [623.07, 157.75], [651.24, 157.7], [651.82, 156.74], [622.51, 156.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[675.47, 235.9], [704.3, 236.09], [706.6, 239.36], [676.05, 239.14], [675.47, 160.92], [704.3, 160.84], [706.6, 159.6], [676.05, 159.69]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[723.88, 218.98], [753.54, 219.06], [759.08, 220.84], [728.01, 220.76], [723.88, 155.26], [753.54, 155.21], [759.08, 153.96], [728.01, 154.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[741.82, 210.05], [766.41, 210.24], [767.23, 211.43], [741.62, 211.22], [741.82, 153.8], [766.41, 153.62], [767.23, 152.47], [741.62, 152.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[375.43, 245.07], [399.49, 242.67], [426.33, 244.24], [402.71, 246.76], [375.43, 164.54], [399.49, 165.18], [426.33, 164.76], [402.71, 164.09]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[700.24, 226.32], [727.77, 226.45], [731.46, 228.85], [702.46, 228.7], [700.24, 162.28], [727.77, 162.22], [731.46, 161.19], [702.46, 161.26]]}]\n```" - } - ] - }, - { - "id": 2578, - "data_source": "Mono3DRefer", - "scene_name": "002944", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[592.74, 194.65], [564.03, 194.61], [572.45, 193.42], [598.59, 193.46], [592.74, 167.46], [564.03, 167.5], [572.45, 168.77], [598.59, 168.73]]}]\n```" - } - ] - }, - { - "id": 2579, - "data_source": "Mono3DRefer", - "scene_name": "002944", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[141.95, 425.07], [206.77, 400.8], [319.48, 411.6], [264.25, 438.45], [141.95, 240.71], [206.77, 234.81], [319.48, 237.43], [264.25, 243.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[500.21, 320.53], [438.96, 320.54], [455.71, 306.07], [510.58, 306.06], [500.21, 156.33], [438.96, 156.32], [455.71, 158.95], [510.58, 158.95]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[450.16, 329.36], [378.57, 329.08], [401.06, 314.99], [465.84, 315.22], [450.16, 164.59], [378.57, 164.62], [401.06, 166.24], [465.84, 166.21]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[628.94, 235.14], [662.22, 235.44], [660.55, 239.0], [625.09, 238.66], [628.94, 158.31], [662.22, 158.18], [660.55, 156.64], [625.09, 156.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[623.07, 213.09], [651.24, 213.16], [651.82, 214.43], [622.51, 214.36], [623.07, 157.75], [651.24, 157.7], [651.82, 156.74], [622.51, 156.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[675.47, 235.9], [704.3, 236.09], [706.6, 239.36], [676.05, 239.14], [675.47, 160.92], [704.3, 160.84], [706.6, 159.6], [676.05, 159.69]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[723.88, 218.98], [753.54, 219.06], [759.08, 220.84], [728.01, 220.76], [723.88, 155.26], [753.54, 155.21], [759.08, 153.96], [728.01, 154.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[741.82, 210.05], [766.41, 210.24], [767.23, 211.43], [741.62, 211.22], [741.82, 153.8], [766.41, 153.62], [767.23, 152.47], [741.62, 152.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[375.43, 245.07], [399.49, 242.67], [426.33, 244.24], [402.71, 246.76], [375.43, 164.54], [399.49, 165.18], [426.33, 164.76], [402.71, 164.09]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[700.24, 226.32], [727.77, 226.45], [731.46, 228.85], [702.46, 228.7], [700.24, 162.28], [727.77, 162.22], [731.46, 161.19], [702.46, 161.26]]}]\n```" - } - ] - }, - { - "id": 2580, - "data_source": "Mono3DRefer", - "scene_name": "002944", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[141.95, 425.07], [206.77, 400.8], [319.48, 411.6], [264.25, 438.45], [141.95, 240.71], [206.77, 234.81], [319.48, 237.43], [264.25, 243.96]]}, {\"category\": \"car\", \"corners_3d\": [[592.74, 194.65], [564.03, 194.61], [572.45, 193.42], [598.59, 193.46], [592.74, 167.46], [564.03, 167.5], [572.45, 168.77], [598.59, 168.73]]}]\n```" - } - ] - }, - { - "id": 2581, - "data_source": "Mono3DRefer", - "scene_name": "002944", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[500.21, 320.53], [438.96, 320.54], [455.71, 306.07], [510.58, 306.06], [500.21, 156.33], [438.96, 156.32], [455.71, 158.95], [510.58, 158.95]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[450.16, 329.36], [378.57, 329.08], [401.06, 314.99], [465.84, 315.22], [450.16, 164.59], [378.57, 164.62], [401.06, 166.24], [465.84, 166.21]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[628.94, 235.14], [662.22, 235.44], [660.55, 239.0], [625.09, 238.66], [628.94, 158.31], [662.22, 158.18], [660.55, 156.64], [625.09, 156.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[623.07, 213.09], [651.24, 213.16], [651.82, 214.43], [622.51, 214.36], [623.07, 157.75], [651.24, 157.7], [651.82, 156.74], [622.51, 156.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[675.47, 235.9], [704.3, 236.09], [706.6, 239.36], [676.05, 239.14], [675.47, 160.92], [704.3, 160.84], [706.6, 159.6], [676.05, 159.69]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[723.88, 218.98], [753.54, 219.06], [759.08, 220.84], [728.01, 220.76], [723.88, 155.26], [753.54, 155.21], [759.08, 153.96], [728.01, 154.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[741.82, 210.05], [766.41, 210.24], [767.23, 211.43], [741.62, 211.22], [741.82, 153.8], [766.41, 153.62], [767.23, 152.47], [741.62, 152.67]]}, {\"category\": \"car\", \"corners_3d\": [[592.74, 194.65], [564.03, 194.61], [572.45, 193.42], [598.59, 193.46], [592.74, 167.46], [564.03, 167.5], [572.45, 168.77], [598.59, 168.73]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[375.43, 245.07], [399.49, 242.67], [426.33, 244.24], [402.71, 246.76], [375.43, 164.54], [399.49, 165.18], [426.33, 164.76], [402.71, 164.09]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[700.24, 226.32], [727.77, 226.45], [731.46, 228.85], [702.46, 228.7], [700.24, 162.28], [727.77, 162.22], [731.46, 161.19], [702.46, 161.26]]}]\n```" - } - ] - }, - { - "id": 2582, - "data_source": "Mono3DRefer", - "scene_name": "002944", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[141.95, 425.07], [206.77, 400.8], [319.48, 411.6], [264.25, 438.45], [141.95, 240.71], [206.77, 234.81], [319.48, 237.43], [264.25, 243.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[500.21, 320.53], [438.96, 320.54], [455.71, 306.07], [510.58, 306.06], [500.21, 156.33], [438.96, 156.32], [455.71, 158.95], [510.58, 158.95]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[450.16, 329.36], [378.57, 329.08], [401.06, 314.99], [465.84, 315.22], [450.16, 164.59], [378.57, 164.62], [401.06, 166.24], [465.84, 166.21]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[628.94, 235.14], [662.22, 235.44], [660.55, 239.0], [625.09, 238.66], [628.94, 158.31], [662.22, 158.18], [660.55, 156.64], [625.09, 156.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[623.07, 213.09], [651.24, 213.16], [651.82, 214.43], [622.51, 214.36], [623.07, 157.75], [651.24, 157.7], [651.82, 156.74], [622.51, 156.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[675.47, 235.9], [704.3, 236.09], [706.6, 239.36], [676.05, 239.14], [675.47, 160.92], [704.3, 160.84], [706.6, 159.6], [676.05, 159.69]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[723.88, 218.98], [753.54, 219.06], [759.08, 220.84], [728.01, 220.76], [723.88, 155.26], [753.54, 155.21], [759.08, 153.96], [728.01, 154.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[741.82, 210.05], [766.41, 210.24], [767.23, 211.43], [741.62, 211.22], [741.82, 153.8], [766.41, 153.62], [767.23, 152.47], [741.62, 152.67]]}, {\"category\": \"car\", \"corners_3d\": [[592.74, 194.65], [564.03, 194.61], [572.45, 193.42], [598.59, 193.46], [592.74, 167.46], [564.03, 167.5], [572.45, 168.77], [598.59, 168.73]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[375.43, 245.07], [399.49, 242.67], [426.33, 244.24], [402.71, 246.76], [375.43, 164.54], [399.49, 165.18], [426.33, 164.76], [402.71, 164.09]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[700.24, 226.32], [727.77, 226.45], [731.46, 228.85], [702.46, 228.7], [700.24, 162.28], [727.77, 162.22], [731.46, 161.19], [702.46, 161.26]]}]\n```" - } - ] - }, - { - "id": 2583, - "data_source": "Mono3DRefer", - "scene_name": "002953", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1184.31, 353.72], [1292.61, 361.23], [1360.82, 391.3], [1232.89, 381.27], [1184.31, 105.76], [1292.61, 102.97], [1360.82, 91.82], [1232.89, 95.54]]}]\n```" - } - ] - }, - { - "id": 2584, - "data_source": "Mono3DRefer", - "scene_name": "002953", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-142.01, 258.58], [-86.52, 252.38], [33.27, 252.67], [-13.07, 258.92], [-142.01, 215.55], [-86.52, 212.46], [33.27, 212.6], [-13.07, 215.72]]}, {\"category\": \"car\", \"corners_3d\": [[869.0, 212.91], [853.69, 210.53], [955.47, 210.54], [977.2, 212.92], [869.0, 169.03], [853.69, 169.25], [955.47, 169.25], [977.2, 169.03]]}, {\"category\": \"car\", \"corners_3d\": [[1027.58, 209.33], [994.65, 206.65], [1129.74, 206.38], [1173.02, 209.02], [1027.58, 158.76], [994.65, 159.79], [1129.74, 159.9], [1173.02, 158.88]]}, {\"category\": \"car\", \"corners_3d\": [[324.22, 222.64], [287.91, 224.48], [252.97, 220.21], [287.52, 218.66], [324.22, 187.6], [287.91, 188.14], [252.97, 186.88], [287.52, 186.42]]}]\n```" - } - ] - }, - { - "id": 2585, - "data_source": "Mono3DRefer", - "scene_name": "002953", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[726.98, 193.35], [715.83, 191.72], [903.17, 191.56], [930.24, 193.17], [726.98, 106.37], [715.83, 111.65], [903.17, 112.16], [930.24, 106.96]]}]\n```" - } - ] - }, - { - "id": 2586, - "data_source": "Mono3DRefer", - "scene_name": "002953", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[103.69, 248.06], [145.72, 242.08], [283.65, 242.51], [253.33, 248.56], [103.69, 183.73], [145.72, 182.86], [283.65, 182.93], [253.33, 183.8]]}]\n```" - } - ] - }, - { - "id": 2587, - "data_source": "Mono3DRefer", - "scene_name": "002953", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1184.31, 353.72], [1292.61, 361.23], [1360.82, 391.3], [1232.89, 381.27], [1184.31, 105.76], [1292.61, 102.97], [1360.82, 91.82], [1232.89, 95.54]]}, {\"category\": \"car\", \"corners_3d\": [[-142.01, 258.58], [-86.52, 252.38], [33.27, 252.67], [-13.07, 258.92], [-142.01, 215.55], [-86.52, 212.46], [33.27, 212.6], [-13.07, 215.72]]}, {\"category\": \"car\", \"corners_3d\": [[869.0, 212.91], [853.69, 210.53], [955.47, 210.54], [977.2, 212.92], [869.0, 169.03], [853.69, 169.25], [955.47, 169.25], [977.2, 169.03]]}, {\"category\": \"car\", \"corners_3d\": [[1027.58, 209.33], [994.65, 206.65], [1129.74, 206.38], [1173.02, 209.02], [1027.58, 158.76], [994.65, 159.79], [1129.74, 159.9], [1173.02, 158.88]]}, {\"category\": \"car\", \"corners_3d\": [[324.22, 222.64], [287.91, 224.48], [252.97, 220.21], [287.52, 218.66], [324.22, 187.6], [287.91, 188.14], [252.97, 186.88], [287.52, 186.42]]}]\n```" - } - ] - }, - { - "id": 2588, - "data_source": "Mono3DRefer", - "scene_name": "002953", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1184.31, 353.72], [1292.61, 361.23], [1360.82, 391.3], [1232.89, 381.27], [1184.31, 105.76], [1292.61, 102.97], [1360.82, 91.82], [1232.89, 95.54]]}, {\"category\": \"truck\", \"corners_3d\": [[726.98, 193.35], [715.83, 191.72], [903.17, 191.56], [930.24, 193.17], [726.98, 106.37], [715.83, 111.65], [903.17, 112.16], [930.24, 106.96]]}]\n```" - } - ] - }, - { - "id": 2589, - "data_source": "Mono3DRefer", - "scene_name": "002953", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1184.31, 353.72], [1292.61, 361.23], [1360.82, 391.3], [1232.89, 381.27], [1184.31, 105.76], [1292.61, 102.97], [1360.82, 91.82], [1232.89, 95.54]]}, {\"category\": \"van\", \"corners_3d\": [[103.69, 248.06], [145.72, 242.08], [283.65, 242.51], [253.33, 248.56], [103.69, 183.73], [145.72, 182.86], [283.65, 182.93], [253.33, 183.8]]}]\n```" - } - ] - }, - { - "id": 2590, - "data_source": "Mono3DRefer", - "scene_name": "002953", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-142.01, 258.58], [-86.52, 252.38], [33.27, 252.67], [-13.07, 258.92], [-142.01, 215.55], [-86.52, 212.46], [33.27, 212.6], [-13.07, 215.72]]}, {\"category\": \"car\", \"corners_3d\": [[869.0, 212.91], [853.69, 210.53], [955.47, 210.54], [977.2, 212.92], [869.0, 169.03], [853.69, 169.25], [955.47, 169.25], [977.2, 169.03]]}, {\"category\": \"truck\", \"corners_3d\": [[726.98, 193.35], [715.83, 191.72], [903.17, 191.56], [930.24, 193.17], [726.98, 106.37], [715.83, 111.65], [903.17, 112.16], [930.24, 106.96]]}, {\"category\": \"car\", \"corners_3d\": [[1027.58, 209.33], [994.65, 206.65], [1129.74, 206.38], [1173.02, 209.02], [1027.58, 158.76], [994.65, 159.79], [1129.74, 159.9], [1173.02, 158.88]]}, {\"category\": \"car\", \"corners_3d\": [[324.22, 222.64], [287.91, 224.48], [252.97, 220.21], [287.52, 218.66], [324.22, 187.6], [287.91, 188.14], [252.97, 186.88], [287.52, 186.42]]}]\n```" - } - ] - }, - { - "id": 2591, - "data_source": "Mono3DRefer", - "scene_name": "002953", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-142.01, 258.58], [-86.52, 252.38], [33.27, 252.67], [-13.07, 258.92], [-142.01, 215.55], [-86.52, 212.46], [33.27, 212.6], [-13.07, 215.72]]}, {\"category\": \"car\", \"corners_3d\": [[869.0, 212.91], [853.69, 210.53], [955.47, 210.54], [977.2, 212.92], [869.0, 169.03], [853.69, 169.25], [955.47, 169.25], [977.2, 169.03]]}, {\"category\": \"car\", \"corners_3d\": [[1027.58, 209.33], [994.65, 206.65], [1129.74, 206.38], [1173.02, 209.02], [1027.58, 158.76], [994.65, 159.79], [1129.74, 159.9], [1173.02, 158.88]]}, {\"category\": \"car\", \"corners_3d\": [[324.22, 222.64], [287.91, 224.48], [252.97, 220.21], [287.52, 218.66], [324.22, 187.6], [287.91, 188.14], [252.97, 186.88], [287.52, 186.42]]}, {\"category\": \"van\", \"corners_3d\": [[103.69, 248.06], [145.72, 242.08], [283.65, 242.51], [253.33, 248.56], [103.69, 183.73], [145.72, 182.86], [283.65, 182.93], [253.33, 183.8]]}]\n```" - } - ] - }, - { - "id": 2592, - "data_source": "Mono3DRefer", - "scene_name": "002953", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[726.98, 193.35], [715.83, 191.72], [903.17, 191.56], [930.24, 193.17], [726.98, 106.37], [715.83, 111.65], [903.17, 112.16], [930.24, 106.96]]}, {\"category\": \"van\", \"corners_3d\": [[103.69, 248.06], [145.72, 242.08], [283.65, 242.51], [253.33, 248.56], [103.69, 183.73], [145.72, 182.86], [283.65, 182.93], [253.33, 183.8]]}]\n```" - } - ] - }, - { - "id": 2593, - "data_source": "Mono3DRefer", - "scene_name": "002953", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1184.31, 353.72], [1292.61, 361.23], [1360.82, 391.3], [1232.89, 381.27], [1184.31, 105.76], [1292.61, 102.97], [1360.82, 91.82], [1232.89, 95.54]]}, {\"category\": \"car\", \"corners_3d\": [[-142.01, 258.58], [-86.52, 252.38], [33.27, 252.67], [-13.07, 258.92], [-142.01, 215.55], [-86.52, 212.46], [33.27, 212.6], [-13.07, 215.72]]}, {\"category\": \"car\", \"corners_3d\": [[869.0, 212.91], [853.69, 210.53], [955.47, 210.54], [977.2, 212.92], [869.0, 169.03], [853.69, 169.25], [955.47, 169.25], [977.2, 169.03]]}, {\"category\": \"truck\", \"corners_3d\": [[726.98, 193.35], [715.83, 191.72], [903.17, 191.56], [930.24, 193.17], [726.98, 106.37], [715.83, 111.65], [903.17, 112.16], [930.24, 106.96]]}, {\"category\": \"car\", \"corners_3d\": [[1027.58, 209.33], [994.65, 206.65], [1129.74, 206.38], [1173.02, 209.02], [1027.58, 158.76], [994.65, 159.79], [1129.74, 159.9], [1173.02, 158.88]]}, {\"category\": \"car\", \"corners_3d\": [[324.22, 222.64], [287.91, 224.48], [252.97, 220.21], [287.52, 218.66], [324.22, 187.6], [287.91, 188.14], [252.97, 186.88], [287.52, 186.42]]}]\n```" - } - ] - }, - { - "id": 2594, - "data_source": "Mono3DRefer", - "scene_name": "002953", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1184.31, 353.72], [1292.61, 361.23], [1360.82, 391.3], [1232.89, 381.27], [1184.31, 105.76], [1292.61, 102.97], [1360.82, 91.82], [1232.89, 95.54]]}, {\"category\": \"car\", \"corners_3d\": [[-142.01, 258.58], [-86.52, 252.38], [33.27, 252.67], [-13.07, 258.92], [-142.01, 215.55], [-86.52, 212.46], [33.27, 212.6], [-13.07, 215.72]]}, {\"category\": \"car\", \"corners_3d\": [[869.0, 212.91], [853.69, 210.53], [955.47, 210.54], [977.2, 212.92], [869.0, 169.03], [853.69, 169.25], [955.47, 169.25], [977.2, 169.03]]}, {\"category\": \"car\", \"corners_3d\": [[1027.58, 209.33], [994.65, 206.65], [1129.74, 206.38], [1173.02, 209.02], [1027.58, 158.76], [994.65, 159.79], [1129.74, 159.9], [1173.02, 158.88]]}, {\"category\": \"car\", \"corners_3d\": [[324.22, 222.64], [287.91, 224.48], [252.97, 220.21], [287.52, 218.66], [324.22, 187.6], [287.91, 188.14], [252.97, 186.88], [287.52, 186.42]]}, {\"category\": \"van\", \"corners_3d\": [[103.69, 248.06], [145.72, 242.08], [283.65, 242.51], [253.33, 248.56], [103.69, 183.73], [145.72, 182.86], [283.65, 182.93], [253.33, 183.8]]}]\n```" - } - ] - }, - { - "id": 2595, - "data_source": "Mono3DRefer", - "scene_name": "002953", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1184.31, 353.72], [1292.61, 361.23], [1360.82, 391.3], [1232.89, 381.27], [1184.31, 105.76], [1292.61, 102.97], [1360.82, 91.82], [1232.89, 95.54]]}, {\"category\": \"truck\", \"corners_3d\": [[726.98, 193.35], [715.83, 191.72], [903.17, 191.56], [930.24, 193.17], [726.98, 106.37], [715.83, 111.65], [903.17, 112.16], [930.24, 106.96]]}, {\"category\": \"van\", \"corners_3d\": [[103.69, 248.06], [145.72, 242.08], [283.65, 242.51], [253.33, 248.56], [103.69, 183.73], [145.72, 182.86], [283.65, 182.93], [253.33, 183.8]]}]\n```" - } - ] - }, - { - "id": 2596, - "data_source": "Mono3DRefer", - "scene_name": "002953", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-142.01, 258.58], [-86.52, 252.38], [33.27, 252.67], [-13.07, 258.92], [-142.01, 215.55], [-86.52, 212.46], [33.27, 212.6], [-13.07, 215.72]]}, {\"category\": \"car\", \"corners_3d\": [[869.0, 212.91], [853.69, 210.53], [955.47, 210.54], [977.2, 212.92], [869.0, 169.03], [853.69, 169.25], [955.47, 169.25], [977.2, 169.03]]}, {\"category\": \"truck\", \"corners_3d\": [[726.98, 193.35], [715.83, 191.72], [903.17, 191.56], [930.24, 193.17], [726.98, 106.37], [715.83, 111.65], [903.17, 112.16], [930.24, 106.96]]}, {\"category\": \"car\", \"corners_3d\": [[1027.58, 209.33], [994.65, 206.65], [1129.74, 206.38], [1173.02, 209.02], [1027.58, 158.76], [994.65, 159.79], [1129.74, 159.9], [1173.02, 158.88]]}, {\"category\": \"car\", \"corners_3d\": [[324.22, 222.64], [287.91, 224.48], [252.97, 220.21], [287.52, 218.66], [324.22, 187.6], [287.91, 188.14], [252.97, 186.88], [287.52, 186.42]]}, {\"category\": \"van\", \"corners_3d\": [[103.69, 248.06], [145.72, 242.08], [283.65, 242.51], [253.33, 248.56], [103.69, 183.73], [145.72, 182.86], [283.65, 182.93], [253.33, 183.8]]}]\n```" - } - ] - }, - { - "id": 2597, - "data_source": "Mono3DRefer", - "scene_name": "002953", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1184.31, 353.72], [1292.61, 361.23], [1360.82, 391.3], [1232.89, 381.27], [1184.31, 105.76], [1292.61, 102.97], [1360.82, 91.82], [1232.89, 95.54]]}, {\"category\": \"car\", \"corners_3d\": [[-142.01, 258.58], [-86.52, 252.38], [33.27, 252.67], [-13.07, 258.92], [-142.01, 215.55], [-86.52, 212.46], [33.27, 212.6], [-13.07, 215.72]]}, {\"category\": \"car\", \"corners_3d\": [[869.0, 212.91], [853.69, 210.53], [955.47, 210.54], [977.2, 212.92], [869.0, 169.03], [853.69, 169.25], [955.47, 169.25], [977.2, 169.03]]}, {\"category\": \"truck\", \"corners_3d\": [[726.98, 193.35], [715.83, 191.72], [903.17, 191.56], [930.24, 193.17], [726.98, 106.37], [715.83, 111.65], [903.17, 112.16], [930.24, 106.96]]}, {\"category\": \"car\", \"corners_3d\": [[1027.58, 209.33], [994.65, 206.65], [1129.74, 206.38], [1173.02, 209.02], [1027.58, 158.76], [994.65, 159.79], [1129.74, 159.9], [1173.02, 158.88]]}, {\"category\": \"car\", \"corners_3d\": [[324.22, 222.64], [287.91, 224.48], [252.97, 220.21], [287.52, 218.66], [324.22, 187.6], [287.91, 188.14], [252.97, 186.88], [287.52, 186.42]]}, {\"category\": \"van\", \"corners_3d\": [[103.69, 248.06], [145.72, 242.08], [283.65, 242.51], [253.33, 248.56], [103.69, 183.73], [145.72, 182.86], [283.65, 182.93], [253.33, 183.8]]}]\n```" - } - ] - }, - { - "id": 2598, - "data_source": "Mono3DRefer", - "scene_name": "002956", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[734.87, 338.06], [709.08, 324.41], [811.09, 319.33], [845.19, 332.0], [734.87, 144.81], [709.08, 147.9], [811.09, 149.05], [845.19, 146.18]]}]\n```" - } - ] - }, - { - "id": 2599, - "data_source": "Mono3DRefer", - "scene_name": "002967", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[253.7, 539.63], [-89.59, 541.5], [249.74, 360.56], [424.09, 360.08], [253.7, 239.48], [-89.59, 239.82], [249.74, 206.95], [424.09, 206.86]]}, {\"category\": \"car\", \"corners_3d\": [[831.81, 372.32], [1002.51, 369.47], [1453.11, 572.81], [1114.08, 584.77], [831.81, 195.61], [1002.51, 195.29], [1453.11, 218.49], [1114.08, 219.86]]}, {\"category\": \"car\", \"corners_3d\": [[715.55, 283.05], [814.09, 283.06], [911.6, 335.75], [765.94, 335.72], [715.55, 191.92], [814.09, 191.92], [911.6, 201.04], [765.94, 201.03]]}, {\"category\": \"car\", \"corners_3d\": [[540.89, 250.12], [481.47, 249.87], [506.89, 237.35], [556.7, 237.52], [540.89, 189.84], [481.47, 189.79], [506.89, 187.03], [556.7, 187.07]]}, {\"category\": \"car\", \"corners_3d\": [[578.13, 233.31], [527.99, 233.27], [541.24, 224.2], [583.85, 224.23], [578.13, 175.56], [527.99, 175.56], [541.24, 175.16], [583.85, 175.16]]}, {\"category\": \"car\", \"corners_3d\": [[602.12, 223.5], [561.5, 223.27], [574.65, 216.98], [610.23, 217.15], [602.12, 179.15], [561.5, 179.12], [574.65, 178.34], [610.23, 178.36]]}, {\"category\": \"car\", \"corners_3d\": [[643.96, 209.54], [618.94, 209.17], [634.94, 206.34], [658.15, 206.65], [643.96, 185.45], [618.94, 185.32], [634.94, 184.35], [658.15, 184.46]]}]\n```" - } - ] - }, - { - "id": 2600, - "data_source": "Mono3DRefer", - "scene_name": "002972", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1081.29, 238.12], [1144.6, 236.19], [1371.22, 255.75], [1296.7, 259.08], [1081.29, 140.19], [1144.6, 141.16], [1371.22, 131.37], [1296.7, 129.71]]}, {\"category\": \"van\", \"corners_3d\": [[938.97, 226.19], [1010.2, 225.15], [1189.92, 242.83], [1098.77, 244.71], [938.97, 140.64], [1010.2, 141.27], [1189.92, 130.59], [1098.77, 129.46]]}, {\"category\": \"van\", \"corners_3d\": [[572.27, 207.63], [534.9, 207.5], [549.0, 203.58], [582.17, 203.68], [572.27, 165.89], [534.9, 165.92], [549.0, 166.7], [582.17, 166.68]]}]\n```" - } - ] - }, - { - "id": 2601, - "data_source": "Mono3DRefer", - "scene_name": "002972", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[157.49, 541.3], [-156.17, 520.56], [303.14, 333.14], [455.47, 337.41], [157.49, 201.46], [-156.17, 199.85], [303.14, 185.3], [455.47, 185.63]]}, {\"category\": \"car\", \"corners_3d\": [[840.49, 351.15], [1025.36, 352.1], [1464.37, 548.43], [1074.41, 544.31], [840.49, 178.29], [1025.36, 178.31], [1464.37, 184.29], [1074.41, 184.17]]}, {\"category\": \"car\", \"corners_3d\": [[1220.52, 222.92], [1285.72, 221.12], [1550.52, 235.46], [1477.95, 238.53], [1220.52, 133.45], [1285.72, 134.87], [1550.52, 123.58], [1477.95, 121.17]]}, {\"category\": \"car\", \"corners_3d\": [[455.14, 268.49], [362.58, 268.62], [415.45, 247.45], [487.51, 247.37], [455.14, 182.0], [362.58, 182.01], [415.45, 179.99], [487.51, 179.98]]}, {\"category\": \"car\", \"corners_3d\": [[498.92, 241.73], [433.8, 241.29], [477.16, 228.14], [529.93, 228.43], [498.92, 179.48], [433.8, 179.44], [477.16, 178.17], [529.93, 178.2]]}, {\"category\": \"car\", \"corners_3d\": [[967.22, 215.25], [916.52, 212.88], [971.21, 209.46], [1020.28, 211.42], [967.22, 158.6], [916.52, 159.39], [971.21, 160.54], [1020.28, 159.88]]}, {\"category\": \"car\", \"corners_3d\": [[884.5, 211.64], [842.47, 209.64], [896.81, 206.91], [938.3, 208.62], [884.5, 164.46], [842.47, 164.9], [896.81, 165.49], [938.3, 165.12]]}, {\"category\": \"car\", \"corners_3d\": [[844.07, 206.92], [811.36, 205.41], [860.07, 203.49], [892.88, 204.82], [844.07, 163.98], [811.36, 164.37], [860.07, 164.87], [892.88, 164.53]]}, {\"category\": \"car\", \"corners_3d\": [[819.5, 206.16], [793.84, 204.81], [851.05, 203.06], [877.52, 204.26], [819.5, 167.91], [793.84, 168.11], [851.05, 168.37], [877.52, 168.19]]}, {\"category\": \"car\", \"corners_3d\": [[678.66, 216.66], [722.03, 216.69], [739.13, 223.85], [688.67, 223.81], [678.66, 175.94], [722.03, 175.94], [739.13, 176.44], [688.67, 176.44]]}, {\"category\": \"car\", \"corners_3d\": [[551.78, 214.18], [513.85, 213.92], [535.82, 208.97], [569.27, 209.17], [551.78, 174.36], [513.85, 174.35], [535.82, 174.17], [569.27, 174.17]]}, {\"category\": \"car\", \"corners_3d\": [[729.32, 191.37], [726.6, 190.7], [785.3, 190.79], [790.25, 191.46], [729.32, 163.59], [726.6, 163.92], [785.3, 163.88], [790.25, 163.54]]}]\n```" - } - ] - }, - { - "id": 2602, - "data_source": "Mono3DRefer", - "scene_name": "002972", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[640.58, 189.2], [673.5, 189.3], [667.17, 192.66], [627.62, 192.52], [640.58, 149.23], [673.5, 149.09], [667.17, 144.23], [627.62, 144.44]]}]\n```" - } - ] - }, - { - "id": 2603, - "data_source": "Mono3DRefer", - "scene_name": "002972", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[157.49, 541.3], [-156.17, 520.56], [303.14, 333.14], [455.47, 337.41], [157.49, 201.46], [-156.17, 199.85], [303.14, 185.3], [455.47, 185.63]]}, {\"category\": \"car\", \"corners_3d\": [[840.49, 351.15], [1025.36, 352.1], [1464.37, 548.43], [1074.41, 544.31], [840.49, 178.29], [1025.36, 178.31], [1464.37, 184.29], [1074.41, 184.17]]}, {\"category\": \"car\", \"corners_3d\": [[1220.52, 222.92], [1285.72, 221.12], [1550.52, 235.46], [1477.95, 238.53], [1220.52, 133.45], [1285.72, 134.87], [1550.52, 123.58], [1477.95, 121.17]]}, {\"category\": \"van\", \"corners_3d\": [[1081.29, 238.12], [1144.6, 236.19], [1371.22, 255.75], [1296.7, 259.08], [1081.29, 140.19], [1144.6, 141.16], [1371.22, 131.37], [1296.7, 129.71]]}, {\"category\": \"van\", \"corners_3d\": [[938.97, 226.19], [1010.2, 225.15], [1189.92, 242.83], [1098.77, 244.71], [938.97, 140.64], [1010.2, 141.27], [1189.92, 130.59], [1098.77, 129.46]]}, {\"category\": \"car\", \"corners_3d\": [[455.14, 268.49], [362.58, 268.62], [415.45, 247.45], [487.51, 247.37], [455.14, 182.0], [362.58, 182.01], [415.45, 179.99], [487.51, 179.98]]}, {\"category\": \"car\", \"corners_3d\": [[498.92, 241.73], [433.8, 241.29], [477.16, 228.14], [529.93, 228.43], [498.92, 179.48], [433.8, 179.44], [477.16, 178.17], [529.93, 178.2]]}, {\"category\": \"car\", \"corners_3d\": [[967.22, 215.25], [916.52, 212.88], [971.21, 209.46], [1020.28, 211.42], [967.22, 158.6], [916.52, 159.39], [971.21, 160.54], [1020.28, 159.88]]}, {\"category\": \"car\", \"corners_3d\": [[884.5, 211.64], [842.47, 209.64], [896.81, 206.91], [938.3, 208.62], [884.5, 164.46], [842.47, 164.9], [896.81, 165.49], [938.3, 165.12]]}, {\"category\": \"car\", \"corners_3d\": [[844.07, 206.92], [811.36, 205.41], [860.07, 203.49], [892.88, 204.82], [844.07, 163.98], [811.36, 164.37], [860.07, 164.87], [892.88, 164.53]]}, {\"category\": \"car\", \"corners_3d\": [[819.5, 206.16], [793.84, 204.81], [851.05, 203.06], [877.52, 204.26], [819.5, 167.91], [793.84, 168.11], [851.05, 168.37], [877.52, 168.19]]}, {\"category\": \"car\", \"corners_3d\": [[678.66, 216.66], [722.03, 216.69], [739.13, 223.85], [688.67, 223.81], [678.66, 175.94], [722.03, 175.94], [739.13, 176.44], [688.67, 176.44]]}, {\"category\": \"car\", \"corners_3d\": [[551.78, 214.18], [513.85, 213.92], [535.82, 208.97], [569.27, 209.17], [551.78, 174.36], [513.85, 174.35], [535.82, 174.17], [569.27, 174.17]]}, {\"category\": \"van\", \"corners_3d\": [[572.27, 207.63], [534.9, 207.5], [549.0, 203.58], [582.17, 203.68], [572.27, 165.89], [534.9, 165.92], [549.0, 166.7], [582.17, 166.68]]}, {\"category\": \"car\", \"corners_3d\": [[729.32, 191.37], [726.6, 190.7], [785.3, 190.79], [790.25, 191.46], [729.32, 163.59], [726.6, 163.92], [785.3, 163.88], [790.25, 163.54]]}]\n```" - } - ] - }, - { - "id": 2604, - "data_source": "Mono3DRefer", - "scene_name": "002972", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1081.29, 238.12], [1144.6, 236.19], [1371.22, 255.75], [1296.7, 259.08], [1081.29, 140.19], [1144.6, 141.16], [1371.22, 131.37], [1296.7, 129.71]]}, {\"category\": \"van\", \"corners_3d\": [[938.97, 226.19], [1010.2, 225.15], [1189.92, 242.83], [1098.77, 244.71], [938.97, 140.64], [1010.2, 141.27], [1189.92, 130.59], [1098.77, 129.46]]}, {\"category\": \"van\", \"corners_3d\": [[572.27, 207.63], [534.9, 207.5], [549.0, 203.58], [582.17, 203.68], [572.27, 165.89], [534.9, 165.92], [549.0, 166.7], [582.17, 166.68]]}, {\"category\": \"truck\", \"corners_3d\": [[640.58, 189.2], [673.5, 189.3], [667.17, 192.66], [627.62, 192.52], [640.58, 149.23], [673.5, 149.09], [667.17, 144.23], [627.62, 144.44]]}]\n```" - } - ] - }, - { - "id": 2605, - "data_source": "Mono3DRefer", - "scene_name": "002972", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[157.49, 541.3], [-156.17, 520.56], [303.14, 333.14], [455.47, 337.41], [157.49, 201.46], [-156.17, 199.85], [303.14, 185.3], [455.47, 185.63]]}, {\"category\": \"car\", \"corners_3d\": [[840.49, 351.15], [1025.36, 352.1], [1464.37, 548.43], [1074.41, 544.31], [840.49, 178.29], [1025.36, 178.31], [1464.37, 184.29], [1074.41, 184.17]]}, {\"category\": \"car\", \"corners_3d\": [[1220.52, 222.92], [1285.72, 221.12], [1550.52, 235.46], [1477.95, 238.53], [1220.52, 133.45], [1285.72, 134.87], [1550.52, 123.58], [1477.95, 121.17]]}, {\"category\": \"car\", \"corners_3d\": [[455.14, 268.49], [362.58, 268.62], [415.45, 247.45], [487.51, 247.37], [455.14, 182.0], [362.58, 182.01], [415.45, 179.99], [487.51, 179.98]]}, {\"category\": \"car\", \"corners_3d\": [[498.92, 241.73], [433.8, 241.29], [477.16, 228.14], [529.93, 228.43], [498.92, 179.48], [433.8, 179.44], [477.16, 178.17], [529.93, 178.2]]}, {\"category\": \"car\", \"corners_3d\": [[967.22, 215.25], [916.52, 212.88], [971.21, 209.46], [1020.28, 211.42], [967.22, 158.6], [916.52, 159.39], [971.21, 160.54], [1020.28, 159.88]]}, {\"category\": \"car\", \"corners_3d\": [[884.5, 211.64], [842.47, 209.64], [896.81, 206.91], [938.3, 208.62], [884.5, 164.46], [842.47, 164.9], [896.81, 165.49], [938.3, 165.12]]}, {\"category\": \"car\", \"corners_3d\": [[844.07, 206.92], [811.36, 205.41], [860.07, 203.49], [892.88, 204.82], [844.07, 163.98], [811.36, 164.37], [860.07, 164.87], [892.88, 164.53]]}, {\"category\": \"car\", \"corners_3d\": [[819.5, 206.16], [793.84, 204.81], [851.05, 203.06], [877.52, 204.26], [819.5, 167.91], [793.84, 168.11], [851.05, 168.37], [877.52, 168.19]]}, {\"category\": \"car\", \"corners_3d\": [[678.66, 216.66], [722.03, 216.69], [739.13, 223.85], [688.67, 223.81], [678.66, 175.94], [722.03, 175.94], [739.13, 176.44], [688.67, 176.44]]}, {\"category\": \"car\", \"corners_3d\": [[551.78, 214.18], [513.85, 213.92], [535.82, 208.97], [569.27, 209.17], [551.78, 174.36], [513.85, 174.35], [535.82, 174.17], [569.27, 174.17]]}, {\"category\": \"car\", \"corners_3d\": [[729.32, 191.37], [726.6, 190.7], [785.3, 190.79], [790.25, 191.46], [729.32, 163.59], [726.6, 163.92], [785.3, 163.88], [790.25, 163.54]]}, {\"category\": \"truck\", \"corners_3d\": [[640.58, 189.2], [673.5, 189.3], [667.17, 192.66], [627.62, 192.52], [640.58, 149.23], [673.5, 149.09], [667.17, 144.23], [627.62, 144.44]]}]\n```" - } - ] - }, - { - "id": 2606, - "data_source": "Mono3DRefer", - "scene_name": "002972", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[157.49, 541.3], [-156.17, 520.56], [303.14, 333.14], [455.47, 337.41], [157.49, 201.46], [-156.17, 199.85], [303.14, 185.3], [455.47, 185.63]]}, {\"category\": \"car\", \"corners_3d\": [[840.49, 351.15], [1025.36, 352.1], [1464.37, 548.43], [1074.41, 544.31], [840.49, 178.29], [1025.36, 178.31], [1464.37, 184.29], [1074.41, 184.17]]}, {\"category\": \"car\", \"corners_3d\": [[1220.52, 222.92], [1285.72, 221.12], [1550.52, 235.46], [1477.95, 238.53], [1220.52, 133.45], [1285.72, 134.87], [1550.52, 123.58], [1477.95, 121.17]]}, {\"category\": \"van\", \"corners_3d\": [[1081.29, 238.12], [1144.6, 236.19], [1371.22, 255.75], [1296.7, 259.08], [1081.29, 140.19], [1144.6, 141.16], [1371.22, 131.37], [1296.7, 129.71]]}, {\"category\": \"van\", \"corners_3d\": [[938.97, 226.19], [1010.2, 225.15], [1189.92, 242.83], [1098.77, 244.71], [938.97, 140.64], [1010.2, 141.27], [1189.92, 130.59], [1098.77, 129.46]]}, {\"category\": \"car\", \"corners_3d\": [[455.14, 268.49], [362.58, 268.62], [415.45, 247.45], [487.51, 247.37], [455.14, 182.0], [362.58, 182.01], [415.45, 179.99], [487.51, 179.98]]}, {\"category\": \"car\", \"corners_3d\": [[498.92, 241.73], [433.8, 241.29], [477.16, 228.14], [529.93, 228.43], [498.92, 179.48], [433.8, 179.44], [477.16, 178.17], [529.93, 178.2]]}, {\"category\": \"car\", \"corners_3d\": [[967.22, 215.25], [916.52, 212.88], [971.21, 209.46], [1020.28, 211.42], [967.22, 158.6], [916.52, 159.39], [971.21, 160.54], [1020.28, 159.88]]}, {\"category\": \"car\", \"corners_3d\": [[884.5, 211.64], [842.47, 209.64], [896.81, 206.91], [938.3, 208.62], [884.5, 164.46], [842.47, 164.9], [896.81, 165.49], [938.3, 165.12]]}, {\"category\": \"car\", \"corners_3d\": [[844.07, 206.92], [811.36, 205.41], [860.07, 203.49], [892.88, 204.82], [844.07, 163.98], [811.36, 164.37], [860.07, 164.87], [892.88, 164.53]]}, {\"category\": \"car\", \"corners_3d\": [[819.5, 206.16], [793.84, 204.81], [851.05, 203.06], [877.52, 204.26], [819.5, 167.91], [793.84, 168.11], [851.05, 168.37], [877.52, 168.19]]}, {\"category\": \"car\", \"corners_3d\": [[678.66, 216.66], [722.03, 216.69], [739.13, 223.85], [688.67, 223.81], [678.66, 175.94], [722.03, 175.94], [739.13, 176.44], [688.67, 176.44]]}, {\"category\": \"car\", \"corners_3d\": [[551.78, 214.18], [513.85, 213.92], [535.82, 208.97], [569.27, 209.17], [551.78, 174.36], [513.85, 174.35], [535.82, 174.17], [569.27, 174.17]]}, {\"category\": \"van\", \"corners_3d\": [[572.27, 207.63], [534.9, 207.5], [549.0, 203.58], [582.17, 203.68], [572.27, 165.89], [534.9, 165.92], [549.0, 166.7], [582.17, 166.68]]}, {\"category\": \"car\", \"corners_3d\": [[729.32, 191.37], [726.6, 190.7], [785.3, 190.79], [790.25, 191.46], [729.32, 163.59], [726.6, 163.92], [785.3, 163.88], [790.25, 163.54]]}, {\"category\": \"truck\", \"corners_3d\": [[640.58, 189.2], [673.5, 189.3], [667.17, 192.66], [627.62, 192.52], [640.58, 149.23], [673.5, 149.09], [667.17, 144.23], [627.62, 144.44]]}]\n```" - } - ] - }, - { - "id": 2607, - "data_source": "Mono3DRefer", - "scene_name": "002996", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[361.91, 425.12], [127.89, 426.0], [315.99, 325.52], [456.92, 325.2], [361.91, 219.84], [127.89, 220.01], [315.99, 201.29], [456.92, 201.23]]}, {\"category\": \"car\", \"corners_3d\": [[784.92, 332.78], [921.93, 330.95], [1165.15, 439.22], [937.23, 444.45], [784.92, 191.91], [921.93, 191.69], [1165.15, 204.59], [937.23, 205.21]]}, {\"category\": \"car\", \"corners_3d\": [[700.22, 268.78], [786.76, 268.68], [859.82, 306.76], [738.95, 306.97], [700.22, 188.56], [786.76, 188.54], [859.82, 194.77], [738.95, 194.81]]}, {\"category\": \"car\", \"corners_3d\": [[545.18, 243.38], [490.36, 243.17], [512.76, 232.49], [559.29, 232.64], [545.18, 187.8], [490.36, 187.76], [512.76, 185.5], [559.29, 185.53]]}, {\"category\": \"car\", \"corners_3d\": [[577.8, 228.37], [530.88, 228.31], [544.02, 220.43], [584.29, 220.48], [577.8, 174.26], [530.88, 174.26], [544.02, 174.06], [584.29, 174.06]]}, {\"category\": \"car\", \"corners_3d\": [[601.57, 219.79], [563.0, 219.59], [575.39, 214.02], [609.39, 214.17], [601.57, 177.68], [563.0, 177.66], [575.39, 177.08], [609.39, 177.1]]}, {\"category\": \"car\", \"corners_3d\": [[641.56, 209.06], [617.4, 208.71], [633.02, 206.0], [655.48, 206.31], [641.56, 185.79], [617.4, 185.67], [633.02, 184.7], [655.48, 184.81]]}]\n```" - } - ] - }, - { - "id": 2608, - "data_source": "Mono3DRefer", - "scene_name": "003002", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[480.38, 188.08], [470.83, 187.99], [497.74, 187.7], [507.25, 187.79], [480.38, 143.88], [470.83, 144.35], [497.74, 145.73], [507.25, 145.3]]}]\n```" - } - ] - }, - { - "id": 2609, - "data_source": "Mono3DRefer", - "scene_name": "003002", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[310.98, 322.03], [345.65, 311.53], [428.3, 315.68], [399.3, 326.89], [310.98, 151.15], [345.65, 153.33], [428.3, 152.47], [399.3, 150.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[628.37, 234.85], [605.55, 235.54], [591.26, 232.7], [613.09, 232.08], [628.37, 143.39], [605.55, 142.92], [591.26, 144.86], [613.09, 145.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[561.75, 212.99], [535.81, 213.23], [532.72, 211.9], [557.63, 211.68], [561.75, 132.62], [535.81, 132.27], [532.72, 134.24], [557.63, 134.57]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[592.67, 213.26], [568.42, 213.53], [564.67, 212.58], [588.26, 212.33], [592.67, 143.09], [568.42, 142.79], [564.67, 143.88], [588.26, 144.16]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[645.46, 212.42], [627.4, 212.64], [619.67, 211.47], [637.14, 211.27], [645.46, 147.22], [627.4, 146.99], [619.67, 148.21], [637.14, 148.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1049.97, 439.25], [1023.74, 463.32], [853.58, 419.82], [888.49, 402.36], [1049.97, 136.14], [1023.74, 132.01], [853.58, 139.47], [888.49, 142.46]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[954.76, 451.12], [899.37, 486.61], [764.13, 442.66], [826.03, 416.19], [954.76, 116.65], [899.37, 108.27], [764.13, 118.64], [826.03, 124.89]]}]\n```" - } - ] - }, - { - "id": 2610, - "data_source": "Mono3DRefer", - "scene_name": "003002", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[310.98, 322.03], [345.65, 311.53], [428.3, 315.68], [399.3, 326.89], [310.98, 151.15], [345.65, 153.33], [428.3, 152.47], [399.3, 150.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[628.37, 234.85], [605.55, 235.54], [591.26, 232.7], [613.09, 232.08], [628.37, 143.39], [605.55, 142.92], [591.26, 144.86], [613.09, 145.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[561.75, 212.99], [535.81, 213.23], [532.72, 211.9], [557.63, 211.68], [561.75, 132.62], [535.81, 132.27], [532.72, 134.24], [557.63, 134.57]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[592.67, 213.26], [568.42, 213.53], [564.67, 212.58], [588.26, 212.33], [592.67, 143.09], [568.42, 142.79], [564.67, 143.88], [588.26, 144.16]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[645.46, 212.42], [627.4, 212.64], [619.67, 211.47], [637.14, 211.27], [645.46, 147.22], [627.4, 146.99], [619.67, 148.21], [637.14, 148.42]]}, {\"category\": \"cyclist\", \"corners_3d\": [[480.38, 188.08], [470.83, 187.99], [497.74, 187.7], [507.25, 187.79], [480.38, 143.88], [470.83, 144.35], [497.74, 145.73], [507.25, 145.3]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1049.97, 439.25], [1023.74, 463.32], [853.58, 419.82], [888.49, 402.36], [1049.97, 136.14], [1023.74, 132.01], [853.58, 139.47], [888.49, 142.46]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[954.76, 451.12], [899.37, 486.61], [764.13, 442.66], [826.03, 416.19], [954.76, 116.65], [899.37, 108.27], [764.13, 118.64], [826.03, 124.89]]}]\n```" - } - ] - }, - { - "id": 2611, - "data_source": "Mono3DRefer", - "scene_name": "003032", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[130.18, 486.21], [-128.0, 487.41], [202.09, 345.13], [343.04, 344.77], [130.18, 275.61], [-128.0, 276.0], [202.09, 229.35], [343.04, 229.23]]}, {\"category\": \"car\", \"corners_3d\": [[409.98, 297.39], [295.27, 297.6], [378.96, 263.54], [462.28, 263.43], [409.98, 184.59], [295.27, 184.61], [378.96, 181.4], [462.28, 181.39]]}, {\"category\": \"car\", \"corners_3d\": [[616.62, 211.78], [653.83, 211.76], [659.92, 216.42], [618.26, 216.44], [616.62, 176.25], [653.83, 176.25], [659.92, 176.66], [618.26, 176.66]]}]\n```" - } - ] - }, - { - "id": 2612, - "data_source": "Mono3DRefer", - "scene_name": "003032", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[534.74, 217.01], [484.01, 216.76], [518.49, 208.48], [559.77, 208.64], [534.74, 139.01], [484.01, 139.2], [518.49, 145.54], [559.77, 145.42]]}]\n```" - } - ] - }, - { - "id": 2613, - "data_source": "Mono3DRefer", - "scene_name": "003032", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[605.03, 216.71], [615.11, 216.24], [621.78, 216.75], [611.66, 217.24], [605.03, 173.61], [615.11, 173.61], [621.78, 173.62], [611.66, 173.62]]}]\n```" - } - ] - }, - { - "id": 2614, - "data_source": "Mono3DRefer", - "scene_name": "003032", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[130.18, 486.21], [-128.0, 487.41], [202.09, 345.13], [343.04, 344.77], [130.18, 275.61], [-128.0, 276.0], [202.09, 229.35], [343.04, 229.23]]}, {\"category\": \"car\", \"corners_3d\": [[409.98, 297.39], [295.27, 297.6], [378.96, 263.54], [462.28, 263.43], [409.98, 184.59], [295.27, 184.61], [378.96, 181.4], [462.28, 181.39]]}, {\"category\": \"truck\", \"corners_3d\": [[534.74, 217.01], [484.01, 216.76], [518.49, 208.48], [559.77, 208.64], [534.74, 139.01], [484.01, 139.2], [518.49, 145.54], [559.77, 145.42]]}, {\"category\": \"car\", \"corners_3d\": [[616.62, 211.78], [653.83, 211.76], [659.92, 216.42], [618.26, 216.44], [616.62, 176.25], [653.83, 176.25], [659.92, 176.66], [618.26, 176.66]]}]\n```" - } - ] - }, - { - "id": 2615, - "data_source": "Mono3DRefer", - "scene_name": "003032", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[130.18, 486.21], [-128.0, 487.41], [202.09, 345.13], [343.04, 344.77], [130.18, 275.61], [-128.0, 276.0], [202.09, 229.35], [343.04, 229.23]]}, {\"category\": \"car\", \"corners_3d\": [[409.98, 297.39], [295.27, 297.6], [378.96, 263.54], [462.28, 263.43], [409.98, 184.59], [295.27, 184.61], [378.96, 181.4], [462.28, 181.39]]}, {\"category\": \"car\", \"corners_3d\": [[616.62, 211.78], [653.83, 211.76], [659.92, 216.42], [618.26, 216.44], [616.62, 176.25], [653.83, 176.25], [659.92, 176.66], [618.26, 176.66]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[605.03, 216.71], [615.11, 216.24], [621.78, 216.75], [611.66, 217.24], [605.03, 173.61], [615.11, 173.61], [621.78, 173.62], [611.66, 173.62]]}]\n```" - } - ] - }, - { - "id": 2616, - "data_source": "Mono3DRefer", - "scene_name": "003032", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[534.74, 217.01], [484.01, 216.76], [518.49, 208.48], [559.77, 208.64], [534.74, 139.01], [484.01, 139.2], [518.49, 145.54], [559.77, 145.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[605.03, 216.71], [615.11, 216.24], [621.78, 216.75], [611.66, 217.24], [605.03, 173.61], [615.11, 173.61], [621.78, 173.62], [611.66, 173.62]]}]\n```" - } - ] - }, - { - "id": 2617, - "data_source": "Mono3DRefer", - "scene_name": "003032", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[130.18, 486.21], [-128.0, 487.41], [202.09, 345.13], [343.04, 344.77], [130.18, 275.61], [-128.0, 276.0], [202.09, 229.35], [343.04, 229.23]]}, {\"category\": \"car\", \"corners_3d\": [[409.98, 297.39], [295.27, 297.6], [378.96, 263.54], [462.28, 263.43], [409.98, 184.59], [295.27, 184.61], [378.96, 181.4], [462.28, 181.39]]}, {\"category\": \"truck\", \"corners_3d\": [[534.74, 217.01], [484.01, 216.76], [518.49, 208.48], [559.77, 208.64], [534.74, 139.01], [484.01, 139.2], [518.49, 145.54], [559.77, 145.42]]}, {\"category\": \"car\", \"corners_3d\": [[616.62, 211.78], [653.83, 211.76], [659.92, 216.42], [618.26, 216.44], [616.62, 176.25], [653.83, 176.25], [659.92, 176.66], [618.26, 176.66]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[605.03, 216.71], [615.11, 216.24], [621.78, 216.75], [611.66, 217.24], [605.03, 173.61], [615.11, 173.61], [621.78, 173.62], [611.66, 173.62]]}]\n```" - } - ] - }, - { - "id": 2618, - "data_source": "Mono3DRefer", - "scene_name": "003044", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[53.08, 234.12], [96.07, 233.95], [19.58, 243.73], [-30.56, 243.96], [53.08, 197.16], [96.07, 197.09], [19.58, 200.97], [-30.56, 201.06]]}, {\"category\": \"car\", \"corners_3d\": [[117.85, 233.04], [161.06, 232.73], [99.87, 242.27], [49.38, 242.68], [117.85, 198.71], [161.06, 198.58], [99.87, 202.68], [49.38, 202.86]]}]\n```" - } - ] - }, - { - "id": 2619, - "data_source": "Mono3DRefer", - "scene_name": "003045", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[554.58, 228.39], [511.08, 228.39], [523.79, 221.18], [561.64, 221.18], [554.58, 191.19], [511.08, 191.19], [523.79, 188.81], [561.64, 188.8]]}, {\"category\": \"car\", \"corners_3d\": [[590.56, 190.19], [572.02, 190.2], [573.34, 189.18], [590.79, 189.17], [590.56, 170.67], [572.02, 170.67], [573.34, 170.8], [590.79, 170.8]]}]\n```" - } - ] - }, - { - "id": 2620, - "data_source": "Mono3DRefer", - "scene_name": "003047", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[577.32, 192.41], [551.42, 192.38], [558.95, 191.49], [582.72, 191.51], [577.32, 167.92], [551.42, 167.95], [558.95, 169.06], [582.72, 169.04]]}]\n```" - } - ] - }, - { - "id": 2621, - "data_source": "Mono3DRefer", - "scene_name": "003047", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[931.64, 366.34], [1079.18, 382.61], [1097.67, 421.78], [922.37, 398.91], [931.64, 134.08], [1079.18, 129.9], [1097.67, 119.85], [922.37, 125.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[991.0, 318.03], [1066.67, 311.24], [1164.27, 328.45], [1083.8, 337.21], [991.0, 115.59], [1066.67, 118.87], [1164.27, 110.56], [1083.8, 106.33]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[528.88, 258.55], [493.17, 258.32], [502.73, 253.42], [536.19, 253.62], [528.88, 161.74], [493.17, 161.8], [502.73, 163.06], [536.19, 163.01]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[506.05, 255.64], [465.79, 255.93], [470.42, 251.82], [508.44, 251.55], [506.05, 164.52], [465.79, 164.45], [470.42, 165.39], [508.44, 165.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[585.12, 230.73], [614.66, 231.0], [609.96, 233.9], [578.72, 233.6], [585.12, 161.9], [614.66, 161.79], [609.96, 160.64], [578.72, 160.76]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[607.84, 211.84], [633.99, 211.77], [636.84, 212.9], [609.72, 212.97], [607.84, 160.13], [633.99, 160.18], [636.84, 159.38], [609.72, 159.33]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[643.46, 229.58], [669.02, 229.7], [669.96, 232.26], [643.04, 232.13], [643.46, 162.76], [669.02, 162.71], [669.96, 161.71], [643.04, 161.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[699.34, 215.27], [726.41, 215.29], [731.24, 216.76], [702.99, 216.73], [699.34, 156.76], [726.41, 156.74], [731.24, 155.67], [702.99, 155.69]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[451.93, 233.46], [464.94, 231.56], [493.82, 232.18], [481.75, 234.12], [451.93, 163.21], [464.94, 163.88], [493.82, 163.66], [481.75, 162.98]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[668.5, 222.98], [693.87, 223.08], [695.96, 225.13], [669.34, 225.02], [668.5, 163.72], [693.87, 163.67], [695.96, 162.79], [669.34, 162.84]]}]\n```" - } - ] - }, - { - "id": 2622, - "data_source": "Mono3DRefer", - "scene_name": "003047", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[21.5, 539.25], [112.97, 492.73], [211.44, 501.11], [132.3, 550.36], [21.5, 272.37], [112.97, 260.55], [211.44, 262.68], [132.3, 275.19]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[321.71, 314.11], [348.12, 306.29], [418.17, 309.42], [395.56, 317.65], [321.71, 210.5], [348.12, 208.79], [418.17, 209.47], [395.56, 211.27]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[98.03, 478.79], [171.77, 439.73], [278.11, 443.01], [219.62, 483.16], [98.03, 264.41], [171.77, 253.51], [278.11, 254.43], [219.62, 265.62]]}]\n```" - } - ] - }, - { - "id": 2623, - "data_source": "Mono3DRefer", - "scene_name": "003047", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[931.64, 366.34], [1079.18, 382.61], [1097.67, 421.78], [922.37, 398.91], [931.64, 134.08], [1079.18, 129.9], [1097.67, 119.85], [922.37, 125.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[991.0, 318.03], [1066.67, 311.24], [1164.27, 328.45], [1083.8, 337.21], [991.0, 115.59], [1066.67, 118.87], [1164.27, 110.56], [1083.8, 106.33]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[528.88, 258.55], [493.17, 258.32], [502.73, 253.42], [536.19, 253.62], [528.88, 161.74], [493.17, 161.8], [502.73, 163.06], [536.19, 163.01]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[506.05, 255.64], [465.79, 255.93], [470.42, 251.82], [508.44, 251.55], [506.05, 164.52], [465.79, 164.45], [470.42, 165.39], [508.44, 165.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[585.12, 230.73], [614.66, 231.0], [609.96, 233.9], [578.72, 233.6], [585.12, 161.9], [614.66, 161.79], [609.96, 160.64], [578.72, 160.76]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[607.84, 211.84], [633.99, 211.77], [636.84, 212.9], [609.72, 212.97], [607.84, 160.13], [633.99, 160.18], [636.84, 159.38], [609.72, 159.33]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[643.46, 229.58], [669.02, 229.7], [669.96, 232.26], [643.04, 232.13], [643.46, 162.76], [669.02, 162.71], [669.96, 161.71], [643.04, 161.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[699.34, 215.27], [726.41, 215.29], [731.24, 216.76], [702.99, 216.73], [699.34, 156.76], [726.41, 156.74], [731.24, 155.67], [702.99, 155.69]]}, {\"category\": \"car\", \"corners_3d\": [[577.32, 192.41], [551.42, 192.38], [558.95, 191.49], [582.72, 191.51], [577.32, 167.92], [551.42, 167.95], [558.95, 169.06], [582.72, 169.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[451.93, 233.46], [464.94, 231.56], [493.82, 232.18], [481.75, 234.12], [451.93, 163.21], [464.94, 163.88], [493.82, 163.66], [481.75, 162.98]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[668.5, 222.98], [693.87, 223.08], [695.96, 225.13], [669.34, 225.02], [668.5, 163.72], [693.87, 163.67], [695.96, 162.79], [669.34, 162.84]]}]\n```" - } - ] - }, - { - "id": 2624, - "data_source": "Mono3DRefer", - "scene_name": "003047", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[21.5, 539.25], [112.97, 492.73], [211.44, 501.11], [132.3, 550.36], [21.5, 272.37], [112.97, 260.55], [211.44, 262.68], [132.3, 275.19]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[321.71, 314.11], [348.12, 306.29], [418.17, 309.42], [395.56, 317.65], [321.71, 210.5], [348.12, 208.79], [418.17, 209.47], [395.56, 211.27]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[98.03, 478.79], [171.77, 439.73], [278.11, 443.01], [219.62, 483.16], [98.03, 264.41], [171.77, 253.51], [278.11, 254.43], [219.62, 265.62]]}, {\"category\": \"car\", \"corners_3d\": [[577.32, 192.41], [551.42, 192.38], [558.95, 191.49], [582.72, 191.51], [577.32, 167.92], [551.42, 167.95], [558.95, 169.06], [582.72, 169.04]]}]\n```" - } - ] - }, - { - "id": 2625, - "data_source": "Mono3DRefer", - "scene_name": "003047", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[931.64, 366.34], [1079.18, 382.61], [1097.67, 421.78], [922.37, 398.91], [931.64, 134.08], [1079.18, 129.9], [1097.67, 119.85], [922.37, 125.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[991.0, 318.03], [1066.67, 311.24], [1164.27, 328.45], [1083.8, 337.21], [991.0, 115.59], [1066.67, 118.87], [1164.27, 110.56], [1083.8, 106.33]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[21.5, 539.25], [112.97, 492.73], [211.44, 501.11], [132.3, 550.36], [21.5, 272.37], [112.97, 260.55], [211.44, 262.68], [132.3, 275.19]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[321.71, 314.11], [348.12, 306.29], [418.17, 309.42], [395.56, 317.65], [321.71, 210.5], [348.12, 208.79], [418.17, 209.47], [395.56, 211.27]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[98.03, 478.79], [171.77, 439.73], [278.11, 443.01], [219.62, 483.16], [98.03, 264.41], [171.77, 253.51], [278.11, 254.43], [219.62, 265.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[528.88, 258.55], [493.17, 258.32], [502.73, 253.42], [536.19, 253.62], [528.88, 161.74], [493.17, 161.8], [502.73, 163.06], [536.19, 163.01]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[506.05, 255.64], [465.79, 255.93], [470.42, 251.82], [508.44, 251.55], [506.05, 164.52], [465.79, 164.45], [470.42, 165.39], [508.44, 165.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[585.12, 230.73], [614.66, 231.0], [609.96, 233.9], [578.72, 233.6], [585.12, 161.9], [614.66, 161.79], [609.96, 160.64], [578.72, 160.76]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[607.84, 211.84], [633.99, 211.77], [636.84, 212.9], [609.72, 212.97], [607.84, 160.13], [633.99, 160.18], [636.84, 159.38], [609.72, 159.33]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[643.46, 229.58], [669.02, 229.7], [669.96, 232.26], [643.04, 232.13], [643.46, 162.76], [669.02, 162.71], [669.96, 161.71], [643.04, 161.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[699.34, 215.27], [726.41, 215.29], [731.24, 216.76], [702.99, 216.73], [699.34, 156.76], [726.41, 156.74], [731.24, 155.67], [702.99, 155.69]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[451.93, 233.46], [464.94, 231.56], [493.82, 232.18], [481.75, 234.12], [451.93, 163.21], [464.94, 163.88], [493.82, 163.66], [481.75, 162.98]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[668.5, 222.98], [693.87, 223.08], [695.96, 225.13], [669.34, 225.02], [668.5, 163.72], [693.87, 163.67], [695.96, 162.79], [669.34, 162.84]]}]\n```" - } - ] - }, - { - "id": 2626, - "data_source": "Mono3DRefer", - "scene_name": "003047", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[931.64, 366.34], [1079.18, 382.61], [1097.67, 421.78], [922.37, 398.91], [931.64, 134.08], [1079.18, 129.9], [1097.67, 119.85], [922.37, 125.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[991.0, 318.03], [1066.67, 311.24], [1164.27, 328.45], [1083.8, 337.21], [991.0, 115.59], [1066.67, 118.87], [1164.27, 110.56], [1083.8, 106.33]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[21.5, 539.25], [112.97, 492.73], [211.44, 501.11], [132.3, 550.36], [21.5, 272.37], [112.97, 260.55], [211.44, 262.68], [132.3, 275.19]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[321.71, 314.11], [348.12, 306.29], [418.17, 309.42], [395.56, 317.65], [321.71, 210.5], [348.12, 208.79], [418.17, 209.47], [395.56, 211.27]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[98.03, 478.79], [171.77, 439.73], [278.11, 443.01], [219.62, 483.16], [98.03, 264.41], [171.77, 253.51], [278.11, 254.43], [219.62, 265.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[528.88, 258.55], [493.17, 258.32], [502.73, 253.42], [536.19, 253.62], [528.88, 161.74], [493.17, 161.8], [502.73, 163.06], [536.19, 163.01]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[506.05, 255.64], [465.79, 255.93], [470.42, 251.82], [508.44, 251.55], [506.05, 164.52], [465.79, 164.45], [470.42, 165.39], [508.44, 165.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[585.12, 230.73], [614.66, 231.0], [609.96, 233.9], [578.72, 233.6], [585.12, 161.9], [614.66, 161.79], [609.96, 160.64], [578.72, 160.76]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[607.84, 211.84], [633.99, 211.77], [636.84, 212.9], [609.72, 212.97], [607.84, 160.13], [633.99, 160.18], [636.84, 159.38], [609.72, 159.33]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[643.46, 229.58], [669.02, 229.7], [669.96, 232.26], [643.04, 232.13], [643.46, 162.76], [669.02, 162.71], [669.96, 161.71], [643.04, 161.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[699.34, 215.27], [726.41, 215.29], [731.24, 216.76], [702.99, 216.73], [699.34, 156.76], [726.41, 156.74], [731.24, 155.67], [702.99, 155.69]]}, {\"category\": \"car\", \"corners_3d\": [[577.32, 192.41], [551.42, 192.38], [558.95, 191.49], [582.72, 191.51], [577.32, 167.92], [551.42, 167.95], [558.95, 169.06], [582.72, 169.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[451.93, 233.46], [464.94, 231.56], [493.82, 232.18], [481.75, 234.12], [451.93, 163.21], [464.94, 163.88], [493.82, 163.66], [481.75, 162.98]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[668.5, 222.98], [693.87, 223.08], [695.96, 225.13], [669.34, 225.02], [668.5, 163.72], [693.87, 163.67], [695.96, 162.79], [669.34, 162.84]]}]\n```" - } - ] - }, - { - "id": 2627, - "data_source": "Mono3DRefer", - "scene_name": "003053", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[535.0, 194.25], [534.66, 195.02], [461.76, 195.22], [464.66, 194.43], [535.0, 166.42], [534.66, 166.18], [461.76, 166.12], [464.66, 166.36]]}]\n```" - } - ] - }, - { - "id": 2628, - "data_source": "Mono3DRefer", - "scene_name": "003053", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[307.48, 227.41], [322.53, 223.92], [451.79, 223.11], [445.66, 226.48], [307.48, 179.95], [322.53, 179.49], [451.79, 179.39], [445.66, 179.83]]}, {\"category\": \"car\", \"corners_3d\": [[645.94, 224.04], [697.31, 223.82], [726.96, 235.4], [664.01, 235.72], [645.94, 171.56], [697.31, 171.57], [726.96, 171.28], [664.01, 171.27]]}, {\"category\": \"car\", \"corners_3d\": [[501.92, 251.63], [430.67, 251.72], [462.67, 236.97], [520.58, 236.91], [501.92, 186.37], [430.67, 186.39], [462.67, 183.86], [520.58, 183.85]]}, {\"category\": \"car\", \"corners_3d\": [[-333.67, 353.56], [-167.57, 318.75], [220.55, 310.9], [151.87, 341.67], [-333.67, 194.33], [-167.57, 190.2], [220.55, 189.26], [151.87, 192.92]]}]\n```" - } - ] - }, - { - "id": 2629, - "data_source": "Mono3DRefer", - "scene_name": "003053", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[535.0, 194.25], [534.66, 195.02], [461.76, 195.22], [464.66, 194.43], [535.0, 166.42], [534.66, 166.18], [461.76, 166.12], [464.66, 166.36]]}, {\"category\": \"car\", \"corners_3d\": [[307.48, 227.41], [322.53, 223.92], [451.79, 223.11], [445.66, 226.48], [307.48, 179.95], [322.53, 179.49], [451.79, 179.39], [445.66, 179.83]]}, {\"category\": \"car\", \"corners_3d\": [[645.94, 224.04], [697.31, 223.82], [726.96, 235.4], [664.01, 235.72], [645.94, 171.56], [697.31, 171.57], [726.96, 171.28], [664.01, 171.27]]}, {\"category\": \"car\", \"corners_3d\": [[501.92, 251.63], [430.67, 251.72], [462.67, 236.97], [520.58, 236.91], [501.92, 186.37], [430.67, 186.39], [462.67, 183.86], [520.58, 183.85]]}, {\"category\": \"car\", \"corners_3d\": [[-333.67, 353.56], [-167.57, 318.75], [220.55, 310.9], [151.87, 341.67], [-333.67, 194.33], [-167.57, 190.2], [220.55, 189.26], [151.87, 192.92]]}]\n```" - } - ] - }, - { - "id": 2630, - "data_source": "Mono3DRefer", - "scene_name": "003068", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[217.65, 384.29], [17.8, 384.92], [247.21, 301.61], [368.28, 301.38], [217.65, 191.04], [17.8, 191.1], [247.21, 183.93], [368.28, 183.91]]}, {\"category\": \"car\", \"corners_3d\": [[658.8, 199.13], [692.57, 199.21], [696.45, 201.82], [659.33, 201.73], [658.8, 167.12], [692.57, 167.1], [696.45, 166.53], [659.33, 166.55]]}]\n```" - } - ] - }, - { - "id": 2631, - "data_source": "Mono3DRefer", - "scene_name": "003069", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[524.94, 205.2], [486.99, 205.12], [502.21, 202.82], [536.49, 202.88], [524.94, 174.58], [486.99, 174.6], [502.21, 175.28], [536.49, 175.26]]}]\n```" - } - ] - }, - { - "id": 2632, - "data_source": "Mono3DRefer", - "scene_name": "003081", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[423.56, 355.58], [260.21, 353.52], [404.1, 285.98], [506.97, 286.78], [423.56, 184.78], [260.21, 184.65], [404.1, 180.24], [506.97, 180.29]]}]\n```" - } - ] - }, - { - "id": 2633, - "data_source": "Mono3DRefer", - "scene_name": "003087", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[601.16, 217.73], [586.29, 217.72], [587.57, 215.81], [601.8, 215.81], [601.16, 168.79], [586.29, 168.79], [587.57, 168.96], [601.8, 168.96]]}]\n```" - } - ] - }, - { - "id": 2634, - "data_source": "Mono3DRefer", - "scene_name": "003087", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[59.92, 409.51], [298.55, 408.03], [-572.44, 1097.37], [-1550.14, 1120.69], [59.92, 192.03], [298.55, 191.91], [-572.44, 247.78], [-1550.14, 249.67]]}, {\"category\": \"car\", \"corners_3d\": [[529.38, 230.1], [473.06, 229.88], [501.31, 220.52], [548.44, 220.67], [529.38, 170.75], [473.06, 170.75], [501.31, 171.1], [548.44, 171.09]]}, {\"category\": \"car\", \"corners_3d\": [[708.52, 272.67], [806.53, 272.68], [859.04, 299.37], [734.82, 299.35], [708.52, 174.68], [806.53, 174.68], [859.04, 175.17], [734.82, 175.17]]}, {\"category\": \"car\", \"corners_3d\": [[686.66, 242.67], [755.16, 243.14], [777.43, 259.71], [692.73, 259.0], [686.66, 184.34], [755.16, 184.41], [777.43, 187.14], [692.73, 187.02]]}, {\"category\": \"car\", \"corners_3d\": [[529.28, 207.99], [566.31, 208.07], [556.88, 212.64], [515.08, 212.54], [529.28, 176.85], [566.31, 176.86], [556.88, 177.38], [515.08, 177.37]]}, {\"category\": \"car\", \"corners_3d\": [[660.21, 197.49], [634.78, 197.49], [632.83, 195.63], [656.34, 195.63], [660.21, 173.95], [634.78, 173.95], [632.83, 173.87], [656.34, 173.87]]}, {\"category\": \"car\", \"corners_3d\": [[178.12, 240.94], [218.86, 234.53], [403.41, 234.55], [381.86, 240.97], [178.12, 167.78], [218.86, 168.25], [403.41, 168.25], [381.86, 167.77]]}]\n```" - } - ] - }, - { - "id": 2635, - "data_source": "Mono3DRefer", - "scene_name": "003087", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[59.92, 409.51], [298.55, 408.03], [-572.44, 1097.37], [-1550.14, 1120.69], [59.92, 192.03], [298.55, 191.91], [-572.44, 247.78], [-1550.14, 249.67]]}, {\"category\": \"car\", \"corners_3d\": [[529.38, 230.1], [473.06, 229.88], [501.31, 220.52], [548.44, 220.67], [529.38, 170.75], [473.06, 170.75], [501.31, 171.1], [548.44, 171.09]]}, {\"category\": \"car\", \"corners_3d\": [[708.52, 272.67], [806.53, 272.68], [859.04, 299.37], [734.82, 299.35], [708.52, 174.68], [806.53, 174.68], [859.04, 175.17], [734.82, 175.17]]}, {\"category\": \"cyclist\", \"corners_3d\": [[601.16, 217.73], [586.29, 217.72], [587.57, 215.81], [601.8, 215.81], [601.16, 168.79], [586.29, 168.79], [587.57, 168.96], [601.8, 168.96]]}, {\"category\": \"car\", \"corners_3d\": [[686.66, 242.67], [755.16, 243.14], [777.43, 259.71], [692.73, 259.0], [686.66, 184.34], [755.16, 184.41], [777.43, 187.14], [692.73, 187.02]]}, {\"category\": \"car\", \"corners_3d\": [[529.28, 207.99], [566.31, 208.07], [556.88, 212.64], [515.08, 212.54], [529.28, 176.85], [566.31, 176.86], [556.88, 177.38], [515.08, 177.37]]}, {\"category\": \"car\", \"corners_3d\": [[660.21, 197.49], [634.78, 197.49], [632.83, 195.63], [656.34, 195.63], [660.21, 173.95], [634.78, 173.95], [632.83, 173.87], [656.34, 173.87]]}, {\"category\": \"car\", \"corners_3d\": [[178.12, 240.94], [218.86, 234.53], [403.41, 234.55], [381.86, 240.97], [178.12, 167.78], [218.86, 168.25], [403.41, 168.25], [381.86, 167.77]]}]\n```" - } - ] - }, - { - "id": 2636, - "data_source": "Mono3DRefer", - "scene_name": "003088", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[475.51, 195.39], [456.38, 195.39], [465.2, 194.03], [483.17, 194.02], [475.51, 179.65], [456.38, 179.65], [465.2, 179.24], [483.17, 179.24]]}]\n```" - } - ] - }, - { - "id": 2637, - "data_source": "Mono3DRefer", - "scene_name": "003088", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-110.63, 268.99], [-201.5, 269.0], [-22.89, 247.81], [47.95, 247.81], [-110.63, 172.84], [-201.5, 172.84], [-22.89, 172.84], [47.95, 172.84]]}]\n```" - } - ] - }, - { - "id": 2638, - "data_source": "Mono3DRefer", - "scene_name": "003088", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-110.63, 268.99], [-201.5, 269.0], [-22.89, 247.81], [47.95, 247.81], [-110.63, 172.84], [-201.5, 172.84], [-22.89, 172.84], [47.95, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[475.51, 195.39], [456.38, 195.39], [465.2, 194.03], [483.17, 194.02], [475.51, 179.65], [456.38, 179.65], [465.2, 179.24], [483.17, 179.24]]}]\n```" - } - ] - }, - { - "id": 2639, - "data_source": "Mono3DRefer", - "scene_name": "003102", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[602.11, 258.22], [623.75, 256.33], [637.87, 258.47], [615.99, 260.46], [602.11, 166.27], [623.75, 166.41], [637.87, 166.25], [615.99, 166.09]]}]\n```" - } - ] - }, - { - "id": 2640, - "data_source": "Mono3DRefer", - "scene_name": "003102", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[623.44, 247.61], [699.12, 247.54], [725.83, 268.28], [629.15, 268.4], [623.44, 175.31], [699.12, 175.31], [725.83, 175.99], [629.15, 176.0]]}]\n```" - } - ] - }, - { - "id": 2641, - "data_source": "Mono3DRefer", - "scene_name": "003102", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[239.87, 402.33], [-73.2, 407.27], [302.17, 271.07], [431.62, 270.2], [239.87, -49.77], [-73.2, -54.56], [302.17, 77.57], [431.62, 78.42]]}]\n```" - } - ] - }, - { - "id": 2642, - "data_source": "Mono3DRefer", - "scene_name": "003102", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[623.44, 247.61], [699.12, 247.54], [725.83, 268.28], [629.15, 268.4], [623.44, 175.31], [699.12, 175.31], [725.83, 175.99], [629.15, 176.0]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[602.11, 258.22], [623.75, 256.33], [637.87, 258.47], [615.99, 260.46], [602.11, 166.27], [623.75, 166.41], [637.87, 166.25], [615.99, 166.09]]}]\n```" - } - ] - }, - { - "id": 2643, - "data_source": "Mono3DRefer", - "scene_name": "003102", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[239.87, 402.33], [-73.2, 407.27], [302.17, 271.07], [431.62, 270.2], [239.87, -49.77], [-73.2, -54.56], [302.17, 77.57], [431.62, 78.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[602.11, 258.22], [623.75, 256.33], [637.87, 258.47], [615.99, 260.46], [602.11, 166.27], [623.75, 166.41], [637.87, 166.25], [615.99, 166.09]]}]\n```" - } - ] - }, - { - "id": 2644, - "data_source": "Mono3DRefer", - "scene_name": "003102", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[239.87, 402.33], [-73.2, 407.27], [302.17, 271.07], [431.62, 270.2], [239.87, -49.77], [-73.2, -54.56], [302.17, 77.57], [431.62, 78.42]]}, {\"category\": \"car\", \"corners_3d\": [[623.44, 247.61], [699.12, 247.54], [725.83, 268.28], [629.15, 268.4], [623.44, 175.31], [699.12, 175.31], [725.83, 175.99], [629.15, 176.0]]}]\n```" - } - ] - }, - { - "id": 2645, - "data_source": "Mono3DRefer", - "scene_name": "003102", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[239.87, 402.33], [-73.2, 407.27], [302.17, 271.07], [431.62, 270.2], [239.87, -49.77], [-73.2, -54.56], [302.17, 77.57], [431.62, 78.42]]}, {\"category\": \"car\", \"corners_3d\": [[623.44, 247.61], [699.12, 247.54], [725.83, 268.28], [629.15, 268.4], [623.44, 175.31], [699.12, 175.31], [725.83, 175.99], [629.15, 176.0]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[602.11, 258.22], [623.75, 256.33], [637.87, 258.47], [615.99, 260.46], [602.11, 166.27], [623.75, 166.41], [637.87, 166.25], [615.99, 166.09]]}]\n```" - } - ] - }, - { - "id": 2646, - "data_source": "Mono3DRefer", - "scene_name": "003104", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[115.04, 595.07], [-301.27, 611.54], [200.02, 355.73], [369.43, 352.81], [115.04, 154.31], [-301.27, 153.59], [200.02, 164.82], [369.43, 164.95]]}, {\"category\": \"car\", \"corners_3d\": [[448.9, 332.62], [312.75, 333.24], [403.59, 281.54], [495.73, 281.26], [448.9, 186.41], [312.75, 186.46], [403.59, 182.07], [495.73, 182.05]]}, {\"category\": \"car\", \"corners_3d\": [[570.04, 241.86], [524.57, 240.93], [557.0, 231.66], [596.59, 232.35], [570.04, 197.7], [524.57, 197.37], [557.0, 194.03], [596.59, 194.28]]}, {\"category\": \"car\", \"corners_3d\": [[607.16, 231.21], [575.64, 230.45], [600.14, 224.78], [628.81, 225.39], [607.16, 194.85], [575.64, 194.56], [600.14, 192.42], [628.81, 192.66]]}, {\"category\": \"car\", \"corners_3d\": [[651.33, 219.38], [622.95, 218.74], [642.73, 214.98], [669.0, 215.51], [651.33, 190.27], [622.95, 190.03], [642.73, 188.63], [669.0, 188.82]]}, {\"category\": \"car\", \"corners_3d\": [[804.43, 198.41], [823.46, 198.79], [806.35, 200.43], [786.4, 200.0], [804.43, 179.08], [823.46, 179.17], [806.35, 179.57], [786.4, 179.46]]}]\n```" - } - ] - }, - { - "id": 2647, - "data_source": "Mono3DRefer", - "scene_name": "003108", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[612.06, 187.85], [598.08, 187.78], [607.36, 187.04], [620.69, 187.1], [612.06, 174.67], [598.08, 174.66], [607.36, 174.57], [620.69, 174.58]]}, {\"category\": \"car\", \"corners_3d\": [[720.48, 187.6], [736.56, 187.66], [734.9, 188.56], [717.86, 188.49], [720.48, 173.35], [736.56, 173.35], [734.9, 173.38], [717.86, 173.38]]}]\n```" - } - ] - }, - { - "id": 2648, - "data_source": "Mono3DRefer", - "scene_name": "003121", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[688.83, 293.03], [693.23, 303.05], [600.5, 302.25], [603.66, 292.36], [688.83, 158.46], [693.23, 156.5], [600.5, 156.66], [603.66, 158.59]]}]\n```" - } - ] - }, - { - "id": 2649, - "data_source": "Mono3DRefer", - "scene_name": "003123", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[444.1, 221.84], [430.02, 225.03], [308.04, 224.4], [329.33, 221.28], [444.1, 186.01], [430.02, 186.87], [308.04, 186.7], [329.33, 185.87]]}]\n```" - } - ] - }, - { - "id": 2650, - "data_source": "Mono3DRefer", - "scene_name": "003145", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[188.74, 224.77], [153.09, 225.89], [155.73, 221.64], [188.53, 220.69], [188.74, 197.78], [153.09, 198.32], [155.73, 196.28], [188.53, 195.83]]}]\n```" - } - ] - }, - { - "id": 2651, - "data_source": "Mono3DRefer", - "scene_name": "003152", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[484.32, 265.87], [394.32, 265.87], [441.89, 245.26], [511.94, 245.26], [484.32, 192.15], [394.32, 192.16], [441.89, 187.88], [511.94, 187.88]]}]\n```" - } - ] - }, - { - "id": 2652, - "data_source": "Mono3DRefer", - "scene_name": "003159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1005.22, 442.72], [1253.52, 439.0], [4639.39, 1779.46], [3334.5, 1925.39], [1005.22, 195.88], [1253.52, 195.56], [4639.39, 309.91], [3334.5, 322.36]]}, {\"category\": \"car\", \"corners_3d\": [[-18.56, 244.17], [21.09, 239.78], [123.35, 239.99], [90.31, 244.41], [-18.56, 207.95], [21.09, 205.79], [123.35, 205.89], [90.31, 208.07]]}, {\"category\": \"car\", \"corners_3d\": [[61.78, 262.98], [12.11, 271.27], [-175.64, 271.52], [-110.01, 263.2], [61.78, 213.57], [12.11, 217.32], [-175.64, 217.43], [-110.01, 213.67]]}, {\"category\": \"car\", \"corners_3d\": [[837.32, 205.4], [826.49, 203.74], [914.81, 203.82], [930.43, 205.49], [837.32, 167.96], [826.49, 168.21], [914.81, 168.2], [930.43, 167.95]]}, {\"category\": \"car\", \"corners_3d\": [[969.57, 207.69], [945.18, 205.47], [1063.48, 205.3], [1095.73, 207.5], [969.57, 164.34], [945.18, 164.88], [1063.48, 164.92], [1095.73, 164.39]]}]\n```" - } - ] - }, - { - "id": 2653, - "data_source": "Mono3DRefer", - "scene_name": "003159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[719.78, 190.0], [710.71, 188.76], [881.52, 188.68], [903.75, 189.9], [719.78, 111.71], [710.71, 116.12], [881.52, 116.41], [903.75, 112.05]]}]\n```" - } - ] - }, - { - "id": 2654, - "data_source": "Mono3DRefer", - "scene_name": "003159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[188.74, 236.1], [219.38, 231.8], [337.12, 232.21], [314.9, 236.57], [188.74, 181.76], [219.38, 181.16], [337.12, 181.21], [314.9, 181.83]]}]\n```" - } - ] - }, - { - "id": 2655, - "data_source": "Mono3DRefer", - "scene_name": "003159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[978.99, 288.77], [1041.2, 291.12], [1064.62, 302.7], [995.94, 299.88], [978.99, 128.73], [1041.2, 127.84], [1064.62, 123.43], [995.94, 124.5]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1041.46, 288.49], [1104.41, 290.89], [1131.83, 301.88], [1062.55, 299.02], [1041.46, 124.7], [1104.41, 123.7], [1131.83, 119.12], [1062.55, 120.32]]}]\n```" - } - ] - }, - { - "id": 2656, - "data_source": "Mono3DRefer", - "scene_name": "003159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1005.22, 442.72], [1253.52, 439.0], [4639.39, 1779.46], [3334.5, 1925.39], [1005.22, 195.88], [1253.52, 195.56], [4639.39, 309.91], [3334.5, 322.36]]}, {\"category\": \"car\", \"corners_3d\": [[-18.56, 244.17], [21.09, 239.78], [123.35, 239.99], [90.31, 244.41], [-18.56, 207.95], [21.09, 205.79], [123.35, 205.89], [90.31, 208.07]]}, {\"category\": \"car\", \"corners_3d\": [[61.78, 262.98], [12.11, 271.27], [-175.64, 271.52], [-110.01, 263.2], [61.78, 213.57], [12.11, 217.32], [-175.64, 217.43], [-110.01, 213.67]]}, {\"category\": \"car\", \"corners_3d\": [[837.32, 205.4], [826.49, 203.74], [914.81, 203.82], [930.43, 205.49], [837.32, 167.96], [826.49, 168.21], [914.81, 168.2], [930.43, 167.95]]}, {\"category\": \"truck\", \"corners_3d\": [[719.78, 190.0], [710.71, 188.76], [881.52, 188.68], [903.75, 189.9], [719.78, 111.71], [710.71, 116.12], [881.52, 116.41], [903.75, 112.05]]}, {\"category\": \"car\", \"corners_3d\": [[969.57, 207.69], [945.18, 205.47], [1063.48, 205.3], [1095.73, 207.5], [969.57, 164.34], [945.18, 164.88], [1063.48, 164.92], [1095.73, 164.39]]}]\n```" - } - ] - }, - { - "id": 2657, - "data_source": "Mono3DRefer", - "scene_name": "003159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1005.22, 442.72], [1253.52, 439.0], [4639.39, 1779.46], [3334.5, 1925.39], [1005.22, 195.88], [1253.52, 195.56], [4639.39, 309.91], [3334.5, 322.36]]}, {\"category\": \"car\", \"corners_3d\": [[-18.56, 244.17], [21.09, 239.78], [123.35, 239.99], [90.31, 244.41], [-18.56, 207.95], [21.09, 205.79], [123.35, 205.89], [90.31, 208.07]]}, {\"category\": \"car\", \"corners_3d\": [[61.78, 262.98], [12.11, 271.27], [-175.64, 271.52], [-110.01, 263.2], [61.78, 213.57], [12.11, 217.32], [-175.64, 217.43], [-110.01, 213.67]]}, {\"category\": \"car\", \"corners_3d\": [[837.32, 205.4], [826.49, 203.74], [914.81, 203.82], [930.43, 205.49], [837.32, 167.96], [826.49, 168.21], [914.81, 168.2], [930.43, 167.95]]}, {\"category\": \"car\", \"corners_3d\": [[969.57, 207.69], [945.18, 205.47], [1063.48, 205.3], [1095.73, 207.5], [969.57, 164.34], [945.18, 164.88], [1063.48, 164.92], [1095.73, 164.39]]}, {\"category\": \"van\", \"corners_3d\": [[188.74, 236.1], [219.38, 231.8], [337.12, 232.21], [314.9, 236.57], [188.74, 181.76], [219.38, 181.16], [337.12, 181.21], [314.9, 181.83]]}]\n```" - } - ] - }, - { - "id": 2658, - "data_source": "Mono3DRefer", - "scene_name": "003159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[978.99, 288.77], [1041.2, 291.12], [1064.62, 302.7], [995.94, 299.88], [978.99, 128.73], [1041.2, 127.84], [1064.62, 123.43], [995.94, 124.5]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1041.46, 288.49], [1104.41, 290.89], [1131.83, 301.88], [1062.55, 299.02], [1041.46, 124.7], [1104.41, 123.7], [1131.83, 119.12], [1062.55, 120.32]]}, {\"category\": \"car\", \"corners_3d\": [[1005.22, 442.72], [1253.52, 439.0], [4639.39, 1779.46], [3334.5, 1925.39], [1005.22, 195.88], [1253.52, 195.56], [4639.39, 309.91], [3334.5, 322.36]]}, {\"category\": \"car\", \"corners_3d\": [[-18.56, 244.17], [21.09, 239.78], [123.35, 239.99], [90.31, 244.41], [-18.56, 207.95], [21.09, 205.79], [123.35, 205.89], [90.31, 208.07]]}, {\"category\": \"car\", \"corners_3d\": [[61.78, 262.98], [12.11, 271.27], [-175.64, 271.52], [-110.01, 263.2], [61.78, 213.57], [12.11, 217.32], [-175.64, 217.43], [-110.01, 213.67]]}, {\"category\": \"car\", \"corners_3d\": [[837.32, 205.4], [826.49, 203.74], [914.81, 203.82], [930.43, 205.49], [837.32, 167.96], [826.49, 168.21], [914.81, 168.2], [930.43, 167.95]]}, {\"category\": \"car\", \"corners_3d\": [[969.57, 207.69], [945.18, 205.47], [1063.48, 205.3], [1095.73, 207.5], [969.57, 164.34], [945.18, 164.88], [1063.48, 164.92], [1095.73, 164.39]]}]\n```" - } - ] - }, - { - "id": 2659, - "data_source": "Mono3DRefer", - "scene_name": "003159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[719.78, 190.0], [710.71, 188.76], [881.52, 188.68], [903.75, 189.9], [719.78, 111.71], [710.71, 116.12], [881.52, 116.41], [903.75, 112.05]]}, {\"category\": \"van\", \"corners_3d\": [[188.74, 236.1], [219.38, 231.8], [337.12, 232.21], [314.9, 236.57], [188.74, 181.76], [219.38, 181.16], [337.12, 181.21], [314.9, 181.83]]}]\n```" - } - ] - }, - { - "id": 2660, - "data_source": "Mono3DRefer", - "scene_name": "003159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[978.99, 288.77], [1041.2, 291.12], [1064.62, 302.7], [995.94, 299.88], [978.99, 128.73], [1041.2, 127.84], [1064.62, 123.43], [995.94, 124.5]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1041.46, 288.49], [1104.41, 290.89], [1131.83, 301.88], [1062.55, 299.02], [1041.46, 124.7], [1104.41, 123.7], [1131.83, 119.12], [1062.55, 120.32]]}, {\"category\": \"truck\", \"corners_3d\": [[719.78, 190.0], [710.71, 188.76], [881.52, 188.68], [903.75, 189.9], [719.78, 111.71], [710.71, 116.12], [881.52, 116.41], [903.75, 112.05]]}]\n```" - } - ] - }, - { - "id": 2661, - "data_source": "Mono3DRefer", - "scene_name": "003159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[978.99, 288.77], [1041.2, 291.12], [1064.62, 302.7], [995.94, 299.88], [978.99, 128.73], [1041.2, 127.84], [1064.62, 123.43], [995.94, 124.5]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1041.46, 288.49], [1104.41, 290.89], [1131.83, 301.88], [1062.55, 299.02], [1041.46, 124.7], [1104.41, 123.7], [1131.83, 119.12], [1062.55, 120.32]]}, {\"category\": \"van\", \"corners_3d\": [[188.74, 236.1], [219.38, 231.8], [337.12, 232.21], [314.9, 236.57], [188.74, 181.76], [219.38, 181.16], [337.12, 181.21], [314.9, 181.83]]}]\n```" - } - ] - }, - { - "id": 2662, - "data_source": "Mono3DRefer", - "scene_name": "003159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1005.22, 442.72], [1253.52, 439.0], [4639.39, 1779.46], [3334.5, 1925.39], [1005.22, 195.88], [1253.52, 195.56], [4639.39, 309.91], [3334.5, 322.36]]}, {\"category\": \"car\", \"corners_3d\": [[-18.56, 244.17], [21.09, 239.78], [123.35, 239.99], [90.31, 244.41], [-18.56, 207.95], [21.09, 205.79], [123.35, 205.89], [90.31, 208.07]]}, {\"category\": \"car\", \"corners_3d\": [[61.78, 262.98], [12.11, 271.27], [-175.64, 271.52], [-110.01, 263.2], [61.78, 213.57], [12.11, 217.32], [-175.64, 217.43], [-110.01, 213.67]]}, {\"category\": \"car\", \"corners_3d\": [[837.32, 205.4], [826.49, 203.74], [914.81, 203.82], [930.43, 205.49], [837.32, 167.96], [826.49, 168.21], [914.81, 168.2], [930.43, 167.95]]}, {\"category\": \"truck\", \"corners_3d\": [[719.78, 190.0], [710.71, 188.76], [881.52, 188.68], [903.75, 189.9], [719.78, 111.71], [710.71, 116.12], [881.52, 116.41], [903.75, 112.05]]}, {\"category\": \"car\", \"corners_3d\": [[969.57, 207.69], [945.18, 205.47], [1063.48, 205.3], [1095.73, 207.5], [969.57, 164.34], [945.18, 164.88], [1063.48, 164.92], [1095.73, 164.39]]}, {\"category\": \"van\", \"corners_3d\": [[188.74, 236.1], [219.38, 231.8], [337.12, 232.21], [314.9, 236.57], [188.74, 181.76], [219.38, 181.16], [337.12, 181.21], [314.9, 181.83]]}]\n```" - } - ] - }, - { - "id": 2663, - "data_source": "Mono3DRefer", - "scene_name": "003159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[978.99, 288.77], [1041.2, 291.12], [1064.62, 302.7], [995.94, 299.88], [978.99, 128.73], [1041.2, 127.84], [1064.62, 123.43], [995.94, 124.5]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1041.46, 288.49], [1104.41, 290.89], [1131.83, 301.88], [1062.55, 299.02], [1041.46, 124.7], [1104.41, 123.7], [1131.83, 119.12], [1062.55, 120.32]]}, {\"category\": \"car\", \"corners_3d\": [[1005.22, 442.72], [1253.52, 439.0], [4639.39, 1779.46], [3334.5, 1925.39], [1005.22, 195.88], [1253.52, 195.56], [4639.39, 309.91], [3334.5, 322.36]]}, {\"category\": \"car\", \"corners_3d\": [[-18.56, 244.17], [21.09, 239.78], [123.35, 239.99], [90.31, 244.41], [-18.56, 207.95], [21.09, 205.79], [123.35, 205.89], [90.31, 208.07]]}, {\"category\": \"car\", \"corners_3d\": [[61.78, 262.98], [12.11, 271.27], [-175.64, 271.52], [-110.01, 263.2], [61.78, 213.57], [12.11, 217.32], [-175.64, 217.43], [-110.01, 213.67]]}, {\"category\": \"car\", \"corners_3d\": [[837.32, 205.4], [826.49, 203.74], [914.81, 203.82], [930.43, 205.49], [837.32, 167.96], [826.49, 168.21], [914.81, 168.2], [930.43, 167.95]]}, {\"category\": \"truck\", \"corners_3d\": [[719.78, 190.0], [710.71, 188.76], [881.52, 188.68], [903.75, 189.9], [719.78, 111.71], [710.71, 116.12], [881.52, 116.41], [903.75, 112.05]]}, {\"category\": \"car\", \"corners_3d\": [[969.57, 207.69], [945.18, 205.47], [1063.48, 205.3], [1095.73, 207.5], [969.57, 164.34], [945.18, 164.88], [1063.48, 164.92], [1095.73, 164.39]]}]\n```" - } - ] - }, - { - "id": 2664, - "data_source": "Mono3DRefer", - "scene_name": "003159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[978.99, 288.77], [1041.2, 291.12], [1064.62, 302.7], [995.94, 299.88], [978.99, 128.73], [1041.2, 127.84], [1064.62, 123.43], [995.94, 124.5]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1041.46, 288.49], [1104.41, 290.89], [1131.83, 301.88], [1062.55, 299.02], [1041.46, 124.7], [1104.41, 123.7], [1131.83, 119.12], [1062.55, 120.32]]}, {\"category\": \"car\", \"corners_3d\": [[1005.22, 442.72], [1253.52, 439.0], [4639.39, 1779.46], [3334.5, 1925.39], [1005.22, 195.88], [1253.52, 195.56], [4639.39, 309.91], [3334.5, 322.36]]}, {\"category\": \"car\", \"corners_3d\": [[-18.56, 244.17], [21.09, 239.78], [123.35, 239.99], [90.31, 244.41], [-18.56, 207.95], [21.09, 205.79], [123.35, 205.89], [90.31, 208.07]]}, {\"category\": \"car\", \"corners_3d\": [[61.78, 262.98], [12.11, 271.27], [-175.64, 271.52], [-110.01, 263.2], [61.78, 213.57], [12.11, 217.32], [-175.64, 217.43], [-110.01, 213.67]]}, {\"category\": \"car\", \"corners_3d\": [[837.32, 205.4], [826.49, 203.74], [914.81, 203.82], [930.43, 205.49], [837.32, 167.96], [826.49, 168.21], [914.81, 168.2], [930.43, 167.95]]}, {\"category\": \"car\", \"corners_3d\": [[969.57, 207.69], [945.18, 205.47], [1063.48, 205.3], [1095.73, 207.5], [969.57, 164.34], [945.18, 164.88], [1063.48, 164.92], [1095.73, 164.39]]}, {\"category\": \"van\", \"corners_3d\": [[188.74, 236.1], [219.38, 231.8], [337.12, 232.21], [314.9, 236.57], [188.74, 181.76], [219.38, 181.16], [337.12, 181.21], [314.9, 181.83]]}]\n```" - } - ] - }, - { - "id": 2665, - "data_source": "Mono3DRefer", - "scene_name": "003159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[978.99, 288.77], [1041.2, 291.12], [1064.62, 302.7], [995.94, 299.88], [978.99, 128.73], [1041.2, 127.84], [1064.62, 123.43], [995.94, 124.5]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1041.46, 288.49], [1104.41, 290.89], [1131.83, 301.88], [1062.55, 299.02], [1041.46, 124.7], [1104.41, 123.7], [1131.83, 119.12], [1062.55, 120.32]]}, {\"category\": \"truck\", \"corners_3d\": [[719.78, 190.0], [710.71, 188.76], [881.52, 188.68], [903.75, 189.9], [719.78, 111.71], [710.71, 116.12], [881.52, 116.41], [903.75, 112.05]]}, {\"category\": \"van\", \"corners_3d\": [[188.74, 236.1], [219.38, 231.8], [337.12, 232.21], [314.9, 236.57], [188.74, 181.76], [219.38, 181.16], [337.12, 181.21], [314.9, 181.83]]}]\n```" - } - ] - }, - { - "id": 2666, - "data_source": "Mono3DRefer", - "scene_name": "003159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[978.99, 288.77], [1041.2, 291.12], [1064.62, 302.7], [995.94, 299.88], [978.99, 128.73], [1041.2, 127.84], [1064.62, 123.43], [995.94, 124.5]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1041.46, 288.49], [1104.41, 290.89], [1131.83, 301.88], [1062.55, 299.02], [1041.46, 124.7], [1104.41, 123.7], [1131.83, 119.12], [1062.55, 120.32]]}, {\"category\": \"car\", \"corners_3d\": [[1005.22, 442.72], [1253.52, 439.0], [4639.39, 1779.46], [3334.5, 1925.39], [1005.22, 195.88], [1253.52, 195.56], [4639.39, 309.91], [3334.5, 322.36]]}, {\"category\": \"car\", \"corners_3d\": [[-18.56, 244.17], [21.09, 239.78], [123.35, 239.99], [90.31, 244.41], [-18.56, 207.95], [21.09, 205.79], [123.35, 205.89], [90.31, 208.07]]}, {\"category\": \"car\", \"corners_3d\": [[61.78, 262.98], [12.11, 271.27], [-175.64, 271.52], [-110.01, 263.2], [61.78, 213.57], [12.11, 217.32], [-175.64, 217.43], [-110.01, 213.67]]}, {\"category\": \"car\", \"corners_3d\": [[837.32, 205.4], [826.49, 203.74], [914.81, 203.82], [930.43, 205.49], [837.32, 167.96], [826.49, 168.21], [914.81, 168.2], [930.43, 167.95]]}, {\"category\": \"truck\", \"corners_3d\": [[719.78, 190.0], [710.71, 188.76], [881.52, 188.68], [903.75, 189.9], [719.78, 111.71], [710.71, 116.12], [881.52, 116.41], [903.75, 112.05]]}, {\"category\": \"car\", \"corners_3d\": [[969.57, 207.69], [945.18, 205.47], [1063.48, 205.3], [1095.73, 207.5], [969.57, 164.34], [945.18, 164.88], [1063.48, 164.92], [1095.73, 164.39]]}, {\"category\": \"van\", \"corners_3d\": [[188.74, 236.1], [219.38, 231.8], [337.12, 232.21], [314.9, 236.57], [188.74, 181.76], [219.38, 181.16], [337.12, 181.21], [314.9, 181.83]]}]\n```" - } - ] - }, - { - "id": 2667, - "data_source": "Mono3DRefer", - "scene_name": "003161", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[552.21, 240.3], [509.41, 240.38], [520.72, 231.41], [557.82, 231.35], [552.21, 200.77], [509.41, 200.81], [520.72, 197.09], [557.82, 197.07]]}]\n```" - } - ] - }, - { - "id": 2668, - "data_source": "Mono3DRefer", - "scene_name": "003165", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[632.4, 197.48], [670.83, 197.53], [681.42, 205.74], [630.21, 205.64], [632.4, 155.94], [670.83, 155.91], [681.42, 150.27], [630.21, 150.34]]}]\n```" - } - ] - }, - { - "id": 2669, - "data_source": "Mono3DRefer", - "scene_name": "003165", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[247.0, 234.53], [201.41, 234.53], [251.16, 227.0], [291.18, 227.0], [247.0, 194.24], [201.41, 194.24], [251.16, 191.63], [291.18, 191.63]]}, {\"category\": \"car\", \"corners_3d\": [[416.51, 206.75], [388.38, 206.76], [405.73, 204.09], [431.65, 204.09], [416.51, 183.51], [388.38, 183.51], [405.73, 182.67], [431.65, 182.67]]}]\n```" - } - ] - }, - { - "id": 2670, - "data_source": "Mono3DRefer", - "scene_name": "003165", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[632.4, 197.48], [670.83, 197.53], [681.42, 205.74], [630.21, 205.64], [632.4, 155.94], [670.83, 155.91], [681.42, 150.27], [630.21, 150.34]]}, {\"category\": \"car\", \"corners_3d\": [[247.0, 234.53], [201.41, 234.53], [251.16, 227.0], [291.18, 227.0], [247.0, 194.24], [201.41, 194.24], [251.16, 191.63], [291.18, 191.63]]}, {\"category\": \"car\", \"corners_3d\": [[416.51, 206.75], [388.38, 206.76], [405.73, 204.09], [431.65, 204.09], [416.51, 183.51], [388.38, 183.51], [405.73, 182.67], [431.65, 182.67]]}]\n```" - } - ] - }, - { - "id": 2671, - "data_source": "Mono3DRefer", - "scene_name": "003166", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[366.39, 237.41], [389.35, 237.31], [378.58, 241.07], [354.26, 241.18], [366.39, 173.86], [389.35, 173.85], [378.58, 173.91], [354.26, 173.91]]}]\n```" - } - ] - }, - { - "id": 2672, - "data_source": "Mono3DRefer", - "scene_name": "003166", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1356.18, 271.85], [1475.36, 288.12], [1106.17, 286.41], [1039.97, 270.59], [1356.18, 169.21], [1475.36, 168.62], [1106.17, 168.68], [1039.97, 169.26]]}]\n```" - } - ] - }, - { - "id": 2673, - "data_source": "Mono3DRefer", - "scene_name": "003166", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[366.39, 237.41], [389.35, 237.31], [378.58, 241.07], [354.26, 241.18], [366.39, 173.86], [389.35, 173.85], [378.58, 173.91], [354.26, 173.91]]}, {\"category\": \"car\", \"corners_3d\": [[1356.18, 271.85], [1475.36, 288.12], [1106.17, 286.41], [1039.97, 270.59], [1356.18, 169.21], [1475.36, 168.62], [1106.17, 168.68], [1039.97, 169.26]]}]\n```" - } - ] - }, - { - "id": 2674, - "data_source": "Mono3DRefer", - "scene_name": "003172", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[695.28, 207.61], [682.85, 208.02], [671.32, 207.25], [683.61, 206.86], [695.28, 169.21], [682.85, 169.17], [671.32, 169.25], [683.61, 169.29]]}]\n```" - } - ] - }, - { - "id": 2675, - "data_source": "Mono3DRefer", - "scene_name": "003172", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[107.44, 349.12], [-61.85, 349.97], [163.32, 289.21], [274.01, 288.85], [107.44, 197.28], [-61.85, 197.39], [163.32, 188.98], [274.01, 188.92]]}, {\"category\": \"car\", \"corners_3d\": [[829.95, 331.05], [984.9, 329.35], [1365.33, 474.25], [1071.91, 480.62], [829.95, 183.7], [984.9, 183.58], [1365.33, 193.51], [1071.91, 193.95]]}, {\"category\": \"car\", \"corners_3d\": [[519.23, 208.87], [486.93, 209.66], [465.02, 206.48], [495.02, 205.82], [519.23, 178.81], [486.93, 178.95], [465.02, 178.42], [495.02, 178.31]]}, {\"category\": \"car\", \"corners_3d\": [[555.64, 195.18], [544.26, 194.55], [595.22, 193.64], [607.53, 194.22], [555.64, 173.86], [544.26, 173.84], [595.22, 173.79], [607.53, 173.82]]}, {\"category\": \"car\", \"corners_3d\": [[418.6, 204.72], [398.02, 205.96], [342.07, 204.56], [363.82, 203.42], [418.6, 179.99], [398.02, 180.27], [342.07, 179.95], [363.82, 179.7]]}, {\"category\": \"car\", \"corners_3d\": [[658.14, 193.2], [649.33, 192.54], [705.91, 192.06], [716.33, 192.68], [658.14, 173.25], [649.33, 173.24], [705.91, 173.23], [716.33, 173.24]]}]\n```" - } - ] - }, - { - "id": 2676, - "data_source": "Mono3DRefer", - "scene_name": "003172", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[541.59, 194.33], [544.16, 193.5], [624.02, 193.5], [624.65, 194.32], [541.59, 151.85], [544.16, 152.66], [624.02, 152.66], [624.65, 151.85]]}]\n```" - } - ] - }, - { - "id": 2677, - "data_source": "Mono3DRefer", - "scene_name": "003172", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[107.44, 349.12], [-61.85, 349.97], [163.32, 289.21], [274.01, 288.85], [107.44, 197.28], [-61.85, 197.39], [163.32, 188.98], [274.01, 188.92]]}, {\"category\": \"car\", \"corners_3d\": [[829.95, 331.05], [984.9, 329.35], [1365.33, 474.25], [1071.91, 480.62], [829.95, 183.7], [984.9, 183.58], [1365.33, 193.51], [1071.91, 193.95]]}, {\"category\": \"car\", \"corners_3d\": [[519.23, 208.87], [486.93, 209.66], [465.02, 206.48], [495.02, 205.82], [519.23, 178.81], [486.93, 178.95], [465.02, 178.42], [495.02, 178.31]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[695.28, 207.61], [682.85, 208.02], [671.32, 207.25], [683.61, 206.86], [695.28, 169.21], [682.85, 169.17], [671.32, 169.25], [683.61, 169.29]]}, {\"category\": \"car\", \"corners_3d\": [[555.64, 195.18], [544.26, 194.55], [595.22, 193.64], [607.53, 194.22], [555.64, 173.86], [544.26, 173.84], [595.22, 173.79], [607.53, 173.82]]}, {\"category\": \"car\", \"corners_3d\": [[418.6, 204.72], [398.02, 205.96], [342.07, 204.56], [363.82, 203.42], [418.6, 179.99], [398.02, 180.27], [342.07, 179.95], [363.82, 179.7]]}, {\"category\": \"car\", \"corners_3d\": [[658.14, 193.2], [649.33, 192.54], [705.91, 192.06], [716.33, 192.68], [658.14, 173.25], [649.33, 173.24], [705.91, 173.23], [716.33, 173.24]]}]\n```" - } - ] - }, - { - "id": 2678, - "data_source": "Mono3DRefer", - "scene_name": "003172", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[695.28, 207.61], [682.85, 208.02], [671.32, 207.25], [683.61, 206.86], [695.28, 169.21], [682.85, 169.17], [671.32, 169.25], [683.61, 169.29]]}, {\"category\": \"truck\", \"corners_3d\": [[541.59, 194.33], [544.16, 193.5], [624.02, 193.5], [624.65, 194.32], [541.59, 151.85], [544.16, 152.66], [624.02, 152.66], [624.65, 151.85]]}]\n```" - } - ] - }, - { - "id": 2679, - "data_source": "Mono3DRefer", - "scene_name": "003172", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[107.44, 349.12], [-61.85, 349.97], [163.32, 289.21], [274.01, 288.85], [107.44, 197.28], [-61.85, 197.39], [163.32, 188.98], [274.01, 188.92]]}, {\"category\": \"car\", \"corners_3d\": [[829.95, 331.05], [984.9, 329.35], [1365.33, 474.25], [1071.91, 480.62], [829.95, 183.7], [984.9, 183.58], [1365.33, 193.51], [1071.91, 193.95]]}, {\"category\": \"car\", \"corners_3d\": [[519.23, 208.87], [486.93, 209.66], [465.02, 206.48], [495.02, 205.82], [519.23, 178.81], [486.93, 178.95], [465.02, 178.42], [495.02, 178.31]]}, {\"category\": \"car\", \"corners_3d\": [[555.64, 195.18], [544.26, 194.55], [595.22, 193.64], [607.53, 194.22], [555.64, 173.86], [544.26, 173.84], [595.22, 173.79], [607.53, 173.82]]}, {\"category\": \"car\", \"corners_3d\": [[418.6, 204.72], [398.02, 205.96], [342.07, 204.56], [363.82, 203.42], [418.6, 179.99], [398.02, 180.27], [342.07, 179.95], [363.82, 179.7]]}, {\"category\": \"car\", \"corners_3d\": [[658.14, 193.2], [649.33, 192.54], [705.91, 192.06], [716.33, 192.68], [658.14, 173.25], [649.33, 173.24], [705.91, 173.23], [716.33, 173.24]]}, {\"category\": \"truck\", \"corners_3d\": [[541.59, 194.33], [544.16, 193.5], [624.02, 193.5], [624.65, 194.32], [541.59, 151.85], [544.16, 152.66], [624.02, 152.66], [624.65, 151.85]]}]\n```" - } - ] - }, - { - "id": 2680, - "data_source": "Mono3DRefer", - "scene_name": "003172", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[107.44, 349.12], [-61.85, 349.97], [163.32, 289.21], [274.01, 288.85], [107.44, 197.28], [-61.85, 197.39], [163.32, 188.98], [274.01, 188.92]]}, {\"category\": \"car\", \"corners_3d\": [[829.95, 331.05], [984.9, 329.35], [1365.33, 474.25], [1071.91, 480.62], [829.95, 183.7], [984.9, 183.58], [1365.33, 193.51], [1071.91, 193.95]]}, {\"category\": \"car\", \"corners_3d\": [[519.23, 208.87], [486.93, 209.66], [465.02, 206.48], [495.02, 205.82], [519.23, 178.81], [486.93, 178.95], [465.02, 178.42], [495.02, 178.31]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[695.28, 207.61], [682.85, 208.02], [671.32, 207.25], [683.61, 206.86], [695.28, 169.21], [682.85, 169.17], [671.32, 169.25], [683.61, 169.29]]}, {\"category\": \"car\", \"corners_3d\": [[555.64, 195.18], [544.26, 194.55], [595.22, 193.64], [607.53, 194.22], [555.64, 173.86], [544.26, 173.84], [595.22, 173.79], [607.53, 173.82]]}, {\"category\": \"car\", \"corners_3d\": [[418.6, 204.72], [398.02, 205.96], [342.07, 204.56], [363.82, 203.42], [418.6, 179.99], [398.02, 180.27], [342.07, 179.95], [363.82, 179.7]]}, {\"category\": \"car\", \"corners_3d\": [[658.14, 193.2], [649.33, 192.54], [705.91, 192.06], [716.33, 192.68], [658.14, 173.25], [649.33, 173.24], [705.91, 173.23], [716.33, 173.24]]}, {\"category\": \"truck\", \"corners_3d\": [[541.59, 194.33], [544.16, 193.5], [624.02, 193.5], [624.65, 194.32], [541.59, 151.85], [544.16, 152.66], [624.02, 152.66], [624.65, 151.85]]}]\n```" - } - ] - }, - { - "id": 2681, - "data_source": "Mono3DRefer", - "scene_name": "003188", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[554.1, 196.58], [536.45, 196.58], [540.29, 195.19], [556.89, 195.19], [554.1, 177.93], [536.45, 177.94], [540.29, 177.64], [556.89, 177.64]]}]\n```" - } - ] - }, - { - "id": 2682, - "data_source": "Mono3DRefer", - "scene_name": "003191", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[484.39, 262.06], [401.06, 261.96], [445.73, 243.46], [511.78, 243.52], [484.39, 187.79], [401.06, 187.78], [445.73, 184.68], [511.78, 184.69]]}, {\"category\": \"car\", \"corners_3d\": [[539.48, 228.72], [488.13, 228.69], [505.97, 220.91], [550.17, 220.94], [539.48, 180.78], [488.13, 180.78], [505.97, 179.67], [550.17, 179.68]]}]\n```" - } - ] - }, - { - "id": 2683, - "data_source": "Mono3DRefer", - "scene_name": "003205", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[269.22, 561.99], [-178.73, 585.31], [203.55, 364.88], [408.23, 359.67], [269.22, 179.72], [-178.73, 180.13], [203.55, 176.24], [408.23, 176.15]]}, {\"category\": \"car\", \"corners_3d\": [[480.38, 294.95], [371.12, 294.59], [442.65, 260.08], [521.02, 260.26], [480.38, 182.83], [371.12, 182.8], [442.65, 179.98], [521.02, 179.99]]}]\n```" - } - ] - }, - { - "id": 2684, - "data_source": "Mono3DRefer", - "scene_name": "003217", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[284.45, 225.54], [243.4, 225.51], [287.11, 219.33], [323.36, 219.36], [284.45, 191.85], [243.4, 191.84], [287.11, 189.61], [323.36, 189.62]]}]\n```" - } - ] - }, - { - "id": 2685, - "data_source": "Mono3DRefer", - "scene_name": "003217", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[682.13, 223.91], [761.79, 223.97], [916.64, 278.78], [751.42, 278.52], [682.13, 137.79], [761.79, 137.75], [916.64, 100.11], [751.42, 100.29]]}]\n```" - } - ] - }, - { - "id": 2686, - "data_source": "Mono3DRefer", - "scene_name": "003217", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[682.13, 223.91], [761.79, 223.97], [916.64, 278.78], [751.42, 278.52], [682.13, 137.79], [761.79, 137.75], [916.64, 100.11], [751.42, 100.29]]}, {\"category\": \"car\", \"corners_3d\": [[284.45, 225.54], [243.4, 225.51], [287.11, 219.33], [323.36, 219.36], [284.45, 191.85], [243.4, 191.84], [287.11, 189.61], [323.36, 189.62]]}]\n```" - } - ] - }, - { - "id": 2687, - "data_source": "Mono3DRefer", - "scene_name": "003222", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[203.01, 213.9], [191.34, 213.8], [213.8, 212.01], [225.0, 212.09], [203.01, 177.57], [191.34, 177.56], [213.8, 177.35], [225.0, 177.36]]}]\n```" - } - ] - }, - { - "id": 2688, - "data_source": "Mono3DRefer", - "scene_name": "003222", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[162.77, 232.19], [113.07, 232.02], [177.02, 224.81], [220.8, 224.94], [162.77, 189.59], [113.07, 189.54], [177.02, 187.51], [220.8, 187.55]]}]\n```" - } - ] - }, - { - "id": 2689, - "data_source": "Mono3DRefer", - "scene_name": "003222", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[303.49, 205.03], [272.87, 204.99], [308.81, 201.77], [336.39, 201.8], [303.49, 169.78], [272.87, 169.78], [308.81, 170.09], [336.39, 170.08]]}]\n```" - } - ] - }, - { - "id": 2690, - "data_source": "Mono3DRefer", - "scene_name": "003222", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[162.77, 232.19], [113.07, 232.02], [177.02, 224.81], [220.8, 224.94], [162.77, 189.59], [113.07, 189.54], [177.02, 187.51], [220.8, 187.55]]}, {\"category\": \"cyclist\", \"corners_3d\": [[203.01, 213.9], [191.34, 213.8], [213.8, 212.01], [225.0, 212.09], [203.01, 177.57], [191.34, 177.56], [213.8, 177.35], [225.0, 177.36]]}]\n```" - } - ] - }, - { - "id": 2691, - "data_source": "Mono3DRefer", - "scene_name": "003222", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[303.49, 205.03], [272.87, 204.99], [308.81, 201.77], [336.39, 201.8], [303.49, 169.78], [272.87, 169.78], [308.81, 170.09], [336.39, 170.08]]}, {\"category\": \"cyclist\", \"corners_3d\": [[203.01, 213.9], [191.34, 213.8], [213.8, 212.01], [225.0, 212.09], [203.01, 177.57], [191.34, 177.56], [213.8, 177.35], [225.0, 177.36]]}]\n```" - } - ] - }, - { - "id": 2692, - "data_source": "Mono3DRefer", - "scene_name": "003222", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[162.77, 232.19], [113.07, 232.02], [177.02, 224.81], [220.8, 224.94], [162.77, 189.59], [113.07, 189.54], [177.02, 187.51], [220.8, 187.55]]}, {\"category\": \"van\", \"corners_3d\": [[303.49, 205.03], [272.87, 204.99], [308.81, 201.77], [336.39, 201.8], [303.49, 169.78], [272.87, 169.78], [308.81, 170.09], [336.39, 170.08]]}]\n```" - } - ] - }, - { - "id": 2693, - "data_source": "Mono3DRefer", - "scene_name": "003222", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[162.77, 232.19], [113.07, 232.02], [177.02, 224.81], [220.8, 224.94], [162.77, 189.59], [113.07, 189.54], [177.02, 187.51], [220.8, 187.55]]}, {\"category\": \"van\", \"corners_3d\": [[303.49, 205.03], [272.87, 204.99], [308.81, 201.77], [336.39, 201.8], [303.49, 169.78], [272.87, 169.78], [308.81, 170.09], [336.39, 170.08]]}, {\"category\": \"cyclist\", \"corners_3d\": [[203.01, 213.9], [191.34, 213.8], [213.8, 212.01], [225.0, 212.09], [203.01, 177.57], [191.34, 177.56], [213.8, 177.35], [225.0, 177.36]]}]\n```" - } - ] - }, - { - "id": 2694, - "data_source": "Mono3DRefer", - "scene_name": "003223", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[545.6, 203.94], [509.88, 203.7], [534.0, 200.13], [565.72, 200.31], [545.6, 175.54], [509.88, 175.52], [534.0, 175.21], [565.72, 175.23]]}]\n```" - } - ] - }, - { - "id": 2695, - "data_source": "Mono3DRefer", - "scene_name": "003223", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[512.37, 214.52], [455.6, 214.49], [489.19, 205.78], [534.1, 205.8], [512.37, 144.68], [455.6, 144.71], [489.19, 150.59], [534.1, 150.58]]}]\n```" - } - ] - }, - { - "id": 2696, - "data_source": "Mono3DRefer", - "scene_name": "003223", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[545.6, 203.94], [509.88, 203.7], [534.0, 200.13], [565.72, 200.31], [545.6, 175.54], [509.88, 175.52], [534.0, 175.21], [565.72, 175.23]]}, {\"category\": \"van\", \"corners_3d\": [[512.37, 214.52], [455.6, 214.49], [489.19, 205.78], [534.1, 205.8], [512.37, 144.68], [455.6, 144.71], [489.19, 150.59], [534.1, 150.58]]}]\n```" - } - ] - }, - { - "id": 2697, - "data_source": "Mono3DRefer", - "scene_name": "003248", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[491.69, 228.92], [444.37, 229.43], [451.32, 223.18], [493.4, 222.78], [491.69, 185.28], [444.37, 185.39], [451.32, 184.01], [493.4, 183.92]]}]\n```" - } - ] - }, - { - "id": 2698, - "data_source": "Mono3DRefer", - "scene_name": "003269", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[838.48, 305.32], [820.99, 298.8], [898.88, 296.86], [920.29, 303.16], [838.48, 151.51], [820.99, 153.03], [898.88, 153.48], [920.29, 152.02]]}]\n```" - } - ] - }, - { - "id": 2699, - "data_source": "Mono3DRefer", - "scene_name": "003276", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[176.01, 244.43], [119.29, 244.26], [194.52, 233.76], [243.03, 233.88], [176.01, 198.99], [119.29, 198.93], [194.52, 195.09], [243.03, 195.14]]}]\n```" - } - ] - }, - { - "id": 2700, - "data_source": "Mono3DRefer", - "scene_name": "003276", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[599.22, 190.78], [623.28, 190.78], [625.72, 194.1], [597.2, 194.1], [599.22, 164.71], [623.28, 164.71], [625.72, 163.2], [597.2, 163.2]]}]\n```" - } - ] - }, - { - "id": 2701, - "data_source": "Mono3DRefer", - "scene_name": "003276", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[599.22, 190.78], [623.28, 190.78], [625.72, 194.1], [597.2, 194.1], [599.22, 164.71], [623.28, 164.71], [625.72, 163.2], [597.2, 163.2]]}, {\"category\": \"car\", \"corners_3d\": [[176.01, 244.43], [119.29, 244.26], [194.52, 233.76], [243.03, 233.88], [176.01, 198.99], [119.29, 198.93], [194.52, 195.09], [243.03, 195.14]]}]\n```" - } - ] - }, - { - "id": 2702, - "data_source": "Mono3DRefer", - "scene_name": "003279", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[392.3, 270.99], [412.72, 258.33], [647.87, 256.1], [661.85, 268.07], [392.3, 181.92], [412.72, 180.75], [647.87, 180.55], [661.85, 181.65]]}, {\"category\": \"car\", \"corners_3d\": [[536.96, 205.97], [560.2, 205.84], [564.5, 209.03], [539.01, 209.2], [536.96, 184.18], [560.2, 184.13], [564.5, 185.22], [539.01, 185.28]]}, {\"category\": \"car\", \"corners_3d\": [[469.13, 212.93], [441.16, 213.07], [450.77, 209.3], [476.1, 209.19], [469.13, 189.41], [441.16, 189.47], [450.77, 187.91], [476.1, 187.87]]}]\n```" - } - ] - }, - { - "id": 2703, - "data_source": "Mono3DRefer", - "scene_name": "003288", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[698.34, 254.28], [779.88, 254.1], [854.43, 286.99], [740.02, 287.34], [698.34, 169.96], [779.88, 169.97], [854.43, 168.8], [740.02, 168.79]]}, {\"category\": \"car\", \"corners_3d\": [[-252.56, 348.03], [-86.81, 315.48], [212.45, 318.05], [112.83, 351.91], [-252.56, 189.5], [-86.81, 186.41], [212.45, 186.65], [112.83, 189.87]]}, {\"category\": \"car\", \"corners_3d\": [[334.18, 285.98], [287.02, 307.0], [-75.16, 309.0], [30.07, 287.39], [334.18, 185.4], [287.02, 187.73], [-75.16, 187.96], [30.07, 185.56]]}, {\"category\": \"car\", \"corners_3d\": [[98.52, 279.78], [164.57, 266.9], [369.93, 268.33], [331.35, 281.63], [98.52, 188.2], [164.57, 186.35], [369.93, 186.56], [331.35, 188.46]]}, {\"category\": \"car\", \"corners_3d\": [[190.7, 259.34], [234.25, 250.67], [392.04, 251.04], [365.94, 259.8], [190.7, 183.01], [234.25, 181.99], [392.04, 182.04], [365.94, 183.07]]}, {\"category\": \"car\", \"corners_3d\": [[463.71, 236.36], [450.2, 241.97], [273.57, 241.8], [301.34, 236.22], [463.71, 166.77], [450.2, 166.24], [273.57, 166.25], [301.34, 166.79]]}, {\"category\": \"car\", \"corners_3d\": [[258.55, 250.75], [290.15, 243.9], [422.63, 244.05], [403.76, 250.93], [258.55, 177.68], [290.15, 177.26], [422.63, 177.27], [403.76, 177.69]]}, {\"category\": \"car\", \"corners_3d\": [[678.19, 235.12], [737.56, 234.82], [770.58, 246.8], [699.85, 247.24], [678.19, 182.98], [737.56, 182.93], [770.58, 184.88], [699.85, 184.95]]}, {\"category\": \"car\", \"corners_3d\": [[318.6, 232.18], [339.32, 227.94], [452.43, 227.93], [440.43, 232.16], [318.6, 167.67], [339.32, 168.04], [452.43, 168.04], [440.43, 167.67]]}, {\"category\": \"car\", \"corners_3d\": [[358.28, 223.49], [374.76, 220.45], [487.97, 220.68], [478.69, 223.76], [358.28, 175.36], [374.76, 175.21], [487.97, 175.22], [478.69, 175.38]]}, {\"category\": \"car\", \"corners_3d\": [[641.87, 207.03], [672.32, 206.99], [681.16, 210.59], [647.51, 210.64], [641.87, 179.18], [672.32, 179.18], [681.16, 179.84], [647.51, 179.85]]}, {\"category\": \"car\", \"corners_3d\": [[664.06, 221.82], [711.08, 221.82], [728.21, 230.14], [673.22, 230.13], [664.06, 178.47], [711.08, 178.47], [728.21, 179.43], [673.22, 179.43]]}]\n```" - } - ] - }, - { - "id": 2704, - "data_source": "Mono3DRefer", - "scene_name": "003289", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.22, 224.89], [662.87, 224.9], [673.11, 235.0], [603.07, 234.99], [604.22, 177.7], [662.87, 177.7], [673.11, 178.64], [603.07, 178.64]]}, {\"category\": \"car\", \"corners_3d\": [[1076.01, 210.37], [1044.95, 208.06], [1155.55, 207.77], [1193.55, 210.05], [1076.01, 169.15], [1044.95, 169.38], [1155.55, 169.4], [1193.55, 169.18]]}, {\"category\": \"car\", \"corners_3d\": [[407.53, 208.65], [374.87, 208.64], [397.17, 205.33], [426.82, 205.34], [407.53, 178.37], [374.87, 178.37], [397.17, 177.86], [426.82, 177.86]]}, {\"category\": \"car\", \"corners_3d\": [[586.41, 210.83], [624.61, 210.82], [627.42, 215.72], [584.27, 215.74], [586.41, 176.05], [624.61, 176.05], [627.42, 176.46], [584.27, 176.46]]}]\n```" - } - ] - }, - { - "id": 2705, - "data_source": "Mono3DRefer", - "scene_name": "003289", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[652.6, 201.04], [685.26, 201.08], [691.68, 204.47], [655.1, 204.42], [652.6, 163.92], [685.26, 163.9], [691.68, 162.83], [655.1, 162.85]]}, {\"category\": \"van\", \"corners_3d\": [[483.72, 195.47], [464.7, 195.48], [473.61, 194.01], [491.38, 194.0], [483.72, 174.67], [464.7, 174.67], [473.61, 174.56], [491.38, 174.56]]}]\n```" - } - ] - }, - { - "id": 2706, - "data_source": "Mono3DRefer", - "scene_name": "003289", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[758.63, 206.29], [755.74, 205.36], [803.11, 205.49], [807.36, 206.43], [758.63, 166.16], [755.74, 166.34], [803.11, 166.32], [807.36, 166.13]]}]\n```" - } - ] - }, - { - "id": 2707, - "data_source": "Mono3DRefer", - "scene_name": "003289", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[989.27, 232.63], [964.83, 232.67], [946.04, 229.63], [969.25, 229.6], [989.27, 155.94], [964.83, 155.93], [946.04, 156.79], [969.25, 156.8]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[546.0, 208.92], [546.61, 209.63], [531.95, 209.73], [531.63, 209.01], [546.0, 167.79], [546.61, 167.69], [531.95, 167.67], [531.63, 167.77]]}]\n```" - } - ] - }, - { - "id": 2708, - "data_source": "Mono3DRefer", - "scene_name": "003289", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.22, 224.89], [662.87, 224.9], [673.11, 235.0], [603.07, 234.99], [604.22, 177.7], [662.87, 177.7], [673.11, 178.64], [603.07, 178.64]]}, {\"category\": \"car\", \"corners_3d\": [[1076.01, 210.37], [1044.95, 208.06], [1155.55, 207.77], [1193.55, 210.05], [1076.01, 169.15], [1044.95, 169.38], [1155.55, 169.4], [1193.55, 169.18]]}, {\"category\": \"car\", \"corners_3d\": [[407.53, 208.65], [374.87, 208.64], [397.17, 205.33], [426.82, 205.34], [407.53, 178.37], [374.87, 178.37], [397.17, 177.86], [426.82, 177.86]]}, {\"category\": \"van\", \"corners_3d\": [[652.6, 201.04], [685.26, 201.08], [691.68, 204.47], [655.1, 204.42], [652.6, 163.92], [685.26, 163.9], [691.68, 162.83], [655.1, 162.85]]}, {\"category\": \"van\", \"corners_3d\": [[483.72, 195.47], [464.7, 195.48], [473.61, 194.01], [491.38, 194.0], [483.72, 174.67], [464.7, 174.67], [473.61, 174.56], [491.38, 174.56]]}, {\"category\": \"car\", \"corners_3d\": [[586.41, 210.83], [624.61, 210.82], [627.42, 215.72], [584.27, 215.74], [586.41, 176.05], [624.61, 176.05], [627.42, 176.46], [584.27, 176.46]]}]\n```" - } - ] - }, - { - "id": 2709, - "data_source": "Mono3DRefer", - "scene_name": "003289", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.22, 224.89], [662.87, 224.9], [673.11, 235.0], [603.07, 234.99], [604.22, 177.7], [662.87, 177.7], [673.11, 178.64], [603.07, 178.64]]}, {\"category\": \"car\", \"corners_3d\": [[1076.01, 210.37], [1044.95, 208.06], [1155.55, 207.77], [1193.55, 210.05], [1076.01, 169.15], [1044.95, 169.38], [1155.55, 169.4], [1193.55, 169.18]]}, {\"category\": \"car\", \"corners_3d\": [[407.53, 208.65], [374.87, 208.64], [397.17, 205.33], [426.82, 205.34], [407.53, 178.37], [374.87, 178.37], [397.17, 177.86], [426.82, 177.86]]}, {\"category\": \"cyclist\", \"corners_3d\": [[758.63, 206.29], [755.74, 205.36], [803.11, 205.49], [807.36, 206.43], [758.63, 166.16], [755.74, 166.34], [803.11, 166.32], [807.36, 166.13]]}, {\"category\": \"car\", \"corners_3d\": [[586.41, 210.83], [624.61, 210.82], [627.42, 215.72], [584.27, 215.74], [586.41, 176.05], [624.61, 176.05], [627.42, 176.46], [584.27, 176.46]]}]\n```" - } - ] - }, - { - "id": 2710, - "data_source": "Mono3DRefer", - "scene_name": "003289", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.22, 224.89], [662.87, 224.9], [673.11, 235.0], [603.07, 234.99], [604.22, 177.7], [662.87, 177.7], [673.11, 178.64], [603.07, 178.64]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[989.27, 232.63], [964.83, 232.67], [946.04, 229.63], [969.25, 229.6], [989.27, 155.94], [964.83, 155.93], [946.04, 156.79], [969.25, 156.8]]}, {\"category\": \"car\", \"corners_3d\": [[1076.01, 210.37], [1044.95, 208.06], [1155.55, 207.77], [1193.55, 210.05], [1076.01, 169.15], [1044.95, 169.38], [1155.55, 169.4], [1193.55, 169.18]]}, {\"category\": \"car\", \"corners_3d\": [[407.53, 208.65], [374.87, 208.64], [397.17, 205.33], [426.82, 205.34], [407.53, 178.37], [374.87, 178.37], [397.17, 177.86], [426.82, 177.86]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[546.0, 208.92], [546.61, 209.63], [531.95, 209.73], [531.63, 209.01], [546.0, 167.79], [546.61, 167.69], [531.95, 167.67], [531.63, 167.77]]}, {\"category\": \"car\", \"corners_3d\": [[586.41, 210.83], [624.61, 210.82], [627.42, 215.72], [584.27, 215.74], [586.41, 176.05], [624.61, 176.05], [627.42, 176.46], [584.27, 176.46]]}]\n```" - } - ] - }, - { - "id": 2711, - "data_source": "Mono3DRefer", - "scene_name": "003289", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[758.63, 206.29], [755.74, 205.36], [803.11, 205.49], [807.36, 206.43], [758.63, 166.16], [755.74, 166.34], [803.11, 166.32], [807.36, 166.13]]}, {\"category\": \"van\", \"corners_3d\": [[652.6, 201.04], [685.26, 201.08], [691.68, 204.47], [655.1, 204.42], [652.6, 163.92], [685.26, 163.9], [691.68, 162.83], [655.1, 162.85]]}, {\"category\": \"van\", \"corners_3d\": [[483.72, 195.47], [464.7, 195.48], [473.61, 194.01], [491.38, 194.0], [483.72, 174.67], [464.7, 174.67], [473.61, 174.56], [491.38, 174.56]]}]\n```" - } - ] - }, - { - "id": 2712, - "data_source": "Mono3DRefer", - "scene_name": "003289", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[989.27, 232.63], [964.83, 232.67], [946.04, 229.63], [969.25, 229.6], [989.27, 155.94], [964.83, 155.93], [946.04, 156.79], [969.25, 156.8]]}, {\"category\": \"van\", \"corners_3d\": [[652.6, 201.04], [685.26, 201.08], [691.68, 204.47], [655.1, 204.42], [652.6, 163.92], [685.26, 163.9], [691.68, 162.83], [655.1, 162.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[546.0, 208.92], [546.61, 209.63], [531.95, 209.73], [531.63, 209.01], [546.0, 167.79], [546.61, 167.69], [531.95, 167.67], [531.63, 167.77]]}, {\"category\": \"van\", \"corners_3d\": [[483.72, 195.47], [464.7, 195.48], [473.61, 194.01], [491.38, 194.0], [483.72, 174.67], [464.7, 174.67], [473.61, 174.56], [491.38, 174.56]]}]\n```" - } - ] - }, - { - "id": 2713, - "data_source": "Mono3DRefer", - "scene_name": "003289", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[989.27, 232.63], [964.83, 232.67], [946.04, 229.63], [969.25, 229.6], [989.27, 155.94], [964.83, 155.93], [946.04, 156.79], [969.25, 156.8]]}, {\"category\": \"cyclist\", \"corners_3d\": [[758.63, 206.29], [755.74, 205.36], [803.11, 205.49], [807.36, 206.43], [758.63, 166.16], [755.74, 166.34], [803.11, 166.32], [807.36, 166.13]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[546.0, 208.92], [546.61, 209.63], [531.95, 209.73], [531.63, 209.01], [546.0, 167.79], [546.61, 167.69], [531.95, 167.67], [531.63, 167.77]]}]\n```" - } - ] - }, - { - "id": 2714, - "data_source": "Mono3DRefer", - "scene_name": "003289", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.22, 224.89], [662.87, 224.9], [673.11, 235.0], [603.07, 234.99], [604.22, 177.7], [662.87, 177.7], [673.11, 178.64], [603.07, 178.64]]}, {\"category\": \"car\", \"corners_3d\": [[1076.01, 210.37], [1044.95, 208.06], [1155.55, 207.77], [1193.55, 210.05], [1076.01, 169.15], [1044.95, 169.38], [1155.55, 169.4], [1193.55, 169.18]]}, {\"category\": \"car\", \"corners_3d\": [[407.53, 208.65], [374.87, 208.64], [397.17, 205.33], [426.82, 205.34], [407.53, 178.37], [374.87, 178.37], [397.17, 177.86], [426.82, 177.86]]}, {\"category\": \"cyclist\", \"corners_3d\": [[758.63, 206.29], [755.74, 205.36], [803.11, 205.49], [807.36, 206.43], [758.63, 166.16], [755.74, 166.34], [803.11, 166.32], [807.36, 166.13]]}, {\"category\": \"van\", \"corners_3d\": [[652.6, 201.04], [685.26, 201.08], [691.68, 204.47], [655.1, 204.42], [652.6, 163.92], [685.26, 163.9], [691.68, 162.83], [655.1, 162.85]]}, {\"category\": \"van\", \"corners_3d\": [[483.72, 195.47], [464.7, 195.48], [473.61, 194.01], [491.38, 194.0], [483.72, 174.67], [464.7, 174.67], [473.61, 174.56], [491.38, 174.56]]}, {\"category\": \"car\", \"corners_3d\": [[586.41, 210.83], [624.61, 210.82], [627.42, 215.72], [584.27, 215.74], [586.41, 176.05], [624.61, 176.05], [627.42, 176.46], [584.27, 176.46]]}]\n```" - } - ] - }, - { - "id": 2715, - "data_source": "Mono3DRefer", - "scene_name": "003289", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.22, 224.89], [662.87, 224.9], [673.11, 235.0], [603.07, 234.99], [604.22, 177.7], [662.87, 177.7], [673.11, 178.64], [603.07, 178.64]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[989.27, 232.63], [964.83, 232.67], [946.04, 229.63], [969.25, 229.6], [989.27, 155.94], [964.83, 155.93], [946.04, 156.79], [969.25, 156.8]]}, {\"category\": \"car\", \"corners_3d\": [[1076.01, 210.37], [1044.95, 208.06], [1155.55, 207.77], [1193.55, 210.05], [1076.01, 169.15], [1044.95, 169.38], [1155.55, 169.4], [1193.55, 169.18]]}, {\"category\": \"car\", \"corners_3d\": [[407.53, 208.65], [374.87, 208.64], [397.17, 205.33], [426.82, 205.34], [407.53, 178.37], [374.87, 178.37], [397.17, 177.86], [426.82, 177.86]]}, {\"category\": \"van\", \"corners_3d\": [[652.6, 201.04], [685.26, 201.08], [691.68, 204.47], [655.1, 204.42], [652.6, 163.92], [685.26, 163.9], [691.68, 162.83], [655.1, 162.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[546.0, 208.92], [546.61, 209.63], [531.95, 209.73], [531.63, 209.01], [546.0, 167.79], [546.61, 167.69], [531.95, 167.67], [531.63, 167.77]]}, {\"category\": \"van\", \"corners_3d\": [[483.72, 195.47], [464.7, 195.48], [473.61, 194.01], [491.38, 194.0], [483.72, 174.67], [464.7, 174.67], [473.61, 174.56], [491.38, 174.56]]}, {\"category\": \"car\", \"corners_3d\": [[586.41, 210.83], [624.61, 210.82], [627.42, 215.72], [584.27, 215.74], [586.41, 176.05], [624.61, 176.05], [627.42, 176.46], [584.27, 176.46]]}]\n```" - } - ] - }, - { - "id": 2716, - "data_source": "Mono3DRefer", - "scene_name": "003289", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.22, 224.89], [662.87, 224.9], [673.11, 235.0], [603.07, 234.99], [604.22, 177.7], [662.87, 177.7], [673.11, 178.64], [603.07, 178.64]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[989.27, 232.63], [964.83, 232.67], [946.04, 229.63], [969.25, 229.6], [989.27, 155.94], [964.83, 155.93], [946.04, 156.79], [969.25, 156.8]]}, {\"category\": \"car\", \"corners_3d\": [[1076.01, 210.37], [1044.95, 208.06], [1155.55, 207.77], [1193.55, 210.05], [1076.01, 169.15], [1044.95, 169.38], [1155.55, 169.4], [1193.55, 169.18]]}, {\"category\": \"car\", \"corners_3d\": [[407.53, 208.65], [374.87, 208.64], [397.17, 205.33], [426.82, 205.34], [407.53, 178.37], [374.87, 178.37], [397.17, 177.86], [426.82, 177.86]]}, {\"category\": \"cyclist\", \"corners_3d\": [[758.63, 206.29], [755.74, 205.36], [803.11, 205.49], [807.36, 206.43], [758.63, 166.16], [755.74, 166.34], [803.11, 166.32], [807.36, 166.13]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[546.0, 208.92], [546.61, 209.63], [531.95, 209.73], [531.63, 209.01], [546.0, 167.79], [546.61, 167.69], [531.95, 167.67], [531.63, 167.77]]}, {\"category\": \"car\", \"corners_3d\": [[586.41, 210.83], [624.61, 210.82], [627.42, 215.72], [584.27, 215.74], [586.41, 176.05], [624.61, 176.05], [627.42, 176.46], [584.27, 176.46]]}]\n```" - } - ] - }, - { - "id": 2717, - "data_source": "Mono3DRefer", - "scene_name": "003289", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[989.27, 232.63], [964.83, 232.67], [946.04, 229.63], [969.25, 229.6], [989.27, 155.94], [964.83, 155.93], [946.04, 156.79], [969.25, 156.8]]}, {\"category\": \"cyclist\", \"corners_3d\": [[758.63, 206.29], [755.74, 205.36], [803.11, 205.49], [807.36, 206.43], [758.63, 166.16], [755.74, 166.34], [803.11, 166.32], [807.36, 166.13]]}, {\"category\": \"van\", \"corners_3d\": [[652.6, 201.04], [685.26, 201.08], [691.68, 204.47], [655.1, 204.42], [652.6, 163.92], [685.26, 163.9], [691.68, 162.83], [655.1, 162.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[546.0, 208.92], [546.61, 209.63], [531.95, 209.73], [531.63, 209.01], [546.0, 167.79], [546.61, 167.69], [531.95, 167.67], [531.63, 167.77]]}, {\"category\": \"van\", \"corners_3d\": [[483.72, 195.47], [464.7, 195.48], [473.61, 194.01], [491.38, 194.0], [483.72, 174.67], [464.7, 174.67], [473.61, 174.56], [491.38, 174.56]]}]\n```" - } - ] - }, - { - "id": 2718, - "data_source": "Mono3DRefer", - "scene_name": "003289", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.22, 224.89], [662.87, 224.9], [673.11, 235.0], [603.07, 234.99], [604.22, 177.7], [662.87, 177.7], [673.11, 178.64], [603.07, 178.64]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[989.27, 232.63], [964.83, 232.67], [946.04, 229.63], [969.25, 229.6], [989.27, 155.94], [964.83, 155.93], [946.04, 156.79], [969.25, 156.8]]}, {\"category\": \"car\", \"corners_3d\": [[1076.01, 210.37], [1044.95, 208.06], [1155.55, 207.77], [1193.55, 210.05], [1076.01, 169.15], [1044.95, 169.38], [1155.55, 169.4], [1193.55, 169.18]]}, {\"category\": \"car\", \"corners_3d\": [[407.53, 208.65], [374.87, 208.64], [397.17, 205.33], [426.82, 205.34], [407.53, 178.37], [374.87, 178.37], [397.17, 177.86], [426.82, 177.86]]}, {\"category\": \"cyclist\", \"corners_3d\": [[758.63, 206.29], [755.74, 205.36], [803.11, 205.49], [807.36, 206.43], [758.63, 166.16], [755.74, 166.34], [803.11, 166.32], [807.36, 166.13]]}, {\"category\": \"van\", \"corners_3d\": [[652.6, 201.04], [685.26, 201.08], [691.68, 204.47], [655.1, 204.42], [652.6, 163.92], [685.26, 163.9], [691.68, 162.83], [655.1, 162.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[546.0, 208.92], [546.61, 209.63], [531.95, 209.73], [531.63, 209.01], [546.0, 167.79], [546.61, 167.69], [531.95, 167.67], [531.63, 167.77]]}, {\"category\": \"van\", \"corners_3d\": [[483.72, 195.47], [464.7, 195.48], [473.61, 194.01], [491.38, 194.0], [483.72, 174.67], [464.7, 174.67], [473.61, 174.56], [491.38, 174.56]]}, {\"category\": \"car\", \"corners_3d\": [[586.41, 210.83], [624.61, 210.82], [627.42, 215.72], [584.27, 215.74], [586.41, 176.05], [624.61, 176.05], [627.42, 176.46], [584.27, 176.46]]}]\n```" - } - ] - }, - { - "id": 2719, - "data_source": "Mono3DRefer", - "scene_name": "003293", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[694.88, 199.62], [718.59, 199.78], [716.81, 201.88], [691.26, 201.69], [694.88, 177.45], [718.59, 177.48], [716.81, 177.84], [691.26, 177.81]]}, {\"category\": \"car\", \"corners_3d\": [[801.55, 306.87], [934.85, 306.89], [1174.54, 405.82], [942.82, 405.76], [801.55, 180.8], [934.85, 180.81], [1174.54, 186.67], [942.82, 186.67]]}, {\"category\": \"car\", \"corners_3d\": [[577.86, 238.69], [519.55, 238.0], [548.11, 227.88], [597.53, 228.37], [577.86, 183.94], [519.55, 183.82], [548.11, 182.12], [597.53, 182.2]]}, {\"category\": \"car\", \"corners_3d\": [[765.25, 263.54], [845.25, 264.34], [913.58, 299.6], [802.3, 298.07], [765.25, 156.68], [845.25, 156.53], [913.58, 150.24], [802.3, 150.52]]}]\n```" - } - ] - }, - { - "id": 2720, - "data_source": "Mono3DRefer", - "scene_name": "003297", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[784.0, 299.86], [656.16, 297.24], [667.8, 260.88], [757.89, 262.18], [784.0, 183.91], [656.16, 183.68], [667.8, 180.51], [757.89, 180.63]]}]\n```" - } - ] - }, - { - "id": 2721, - "data_source": "Mono3DRefer", - "scene_name": "003315", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[575.05, 268.07], [615.18, 268.02], [617.28, 278.45], [572.37, 278.51], [575.05, 153.85], [615.18, 153.86], [617.28, 150.69], [572.37, 150.67]]}]\n```" - } - ] - }, - { - "id": 2722, - "data_source": "Mono3DRefer", - "scene_name": "003320", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1401.87, 351.95], [1186.07, 321.68], [1293.84, 289.44], [1472.2, 307.22], [1401.87, 178.74], [1186.07, 177.75], [1293.84, 176.69], [1472.2, 177.27]]}, {\"category\": \"car\", \"corners_3d\": [[942.23, 279.6], [887.78, 263.89], [1114.57, 262.38], [1206.34, 277.53], [942.23, 151.92], [887.78, 155.0], [1114.57, 155.3], [1206.34, 152.33]]}, {\"category\": \"car\", \"corners_3d\": [[-40.87, 527.47], [-410.29, 545.83], [118.54, 335.91], [272.94, 332.3], [-40.87, 204.31], [-410.29, 205.94], [118.54, 187.32], [272.94, 187.0]]}, {\"category\": \"car\", \"corners_3d\": [[315.74, 313.54], [183.02, 315.08], [305.04, 269.73], [394.84, 269.01], [315.74, 189.84], [183.02, 190.02], [305.04, 184.55], [394.84, 184.46]]}]\n```" - } - ] - }, - { - "id": 2723, - "data_source": "Mono3DRefer", - "scene_name": "003320", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[862.5, 260.29], [825.29, 248.79], [1083.65, 247.65], [1158.63, 258.77], [862.5, 150.38], [825.29, 153.33], [1083.65, 153.63], [1158.63, 150.77]]}]\n```" - } - ] - }, - { - "id": 2724, - "data_source": "Mono3DRefer", - "scene_name": "003320", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[646.82, 208.58], [704.32, 209.85], [607.02, 225.05], [531.7, 222.55], [646.82, 135.88], [704.32, 134.56], [607.02, 118.82], [531.7, 121.42]]}]\n```" - } - ] - }, - { - "id": 2725, - "data_source": "Mono3DRefer", - "scene_name": "003320", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1401.87, 351.95], [1186.07, 321.68], [1293.84, 289.44], [1472.2, 307.22], [1401.87, 178.74], [1186.07, 177.75], [1293.84, 176.69], [1472.2, 177.27]]}, {\"category\": \"car\", \"corners_3d\": [[942.23, 279.6], [887.78, 263.89], [1114.57, 262.38], [1206.34, 277.53], [942.23, 151.92], [887.78, 155.0], [1114.57, 155.3], [1206.34, 152.33]]}, {\"category\": \"van\", \"corners_3d\": [[862.5, 260.29], [825.29, 248.79], [1083.65, 247.65], [1158.63, 258.77], [862.5, 150.38], [825.29, 153.33], [1083.65, 153.63], [1158.63, 150.77]]}, {\"category\": \"car\", \"corners_3d\": [[-40.87, 527.47], [-410.29, 545.83], [118.54, 335.91], [272.94, 332.3], [-40.87, 204.31], [-410.29, 205.94], [118.54, 187.32], [272.94, 187.0]]}, {\"category\": \"car\", \"corners_3d\": [[315.74, 313.54], [183.02, 315.08], [305.04, 269.73], [394.84, 269.01], [315.74, 189.84], [183.02, 190.02], [305.04, 184.55], [394.84, 184.46]]}]\n```" - } - ] - }, - { - "id": 2726, - "data_source": "Mono3DRefer", - "scene_name": "003320", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1401.87, 351.95], [1186.07, 321.68], [1293.84, 289.44], [1472.2, 307.22], [1401.87, 178.74], [1186.07, 177.75], [1293.84, 176.69], [1472.2, 177.27]]}, {\"category\": \"car\", \"corners_3d\": [[942.23, 279.6], [887.78, 263.89], [1114.57, 262.38], [1206.34, 277.53], [942.23, 151.92], [887.78, 155.0], [1114.57, 155.3], [1206.34, 152.33]]}, {\"category\": \"car\", \"corners_3d\": [[-40.87, 527.47], [-410.29, 545.83], [118.54, 335.91], [272.94, 332.3], [-40.87, 204.31], [-410.29, 205.94], [118.54, 187.32], [272.94, 187.0]]}, {\"category\": \"car\", \"corners_3d\": [[315.74, 313.54], [183.02, 315.08], [305.04, 269.73], [394.84, 269.01], [315.74, 189.84], [183.02, 190.02], [305.04, 184.55], [394.84, 184.46]]}, {\"category\": \"truck\", \"corners_3d\": [[646.82, 208.58], [704.32, 209.85], [607.02, 225.05], [531.7, 222.55], [646.82, 135.88], [704.32, 134.56], [607.02, 118.82], [531.7, 121.42]]}]\n```" - } - ] - }, - { - "id": 2727, - "data_source": "Mono3DRefer", - "scene_name": "003320", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[862.5, 260.29], [825.29, 248.79], [1083.65, 247.65], [1158.63, 258.77], [862.5, 150.38], [825.29, 153.33], [1083.65, 153.63], [1158.63, 150.77]]}, {\"category\": \"truck\", \"corners_3d\": [[646.82, 208.58], [704.32, 209.85], [607.02, 225.05], [531.7, 222.55], [646.82, 135.88], [704.32, 134.56], [607.02, 118.82], [531.7, 121.42]]}]\n```" - } - ] - }, - { - "id": 2728, - "data_source": "Mono3DRefer", - "scene_name": "003320", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1401.87, 351.95], [1186.07, 321.68], [1293.84, 289.44], [1472.2, 307.22], [1401.87, 178.74], [1186.07, 177.75], [1293.84, 176.69], [1472.2, 177.27]]}, {\"category\": \"car\", \"corners_3d\": [[942.23, 279.6], [887.78, 263.89], [1114.57, 262.38], [1206.34, 277.53], [942.23, 151.92], [887.78, 155.0], [1114.57, 155.3], [1206.34, 152.33]]}, {\"category\": \"van\", \"corners_3d\": [[862.5, 260.29], [825.29, 248.79], [1083.65, 247.65], [1158.63, 258.77], [862.5, 150.38], [825.29, 153.33], [1083.65, 153.63], [1158.63, 150.77]]}, {\"category\": \"car\", \"corners_3d\": [[-40.87, 527.47], [-410.29, 545.83], [118.54, 335.91], [272.94, 332.3], [-40.87, 204.31], [-410.29, 205.94], [118.54, 187.32], [272.94, 187.0]]}, {\"category\": \"car\", \"corners_3d\": [[315.74, 313.54], [183.02, 315.08], [305.04, 269.73], [394.84, 269.01], [315.74, 189.84], [183.02, 190.02], [305.04, 184.55], [394.84, 184.46]]}, {\"category\": \"truck\", \"corners_3d\": [[646.82, 208.58], [704.32, 209.85], [607.02, 225.05], [531.7, 222.55], [646.82, 135.88], [704.32, 134.56], [607.02, 118.82], [531.7, 121.42]]}]\n```" - } - ] - }, - { - "id": 2729, - "data_source": "Mono3DRefer", - "scene_name": "003331", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1044.3, 313.59], [1011.78, 304.65], [1104.35, 303.61], [1143.17, 312.4], [1044.3, 154.82], [1011.78, 156.55], [1104.35, 156.75], [1143.17, 155.05]]}]\n```" - } - ] - }, - { - "id": 2730, - "data_source": "Mono3DRefer", - "scene_name": "003332", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[577.22, 201.64], [547.53, 201.64], [553.21, 198.98], [580.15, 198.98], [577.22, 176.25], [547.53, 176.25], [553.21, 175.93], [580.15, 175.93]]}, {\"category\": \"car\", \"corners_3d\": [[475.42, 282.51], [370.22, 282.2], [434.91, 254.26], [513.31, 254.43], [475.42, 181.66], [370.22, 181.63], [434.91, 179.39], [513.31, 179.4]]}]\n```" - } - ] - }, - { - "id": 2731, - "data_source": "Mono3DRefer", - "scene_name": "003346", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[760.22, 301.35], [886.26, 300.92], [1121.04, 404.4], [893.98, 405.82], [760.22, 168.24], [886.26, 168.25], [1121.04, 164.53], [893.98, 164.48]]}, {\"category\": \"van\", \"corners_3d\": [[681.05, 221.84], [733.35, 221.74], [767.41, 233.23], [702.88, 233.39], [681.05, 162.09], [733.35, 162.11], [767.41, 159.58], [702.88, 159.55]]}]\n```" - } - ] - }, - { - "id": 2732, - "data_source": "Mono3DRefer", - "scene_name": "003346", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[663.64, 213.51], [698.64, 213.45], [715.52, 219.67], [675.17, 219.74], [663.64, 176.75], [698.64, 176.75], [715.52, 177.34], [675.17, 177.35]]}, {\"category\": \"car\", \"corners_3d\": [[547.38, 234.34], [485.51, 234.34], [508.24, 223.02], [558.72, 223.02], [547.38, 177.26], [485.51, 177.26], [508.24, 176.45], [558.72, 176.45]]}]\n```" - } - ] - }, - { - "id": 2733, - "data_source": "Mono3DRefer", - "scene_name": "003346", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[663.64, 213.51], [698.64, 213.45], [715.52, 219.67], [675.17, 219.74], [663.64, 176.75], [698.64, 176.75], [715.52, 177.34], [675.17, 177.35]]}, {\"category\": \"van\", \"corners_3d\": [[760.22, 301.35], [886.26, 300.92], [1121.04, 404.4], [893.98, 405.82], [760.22, 168.24], [886.26, 168.25], [1121.04, 164.53], [893.98, 164.48]]}, {\"category\": \"car\", \"corners_3d\": [[547.38, 234.34], [485.51, 234.34], [508.24, 223.02], [558.72, 223.02], [547.38, 177.26], [485.51, 177.26], [508.24, 176.45], [558.72, 176.45]]}, {\"category\": \"van\", \"corners_3d\": [[681.05, 221.84], [733.35, 221.74], [767.41, 233.23], [702.88, 233.39], [681.05, 162.09], [733.35, 162.11], [767.41, 159.58], [702.88, 159.55]]}]\n```" - } - ] - }, - { - "id": 2734, - "data_source": "Mono3DRefer", - "scene_name": "003351", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[422.53, 199.23], [398.76, 199.2], [415.37, 197.31], [437.46, 197.33], [422.53, 179.63], [398.76, 179.63], [415.37, 179.14], [437.46, 179.15]]}, {\"category\": \"car\", \"corners_3d\": [[234.22, 236.1], [182.5, 236.06], [238.78, 227.86], [283.82, 227.89], [234.22, 191.09], [182.5, 191.08], [238.78, 188.71], [283.82, 188.72]]}]\n```" - } - ] - }, - { - "id": 2735, - "data_source": "Mono3DRefer", - "scene_name": "003351", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[667.44, 214.94], [733.35, 214.95], [825.78, 246.52], [710.42, 246.51], [667.44, 143.53], [733.35, 143.52], [825.78, 121.52], [710.42, 121.53]]}]\n```" - } - ] - }, - { - "id": 2736, - "data_source": "Mono3DRefer", - "scene_name": "003351", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[422.53, 199.23], [398.76, 199.2], [415.37, 197.31], [437.46, 197.33], [422.53, 179.63], [398.76, 179.63], [415.37, 179.14], [437.46, 179.15]]}, {\"category\": \"bus\", \"corners_3d\": [[667.44, 214.94], [733.35, 214.95], [825.78, 246.52], [710.42, 246.51], [667.44, 143.53], [733.35, 143.52], [825.78, 121.52], [710.42, 121.53]]}, {\"category\": \"car\", \"corners_3d\": [[234.22, 236.1], [182.5, 236.06], [238.78, 227.86], [283.82, 227.89], [234.22, 191.09], [182.5, 191.08], [238.78, 188.71], [283.82, 188.72]]}]\n```" - } - ] - }, - { - "id": 2737, - "data_source": "Mono3DRefer", - "scene_name": "003358", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[866.48, 385.71], [1061.44, 386.33], [1739.64, 712.34], [1244.16, 708.41], [866.48, 210.0], [1061.44, 210.11], [1739.64, 267.01], [1244.16, 266.32]]}, {\"category\": \"car\", \"corners_3d\": [[553.56, 228.29], [497.02, 228.34], [514.91, 218.86], [561.79, 218.83], [553.56, 176.2], [497.02, 176.21], [514.91, 175.63], [561.79, 175.63]]}, {\"category\": \"car\", \"corners_3d\": [[660.68, 210.84], [693.81, 210.8], [707.91, 216.25], [670.02, 216.3], [660.68, 176.1], [693.81, 176.1], [707.91, 176.57], [670.02, 176.57]]}]\n```" - } - ] - }, - { - "id": 2738, - "data_source": "Mono3DRefer", - "scene_name": "003358", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[737.7, 280.55], [843.19, 280.41], [986.22, 344.43], [818.11, 344.8], [737.7, 169.64], [843.19, 169.65], [986.22, 167.74], [818.11, 167.73]]}, {\"category\": \"van\", \"corners_3d\": [[676.36, 217.7], [724.84, 217.61], [754.06, 227.2], [695.24, 227.33], [676.36, 162.32], [724.84, 162.34], [754.06, 160.09], [695.24, 160.06]]}]\n```" - } - ] - }, - { - "id": 2739, - "data_source": "Mono3DRefer", - "scene_name": "003358", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[866.48, 385.71], [1061.44, 386.33], [1739.64, 712.34], [1244.16, 708.41], [866.48, 210.0], [1061.44, 210.11], [1739.64, 267.01], [1244.16, 266.32]]}, {\"category\": \"van\", \"corners_3d\": [[737.7, 280.55], [843.19, 280.41], [986.22, 344.43], [818.11, 344.8], [737.7, 169.64], [843.19, 169.65], [986.22, 167.74], [818.11, 167.73]]}, {\"category\": \"car\", \"corners_3d\": [[553.56, 228.29], [497.02, 228.34], [514.91, 218.86], [561.79, 218.83], [553.56, 176.2], [497.02, 176.21], [514.91, 175.63], [561.79, 175.63]]}, {\"category\": \"van\", \"corners_3d\": [[676.36, 217.7], [724.84, 217.61], [754.06, 227.2], [695.24, 227.33], [676.36, 162.32], [724.84, 162.34], [754.06, 160.09], [695.24, 160.06]]}, {\"category\": \"car\", \"corners_3d\": [[660.68, 210.84], [693.81, 210.8], [707.91, 216.25], [670.02, 216.3], [660.68, 176.1], [693.81, 176.1], [707.91, 176.57], [670.02, 176.57]]}]\n```" - } - ] - }, - { - "id": 2740, - "data_source": "Mono3DRefer", - "scene_name": "003360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[615.7, 202.61], [592.32, 203.03], [545.83, 198.31], [566.14, 198.01], [615.7, 168.63], [592.32, 168.57], [545.83, 169.24], [566.14, 169.28]]}]\n```" - } - ] - }, - { - "id": 2741, - "data_source": "Mono3DRefer", - "scene_name": "003360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[626.15, 213.63], [658.22, 213.98], [645.76, 219.62], [609.46, 219.17], [626.15, 185.66], [658.22, 185.77], [645.76, 187.55], [609.46, 187.41]]}]\n```" - } - ] - }, - { - "id": 2742, - "data_source": "Mono3DRefer", - "scene_name": "003360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[683.68, 217.7], [720.35, 218.18], [713.57, 225.23], [671.35, 224.59], [683.68, 174.56], [720.35, 174.57], [713.57, 174.84], [671.35, 174.82]]}]\n```" - } - ] - }, - { - "id": 2743, - "data_source": "Mono3DRefer", - "scene_name": "003360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[626.15, 213.63], [658.22, 213.98], [645.76, 219.62], [609.46, 219.17], [626.15, 185.66], [658.22, 185.77], [645.76, 187.55], [609.46, 187.41]]}, {\"category\": \"tram\", \"corners_3d\": [[615.7, 202.61], [592.32, 203.03], [545.83, 198.31], [566.14, 198.01], [615.7, 168.63], [592.32, 168.57], [545.83, 169.24], [566.14, 169.28]]}]\n```" - } - ] - }, - { - "id": 2744, - "data_source": "Mono3DRefer", - "scene_name": "003360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[683.68, 217.7], [720.35, 218.18], [713.57, 225.23], [671.35, 224.59], [683.68, 174.56], [720.35, 174.57], [713.57, 174.84], [671.35, 174.82]]}, {\"category\": \"tram\", \"corners_3d\": [[615.7, 202.61], [592.32, 203.03], [545.83, 198.31], [566.14, 198.01], [615.7, 168.63], [592.32, 168.57], [545.83, 169.24], [566.14, 169.28]]}]\n```" - } - ] - }, - { - "id": 2745, - "data_source": "Mono3DRefer", - "scene_name": "003360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[626.15, 213.63], [658.22, 213.98], [645.76, 219.62], [609.46, 219.17], [626.15, 185.66], [658.22, 185.77], [645.76, 187.55], [609.46, 187.41]]}, {\"category\": \"van\", \"corners_3d\": [[683.68, 217.7], [720.35, 218.18], [713.57, 225.23], [671.35, 224.59], [683.68, 174.56], [720.35, 174.57], [713.57, 174.84], [671.35, 174.82]]}]\n```" - } - ] - }, - { - "id": 2746, - "data_source": "Mono3DRefer", - "scene_name": "003360", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[626.15, 213.63], [658.22, 213.98], [645.76, 219.62], [609.46, 219.17], [626.15, 185.66], [658.22, 185.77], [645.76, 187.55], [609.46, 187.41]]}, {\"category\": \"van\", \"corners_3d\": [[683.68, 217.7], [720.35, 218.18], [713.57, 225.23], [671.35, 224.59], [683.68, 174.56], [720.35, 174.57], [713.57, 174.84], [671.35, 174.82]]}, {\"category\": \"tram\", \"corners_3d\": [[615.7, 202.61], [592.32, 203.03], [545.83, 198.31], [566.14, 198.01], [615.7, 168.63], [592.32, 168.57], [545.83, 169.24], [566.14, 169.28]]}]\n```" - } - ] - }, - { - "id": 2747, - "data_source": "Mono3DRefer", - "scene_name": "003376", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[564.76, 201.0], [539.06, 201.05], [542.27, 198.84], [565.96, 198.8], [564.76, 175.68], [539.06, 175.68], [542.27, 175.46], [565.96, 175.46]]}]\n```" - } - ] - }, - { - "id": 2748, - "data_source": "Mono3DRefer", - "scene_name": "003409", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[530.79, 206.91], [502.27, 206.86], [514.8, 203.72], [540.7, 203.76], [530.79, 181.41], [502.27, 181.39], [514.8, 180.61], [540.7, 180.62]]}, {\"category\": \"car\", \"corners_3d\": [[665.72, 199.79], [695.67, 199.79], [705.21, 202.8], [671.93, 202.8], [665.72, 172.29], [695.67, 172.29], [705.21, 172.22], [671.93, 172.22]]}, {\"category\": \"car\", \"corners_3d\": [[947.29, 407.34], [1193.41, 406.6], [4716.84, 1801.32], [3048.28, 1837.9], [947.29, 175.79], [1193.41, 175.78], [4716.84, 193.23], [3048.28, 193.69]]}, {\"category\": \"car\", \"corners_3d\": [[108.13, 323.98], [230.28, 324.52], [-39.78, 428.4], [-243.48, 426.87], [108.13, 198.91], [230.28, 199.01], [-39.78, 216.92], [-243.48, 216.66]]}, {\"category\": \"car\", \"corners_3d\": [[403.57, 254.91], [330.98, 254.92], [383.17, 239.51], [442.13, 239.51], [403.57, 188.82], [330.98, 188.82], [383.17, 185.82], [442.13, 185.82]]}, {\"category\": \"car\", \"corners_3d\": [[451.5, 234.34], [401.24, 234.22], [435.26, 225.12], [478.11, 225.21], [451.5, 187.89], [401.24, 187.86], [435.26, 185.63], [478.11, 185.66]]}, {\"category\": \"car\", \"corners_3d\": [[483.15, 222.98], [443.58, 223.01], [462.67, 216.96], [497.45, 216.94], [483.15, 184.85], [443.58, 184.85], [462.67, 183.4], [497.45, 183.4]]}, {\"category\": \"car\", \"corners_3d\": [[508.14, 214.61], [471.79, 214.64], [485.98, 210.08], [518.35, 210.06], [508.14, 183.84], [471.79, 183.84], [485.98, 182.64], [518.35, 182.64]]}, {\"category\": \"car\", \"corners_3d\": [[792.52, 195.0], [783.26, 193.88], [875.53, 193.88], [889.74, 195.01], [792.52, 155.07], [783.26, 155.97], [875.53, 155.97], [889.74, 155.07]]}, {\"category\": \"car\", \"corners_3d\": [[764.54, 191.71], [756.65, 190.84], [824.76, 190.8], [835.94, 191.67], [764.54, 156.71], [756.65, 157.45], [824.76, 157.48], [835.94, 156.74]]}, {\"category\": \"car\", \"corners_3d\": [[1209.19, 172.61], [1247.91, 172.59], [1140.36, 172.59], [1108.16, 172.61], [1209.19, 136.41], [1247.91, 134.06], [1140.36, 134.06], [1108.16, 136.41]]}, {\"category\": \"car\", \"corners_3d\": [[1164.47, 174.37], [1196.81, 174.46], [1092.71, 174.46], [1066.2, 174.37], [1164.47, 138.05], [1196.81, 135.99], [1092.71, 136.04], [1066.2, 138.09]]}, {\"category\": \"car\", \"corners_3d\": [[1135.78, 173.05], [1164.15, 173.06], [1060.26, 173.06], [1037.2, 173.05], [1135.78, 142.51], [1164.15, 140.88], [1060.26, 140.88], [1037.2, 142.51]]}, {\"category\": \"car\", \"corners_3d\": [[1103.52, 173.04], [1128.86, 173.04], [1032.4, 173.04], [1011.67, 173.04], [1103.52, 143.56], [1128.86, 142.08], [1032.4, 142.04], [1011.67, 143.52]]}]\n```" - } - ] - }, - { - "id": 2749, - "data_source": "Mono3DRefer", - "scene_name": "003410", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[284.6, 310.47], [430.88, 309.67], [286.21, 435.26], [2.48, 438.23], [284.6, 197.46], [430.88, 197.32], [286.21, 219.78], [2.48, 220.31]]}, {\"category\": \"car\", \"corners_3d\": [[654.98, 220.69], [704.36, 220.52], [735.12, 231.78], [674.17, 232.02], [654.98, 173.16], [704.36, 173.16], [735.12, 173.23], [674.17, 173.23]]}, {\"category\": \"car\", \"corners_3d\": [[440.7, 238.47], [504.57, 238.24], [480.85, 258.46], [397.04, 258.85], [440.7, 179.48], [504.57, 179.46], [480.85, 181.51], [397.04, 181.54]]}, {\"category\": \"car\", \"corners_3d\": [[732.62, 285.97], [846.47, 285.63], [1004.96, 356.91], [819.6, 357.83], [732.62, 171.39], [846.47, 171.39], [1004.96, 170.47], [819.6, 170.45]]}, {\"category\": \"car\", \"corners_3d\": [[630.15, 198.87], [660.49, 198.94], [660.47, 201.52], [627.15, 201.42], [630.15, 173.4], [660.49, 173.4], [660.47, 173.46], [627.15, 173.45]]}]\n```" - } - ] - }, - { - "id": 2750, - "data_source": "Mono3DRefer", - "scene_name": "003410", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[687.84, 182.29], [706.18, 182.4], [687.41, 183.06], [668.07, 182.93], [687.84, 157.09], [706.18, 156.9], [687.41, 155.79], [668.07, 156.01]]}]\n```" - } - ] - }, - { - "id": 2751, - "data_source": "Mono3DRefer", - "scene_name": "003410", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[284.6, 310.47], [430.88, 309.67], [286.21, 435.26], [2.48, 438.23], [284.6, 197.46], [430.88, 197.32], [286.21, 219.78], [2.48, 220.31]]}, {\"category\": \"car\", \"corners_3d\": [[654.98, 220.69], [704.36, 220.52], [735.12, 231.78], [674.17, 232.02], [654.98, 173.16], [704.36, 173.16], [735.12, 173.23], [674.17, 173.23]]}, {\"category\": \"car\", \"corners_3d\": [[440.7, 238.47], [504.57, 238.24], [480.85, 258.46], [397.04, 258.85], [440.7, 179.48], [504.57, 179.46], [480.85, 181.51], [397.04, 181.54]]}, {\"category\": \"car\", \"corners_3d\": [[732.62, 285.97], [846.47, 285.63], [1004.96, 356.91], [819.6, 357.83], [732.62, 171.39], [846.47, 171.39], [1004.96, 170.47], [819.6, 170.45]]}, {\"category\": \"car\", \"corners_3d\": [[630.15, 198.87], [660.49, 198.94], [660.47, 201.52], [627.15, 201.42], [630.15, 173.4], [660.49, 173.4], [660.47, 173.46], [627.15, 173.45]]}, {\"category\": \"van\", \"corners_3d\": [[687.84, 182.29], [706.18, 182.4], [687.41, 183.06], [668.07, 182.93], [687.84, 157.09], [706.18, 156.9], [687.41, 155.79], [668.07, 156.01]]}]\n```" - } - ] - }, - { - "id": 2752, - "data_source": "Mono3DRefer", - "scene_name": "003416", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[456.58, 272.59], [424.24, 287.38], [157.88, 283.47], [222.62, 269.61], [456.58, 194.81], [424.24, 198.07], [157.88, 197.21], [222.62, 194.16]]}, {\"category\": \"car\", \"corners_3d\": [[1197.17, 300.42], [1351.0, 336.47], [882.3, 330.46], [833.21, 296.73], [1197.17, 173.68], [1351.0, 173.91], [882.3, 173.87], [833.21, 173.66]]}, {\"category\": \"car\", \"corners_3d\": [[1033.72, 269.14], [1106.39, 284.83], [799.2, 286.29], [769.25, 270.21], [1033.72, 163.08], [1106.39, 161.49], [799.2, 161.34], [769.25, 162.97]]}, {\"category\": \"car\", \"corners_3d\": [[368.77, 234.7], [390.3, 230.36], [518.5, 231.32], [506.42, 235.81], [368.77, 186.59], [390.3, 185.62], [518.5, 185.84], [506.42, 186.84]]}, {\"category\": \"car\", \"corners_3d\": [[418.54, 221.34], [425.8, 218.52], [536.63, 217.89], [536.22, 220.63], [418.54, 184.36], [425.8, 183.7], [536.63, 183.55], [536.22, 184.2]]}]\n```" - } - ] - }, - { - "id": 2753, - "data_source": "Mono3DRefer", - "scene_name": "003445", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[625.36, 218.12], [587.67, 218.12], [590.4, 212.33], [623.26, 212.33], [625.36, 184.7], [587.67, 184.7], [590.4, 183.18], [623.26, 183.18]]}]\n```" - } - ] - }, - { - "id": 2754, - "data_source": "Mono3DRefer", - "scene_name": "003451", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[481.35, 240.92], [418.01, 241.75], [433.2, 230.03], [485.72, 229.46], [481.35, 177.38], [418.01, 177.43], [433.2, 176.66], [485.72, 176.62]]}, {\"category\": \"car\", \"corners_3d\": [[483.21, 212.33], [445.68, 212.91], [439.79, 208.38], [473.2, 207.92], [483.21, 180.61], [445.68, 180.73], [439.79, 179.84], [473.2, 179.75]]}]\n```" - } - ] - }, - { - "id": 2755, - "data_source": "Mono3DRefer", - "scene_name": "003452", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[387.43, 222.63], [354.73, 223.47], [351.06, 219.83], [381.5, 219.1], [387.43, 193.5], [354.73, 193.85], [351.06, 192.34], [381.5, 192.04]]}, {\"category\": \"car\", \"corners_3d\": [[401.56, 335.67], [243.06, 338.84], [350.14, 279.76], [451.59, 278.43], [401.56, 204.14], [243.06, 204.75], [350.14, 193.4], [451.59, 193.14]]}]\n```" - } - ] - }, - { - "id": 2756, - "data_source": "Mono3DRefer", - "scene_name": "003466", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[468.77, 265.28], [416.86, 264.92], [447.0, 253.8], [492.08, 254.07], [468.77, 158.44], [416.86, 158.54], [447.0, 161.6], [492.08, 161.52]]}, {\"category\": \"cyclist\", \"corners_3d\": [[368.3, 296.23], [313.97, 296.15], [367.62, 275.16], [412.02, 275.21], [368.3, 154.13], [313.97, 154.14], [367.62, 159.15], [412.02, 159.14]]}]\n```" - } - ] - }, - { - "id": 2757, - "data_source": "Mono3DRefer", - "scene_name": "003466", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[584.43, 180.54], [557.85, 180.54], [561.31, 180.62], [585.7, 180.62], [584.43, 156.97], [557.85, 156.97], [561.31, 158.99], [585.7, 158.99]]}]\n```" - } - ] - }, - { - "id": 2758, - "data_source": "Mono3DRefer", - "scene_name": "003466", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[776.55, 259.39], [833.21, 259.21], [855.11, 265.91], [793.61, 266.13], [776.55, 158.15], [833.21, 158.21], [855.11, 156.2], [793.61, 156.13]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[685.1, 213.65], [716.98, 213.66], [719.74, 214.47], [687.05, 214.46], [685.1, 153.97], [716.98, 153.96], [719.74, 153.26], [687.05, 153.28]]}]\n```" - } - ] - }, - { - "id": 2759, - "data_source": "Mono3DRefer", - "scene_name": "003466", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[605.54, 189.16], [646.13, 189.17], [653.42, 190.64], [605.04, 190.63], [605.54, 138.75], [646.13, 138.73], [653.42, 130.51], [605.04, 130.55]]}]\n```" - } - ] - }, - { - "id": 2760, - "data_source": "Mono3DRefer", - "scene_name": "003466", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[468.77, 265.28], [416.86, 264.92], [447.0, 253.8], [492.08, 254.07], [468.77, 158.44], [416.86, 158.54], [447.0, 161.6], [492.08, 161.52]]}, {\"category\": \"car\", \"corners_3d\": [[584.43, 180.54], [557.85, 180.54], [561.31, 180.62], [585.7, 180.62], [584.43, 156.97], [557.85, 156.97], [561.31, 158.99], [585.7, 158.99]]}, {\"category\": \"cyclist\", \"corners_3d\": [[368.3, 296.23], [313.97, 296.15], [367.62, 275.16], [412.02, 275.21], [368.3, 154.13], [313.97, 154.14], [367.62, 159.15], [412.02, 159.14]]}]\n```" - } - ] - }, - { - "id": 2761, - "data_source": "Mono3DRefer", - "scene_name": "003466", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[776.55, 259.39], [833.21, 259.21], [855.11, 265.91], [793.61, 266.13], [776.55, 158.15], [833.21, 158.21], [855.11, 156.2], [793.61, 156.13]]}, {\"category\": \"cyclist\", \"corners_3d\": [[468.77, 265.28], [416.86, 264.92], [447.0, 253.8], [492.08, 254.07], [468.77, 158.44], [416.86, 158.54], [447.0, 161.6], [492.08, 161.52]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[685.1, 213.65], [716.98, 213.66], [719.74, 214.47], [687.05, 214.46], [685.1, 153.97], [716.98, 153.96], [719.74, 153.26], [687.05, 153.28]]}, {\"category\": \"cyclist\", \"corners_3d\": [[368.3, 296.23], [313.97, 296.15], [367.62, 275.16], [412.02, 275.21], [368.3, 154.13], [313.97, 154.14], [367.62, 159.15], [412.02, 159.14]]}]\n```" - } - ] - }, - { - "id": 2762, - "data_source": "Mono3DRefer", - "scene_name": "003466", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[468.77, 265.28], [416.86, 264.92], [447.0, 253.8], [492.08, 254.07], [468.77, 158.44], [416.86, 158.54], [447.0, 161.6], [492.08, 161.52]]}, {\"category\": \"cyclist\", \"corners_3d\": [[368.3, 296.23], [313.97, 296.15], [367.62, 275.16], [412.02, 275.21], [368.3, 154.13], [313.97, 154.14], [367.62, 159.15], [412.02, 159.14]]}, {\"category\": \"van\", \"corners_3d\": [[605.54, 189.16], [646.13, 189.17], [653.42, 190.64], [605.04, 190.63], [605.54, 138.75], [646.13, 138.73], [653.42, 130.51], [605.04, 130.55]]}]\n```" - } - ] - }, - { - "id": 2763, - "data_source": "Mono3DRefer", - "scene_name": "003466", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[776.55, 259.39], [833.21, 259.21], [855.11, 265.91], [793.61, 266.13], [776.55, 158.15], [833.21, 158.21], [855.11, 156.2], [793.61, 156.13]]}, {\"category\": \"car\", \"corners_3d\": [[584.43, 180.54], [557.85, 180.54], [561.31, 180.62], [585.7, 180.62], [584.43, 156.97], [557.85, 156.97], [561.31, 158.99], [585.7, 158.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[685.1, 213.65], [716.98, 213.66], [719.74, 214.47], [687.05, 214.46], [685.1, 153.97], [716.98, 153.96], [719.74, 153.26], [687.05, 153.28]]}]\n```" - } - ] - }, - { - "id": 2764, - "data_source": "Mono3DRefer", - "scene_name": "003466", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[584.43, 180.54], [557.85, 180.54], [561.31, 180.62], [585.7, 180.62], [584.43, 156.97], [557.85, 156.97], [561.31, 158.99], [585.7, 158.99]]}, {\"category\": \"van\", \"corners_3d\": [[605.54, 189.16], [646.13, 189.17], [653.42, 190.64], [605.04, 190.63], [605.54, 138.75], [646.13, 138.73], [653.42, 130.51], [605.04, 130.55]]}]\n```" - } - ] - }, - { - "id": 2765, - "data_source": "Mono3DRefer", - "scene_name": "003466", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[776.55, 259.39], [833.21, 259.21], [855.11, 265.91], [793.61, 266.13], [776.55, 158.15], [833.21, 158.21], [855.11, 156.2], [793.61, 156.13]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[685.1, 213.65], [716.98, 213.66], [719.74, 214.47], [687.05, 214.46], [685.1, 153.97], [716.98, 153.96], [719.74, 153.26], [687.05, 153.28]]}, {\"category\": \"van\", \"corners_3d\": [[605.54, 189.16], [646.13, 189.17], [653.42, 190.64], [605.04, 190.63], [605.54, 138.75], [646.13, 138.73], [653.42, 130.51], [605.04, 130.55]]}]\n```" - } - ] - }, - { - "id": 2766, - "data_source": "Mono3DRefer", - "scene_name": "003466", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[776.55, 259.39], [833.21, 259.21], [855.11, 265.91], [793.61, 266.13], [776.55, 158.15], [833.21, 158.21], [855.11, 156.2], [793.61, 156.13]]}, {\"category\": \"cyclist\", \"corners_3d\": [[468.77, 265.28], [416.86, 264.92], [447.0, 253.8], [492.08, 254.07], [468.77, 158.44], [416.86, 158.54], [447.0, 161.6], [492.08, 161.52]]}, {\"category\": \"car\", \"corners_3d\": [[584.43, 180.54], [557.85, 180.54], [561.31, 180.62], [585.7, 180.62], [584.43, 156.97], [557.85, 156.97], [561.31, 158.99], [585.7, 158.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[685.1, 213.65], [716.98, 213.66], [719.74, 214.47], [687.05, 214.46], [685.1, 153.97], [716.98, 153.96], [719.74, 153.26], [687.05, 153.28]]}, {\"category\": \"cyclist\", \"corners_3d\": [[368.3, 296.23], [313.97, 296.15], [367.62, 275.16], [412.02, 275.21], [368.3, 154.13], [313.97, 154.14], [367.62, 159.15], [412.02, 159.14]]}]\n```" - } - ] - }, - { - "id": 2767, - "data_source": "Mono3DRefer", - "scene_name": "003466", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[468.77, 265.28], [416.86, 264.92], [447.0, 253.8], [492.08, 254.07], [468.77, 158.44], [416.86, 158.54], [447.0, 161.6], [492.08, 161.52]]}, {\"category\": \"car\", \"corners_3d\": [[584.43, 180.54], [557.85, 180.54], [561.31, 180.62], [585.7, 180.62], [584.43, 156.97], [557.85, 156.97], [561.31, 158.99], [585.7, 158.99]]}, {\"category\": \"cyclist\", \"corners_3d\": [[368.3, 296.23], [313.97, 296.15], [367.62, 275.16], [412.02, 275.21], [368.3, 154.13], [313.97, 154.14], [367.62, 159.15], [412.02, 159.14]]}, {\"category\": \"van\", \"corners_3d\": [[605.54, 189.16], [646.13, 189.17], [653.42, 190.64], [605.04, 190.63], [605.54, 138.75], [646.13, 138.73], [653.42, 130.51], [605.04, 130.55]]}]\n```" - } - ] - }, - { - "id": 2768, - "data_source": "Mono3DRefer", - "scene_name": "003466", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[776.55, 259.39], [833.21, 259.21], [855.11, 265.91], [793.61, 266.13], [776.55, 158.15], [833.21, 158.21], [855.11, 156.2], [793.61, 156.13]]}, {\"category\": \"cyclist\", \"corners_3d\": [[468.77, 265.28], [416.86, 264.92], [447.0, 253.8], [492.08, 254.07], [468.77, 158.44], [416.86, 158.54], [447.0, 161.6], [492.08, 161.52]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[685.1, 213.65], [716.98, 213.66], [719.74, 214.47], [687.05, 214.46], [685.1, 153.97], [716.98, 153.96], [719.74, 153.26], [687.05, 153.28]]}, {\"category\": \"cyclist\", \"corners_3d\": [[368.3, 296.23], [313.97, 296.15], [367.62, 275.16], [412.02, 275.21], [368.3, 154.13], [313.97, 154.14], [367.62, 159.15], [412.02, 159.14]]}, {\"category\": \"van\", \"corners_3d\": [[605.54, 189.16], [646.13, 189.17], [653.42, 190.64], [605.04, 190.63], [605.54, 138.75], [646.13, 138.73], [653.42, 130.51], [605.04, 130.55]]}]\n```" - } - ] - }, - { - "id": 2769, - "data_source": "Mono3DRefer", - "scene_name": "003466", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[776.55, 259.39], [833.21, 259.21], [855.11, 265.91], [793.61, 266.13], [776.55, 158.15], [833.21, 158.21], [855.11, 156.2], [793.61, 156.13]]}, {\"category\": \"car\", \"corners_3d\": [[584.43, 180.54], [557.85, 180.54], [561.31, 180.62], [585.7, 180.62], [584.43, 156.97], [557.85, 156.97], [561.31, 158.99], [585.7, 158.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[685.1, 213.65], [716.98, 213.66], [719.74, 214.47], [687.05, 214.46], [685.1, 153.97], [716.98, 153.96], [719.74, 153.26], [687.05, 153.28]]}, {\"category\": \"van\", \"corners_3d\": [[605.54, 189.16], [646.13, 189.17], [653.42, 190.64], [605.04, 190.63], [605.54, 138.75], [646.13, 138.73], [653.42, 130.51], [605.04, 130.55]]}]\n```" - } - ] - }, - { - "id": 2770, - "data_source": "Mono3DRefer", - "scene_name": "003466", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[776.55, 259.39], [833.21, 259.21], [855.11, 265.91], [793.61, 266.13], [776.55, 158.15], [833.21, 158.21], [855.11, 156.2], [793.61, 156.13]]}, {\"category\": \"cyclist\", \"corners_3d\": [[468.77, 265.28], [416.86, 264.92], [447.0, 253.8], [492.08, 254.07], [468.77, 158.44], [416.86, 158.54], [447.0, 161.6], [492.08, 161.52]]}, {\"category\": \"car\", \"corners_3d\": [[584.43, 180.54], [557.85, 180.54], [561.31, 180.62], [585.7, 180.62], [584.43, 156.97], [557.85, 156.97], [561.31, 158.99], [585.7, 158.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[685.1, 213.65], [716.98, 213.66], [719.74, 214.47], [687.05, 214.46], [685.1, 153.97], [716.98, 153.96], [719.74, 153.26], [687.05, 153.28]]}, {\"category\": \"cyclist\", \"corners_3d\": [[368.3, 296.23], [313.97, 296.15], [367.62, 275.16], [412.02, 275.21], [368.3, 154.13], [313.97, 154.14], [367.62, 159.15], [412.02, 159.14]]}, {\"category\": \"van\", \"corners_3d\": [[605.54, 189.16], [646.13, 189.17], [653.42, 190.64], [605.04, 190.63], [605.54, 138.75], [646.13, 138.73], [653.42, 130.51], [605.04, 130.55]]}]\n```" - } - ] - }, - { - "id": 2771, - "data_source": "Mono3DRefer", - "scene_name": "003478", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[586.15, 261.88], [558.81, 261.85], [563.16, 255.1], [588.43, 255.13], [586.15, 171.84], [558.81, 171.84], [563.16, 171.92], [588.43, 171.92]]}]\n```" - } - ] - }, - { - "id": 2772, - "data_source": "Mono3DRefer", - "scene_name": "003478", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[594.49, 192.36], [572.21, 192.29], [580.12, 191.05], [601.0, 191.11], [594.49, 169.66], [572.21, 169.67], [580.12, 169.87], [601.0, 169.86]]}]\n```" - } - ] - }, - { - "id": 2773, - "data_source": "Mono3DRefer", - "scene_name": "003478", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[515.47, 252.71], [439.69, 252.55], [478.01, 235.93], [538.02, 236.03], [515.47, 173.3], [439.69, 173.3], [478.01, 173.21], [538.02, 173.21]]}, {\"category\": \"car\", \"corners_3d\": [[806.09, 364.72], [988.58, 365.73], [1221.97, 489.39], [921.49, 486.7], [806.09, 184.17], [988.58, 184.23], [1221.97, 191.52], [921.49, 191.36]]}, {\"category\": \"car\", \"corners_3d\": [[727.7, 276.37], [828.52, 277.24], [896.65, 317.87], [756.26, 316.2], [727.7, 190.92], [828.52, 191.07], [896.65, 198.16], [756.26, 197.87]]}, {\"category\": \"car\", \"corners_3d\": [[523.61, 218.85], [567.53, 219.02], [553.94, 226.32], [503.16, 226.09], [523.61, 181.83], [567.53, 181.87], [553.94, 183.29], [503.16, 183.25]]}, {\"category\": \"car\", \"corners_3d\": [[676.36, 204.59], [647.66, 204.59], [644.38, 201.9], [670.66, 201.9], [676.36, 178.02], [647.66, 178.02], [644.38, 177.58], [670.66, 177.58]]}, {\"category\": \"car\", \"corners_3d\": [[-29.3, 282.47], [57.93, 267.7], [321.0, 268.1], [274.52, 283.01], [-29.3, 172.83], [57.93, 172.83], [321.0, 172.83], [274.52, 172.83]]}]\n```" - } - ] - }, - { - "id": 2774, - "data_source": "Mono3DRefer", - "scene_name": "003478", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[586.15, 261.88], [558.81, 261.85], [563.16, 255.1], [588.43, 255.13], [586.15, 171.84], [558.81, 171.84], [563.16, 171.92], [588.43, 171.92]]}, {\"category\": \"van\", \"corners_3d\": [[594.49, 192.36], [572.21, 192.29], [580.12, 191.05], [601.0, 191.11], [594.49, 169.66], [572.21, 169.67], [580.12, 169.87], [601.0, 169.86]]}]\n```" - } - ] - }, - { - "id": 2775, - "data_source": "Mono3DRefer", - "scene_name": "003478", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[515.47, 252.71], [439.69, 252.55], [478.01, 235.93], [538.02, 236.03], [515.47, 173.3], [439.69, 173.3], [478.01, 173.21], [538.02, 173.21]]}, {\"category\": \"car\", \"corners_3d\": [[806.09, 364.72], [988.58, 365.73], [1221.97, 489.39], [921.49, 486.7], [806.09, 184.17], [988.58, 184.23], [1221.97, 191.52], [921.49, 191.36]]}, {\"category\": \"cyclist\", \"corners_3d\": [[586.15, 261.88], [558.81, 261.85], [563.16, 255.1], [588.43, 255.13], [586.15, 171.84], [558.81, 171.84], [563.16, 171.92], [588.43, 171.92]]}, {\"category\": \"car\", \"corners_3d\": [[727.7, 276.37], [828.52, 277.24], [896.65, 317.87], [756.26, 316.2], [727.7, 190.92], [828.52, 191.07], [896.65, 198.16], [756.26, 197.87]]}, {\"category\": \"car\", \"corners_3d\": [[523.61, 218.85], [567.53, 219.02], [553.94, 226.32], [503.16, 226.09], [523.61, 181.83], [567.53, 181.87], [553.94, 183.29], [503.16, 183.25]]}, {\"category\": \"car\", \"corners_3d\": [[676.36, 204.59], [647.66, 204.59], [644.38, 201.9], [670.66, 201.9], [676.36, 178.02], [647.66, 178.02], [644.38, 177.58], [670.66, 177.58]]}, {\"category\": \"car\", \"corners_3d\": [[-29.3, 282.47], [57.93, 267.7], [321.0, 268.1], [274.52, 283.01], [-29.3, 172.83], [57.93, 172.83], [321.0, 172.83], [274.52, 172.83]]}]\n```" - } - ] - }, - { - "id": 2776, - "data_source": "Mono3DRefer", - "scene_name": "003478", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[515.47, 252.71], [439.69, 252.55], [478.01, 235.93], [538.02, 236.03], [515.47, 173.3], [439.69, 173.3], [478.01, 173.21], [538.02, 173.21]]}, {\"category\": \"car\", \"corners_3d\": [[806.09, 364.72], [988.58, 365.73], [1221.97, 489.39], [921.49, 486.7], [806.09, 184.17], [988.58, 184.23], [1221.97, 191.52], [921.49, 191.36]]}, {\"category\": \"car\", \"corners_3d\": [[727.7, 276.37], [828.52, 277.24], [896.65, 317.87], [756.26, 316.2], [727.7, 190.92], [828.52, 191.07], [896.65, 198.16], [756.26, 197.87]]}, {\"category\": \"car\", \"corners_3d\": [[523.61, 218.85], [567.53, 219.02], [553.94, 226.32], [503.16, 226.09], [523.61, 181.83], [567.53, 181.87], [553.94, 183.29], [503.16, 183.25]]}, {\"category\": \"car\", \"corners_3d\": [[676.36, 204.59], [647.66, 204.59], [644.38, 201.9], [670.66, 201.9], [676.36, 178.02], [647.66, 178.02], [644.38, 177.58], [670.66, 177.58]]}, {\"category\": \"van\", \"corners_3d\": [[594.49, 192.36], [572.21, 192.29], [580.12, 191.05], [601.0, 191.11], [594.49, 169.66], [572.21, 169.67], [580.12, 169.87], [601.0, 169.86]]}, {\"category\": \"car\", \"corners_3d\": [[-29.3, 282.47], [57.93, 267.7], [321.0, 268.1], [274.52, 283.01], [-29.3, 172.83], [57.93, 172.83], [321.0, 172.83], [274.52, 172.83]]}]\n```" - } - ] - }, - { - "id": 2777, - "data_source": "Mono3DRefer", - "scene_name": "003478", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[515.47, 252.71], [439.69, 252.55], [478.01, 235.93], [538.02, 236.03], [515.47, 173.3], [439.69, 173.3], [478.01, 173.21], [538.02, 173.21]]}, {\"category\": \"car\", \"corners_3d\": [[806.09, 364.72], [988.58, 365.73], [1221.97, 489.39], [921.49, 486.7], [806.09, 184.17], [988.58, 184.23], [1221.97, 191.52], [921.49, 191.36]]}, {\"category\": \"cyclist\", \"corners_3d\": [[586.15, 261.88], [558.81, 261.85], [563.16, 255.1], [588.43, 255.13], [586.15, 171.84], [558.81, 171.84], [563.16, 171.92], [588.43, 171.92]]}, {\"category\": \"car\", \"corners_3d\": [[727.7, 276.37], [828.52, 277.24], [896.65, 317.87], [756.26, 316.2], [727.7, 190.92], [828.52, 191.07], [896.65, 198.16], [756.26, 197.87]]}, {\"category\": \"car\", \"corners_3d\": [[523.61, 218.85], [567.53, 219.02], [553.94, 226.32], [503.16, 226.09], [523.61, 181.83], [567.53, 181.87], [553.94, 183.29], [503.16, 183.25]]}, {\"category\": \"car\", \"corners_3d\": [[676.36, 204.59], [647.66, 204.59], [644.38, 201.9], [670.66, 201.9], [676.36, 178.02], [647.66, 178.02], [644.38, 177.58], [670.66, 177.58]]}, {\"category\": \"van\", \"corners_3d\": [[594.49, 192.36], [572.21, 192.29], [580.12, 191.05], [601.0, 191.11], [594.49, 169.66], [572.21, 169.67], [580.12, 169.87], [601.0, 169.86]]}, {\"category\": \"car\", \"corners_3d\": [[-29.3, 282.47], [57.93, 267.7], [321.0, 268.1], [274.52, 283.01], [-29.3, 172.83], [57.93, 172.83], [321.0, 172.83], [274.52, 172.83]]}]\n```" - } - ] - }, - { - "id": 2778, - "data_source": "Mono3DRefer", - "scene_name": "003483", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[192.76, 495.13], [-125.94, 495.24], [259.86, 326.0], [411.22, 325.98], [192.76, 163.02], [-125.94, 163.01], [259.86, 168.18], [411.22, 168.18]]}, {\"category\": \"car\", \"corners_3d\": [[437.13, 252.4], [514.03, 252.67], [477.01, 277.76], [376.2, 277.3], [437.13, 187.64], [514.03, 187.69], [477.01, 192.35], [376.2, 192.27]]}, {\"category\": \"car\", \"corners_3d\": [[682.33, 220.02], [642.05, 220.12], [634.91, 214.68], [670.58, 214.59], [682.33, 182.63], [642.05, 182.65], [634.91, 181.52], [670.58, 181.51]]}]\n```" - } - ] - }, - { - "id": 2779, - "data_source": "Mono3DRefer", - "scene_name": "003483", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[571.08, 200.12], [543.1, 200.05], [551.73, 197.9], [577.52, 197.96], [571.08, 171.78], [543.1, 171.78], [551.73, 171.86], [577.52, 171.86]]}]\n```" - } - ] - }, - { - "id": 2780, - "data_source": "Mono3DRefer", - "scene_name": "003483", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[192.76, 495.13], [-125.94, 495.24], [259.86, 326.0], [411.22, 325.98], [192.76, 163.02], [-125.94, 163.01], [259.86, 168.18], [411.22, 168.18]]}, {\"category\": \"car\", \"corners_3d\": [[437.13, 252.4], [514.03, 252.67], [477.01, 277.76], [376.2, 277.3], [437.13, 187.64], [514.03, 187.69], [477.01, 192.35], [376.2, 192.27]]}, {\"category\": \"car\", \"corners_3d\": [[682.33, 220.02], [642.05, 220.12], [634.91, 214.68], [670.58, 214.59], [682.33, 182.63], [642.05, 182.65], [634.91, 181.52], [670.58, 181.51]]}, {\"category\": \"van\", \"corners_3d\": [[571.08, 200.12], [543.1, 200.05], [551.73, 197.9], [577.52, 197.96], [571.08, 171.78], [543.1, 171.78], [551.73, 171.86], [577.52, 171.86]]}]\n```" - } - ] - }, - { - "id": 2781, - "data_source": "Mono3DRefer", - "scene_name": "003486", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[698.39, 202.71], [722.41, 202.9], [720.31, 205.27], [694.42, 205.06], [698.39, 180.28], [722.41, 180.33], [720.31, 180.92], [694.42, 180.86]]}, {\"category\": \"car\", \"corners_3d\": [[-356.97, 463.73], [-580.15, 454.59], [9.17, 322.21], [135.69, 324.74], [-356.97, 144.69], [-580.15, 145.58], [9.17, 158.39], [135.69, 158.15]]}, {\"category\": \"car\", \"corners_3d\": [[759.23, 269.33], [855.5, 269.47], [960.56, 312.09], [821.69, 311.81], [759.23, 178.58], [855.5, 178.59], [960.56, 181.11], [821.69, 181.1]]}, {\"category\": \"car\", \"corners_3d\": [[613.03, 222.55], [571.11, 222.03], [590.75, 216.32], [627.94, 216.73], [613.03, 183.15], [571.11, 183.04], [590.75, 181.86], [627.94, 181.94]]}, {\"category\": \"car\", \"corners_3d\": [[745.3, 247.86], [809.01, 248.39], [849.26, 269.95], [767.17, 269.08], [745.3, 162.42], [809.01, 162.35], [849.26, 159.35], [767.17, 159.47]]}]\n```" - } - ] - }, - { - "id": 2782, - "data_source": "Mono3DRefer", - "scene_name": "003494", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[681.74, 239.46], [743.91, 239.23], [786.15, 256.47], [707.95, 256.83], [681.74, 180.33], [743.91, 180.3], [786.15, 182.24], [707.95, 182.28]]}, {\"category\": \"car\", \"corners_3d\": [[647.93, 210.43], [681.5, 210.45], [690.29, 215.6], [652.12, 215.58], [647.93, 178.61], [681.5, 178.62], [690.29, 179.4], [652.12, 179.4]]}, {\"category\": \"car\", \"corners_3d\": [[668.02, 226.14], [716.94, 226.11], [738.66, 236.25], [680.43, 236.3], [668.02, 177.77], [716.94, 177.77], [738.66, 178.71], [680.43, 178.71]]}, {\"category\": \"car\", \"corners_3d\": [[410.85, 219.83], [403.05, 221.82], [322.7, 221.92], [333.79, 219.93], [410.85, 190.32], [403.05, 191.06], [322.7, 191.1], [333.79, 190.36]]}, {\"category\": \"car\", \"corners_3d\": [[304.15, 222.61], [290.44, 224.79], [204.9, 224.73], [222.19, 222.55], [304.15, 191.9], [290.44, 192.73], [204.9, 192.71], [222.19, 191.87]]}, {\"category\": \"car\", \"corners_3d\": [[635.82, 199.28], [657.22, 199.26], [662.81, 201.4], [639.68, 201.43], [635.82, 178.66], [657.22, 178.66], [662.81, 179.13], [639.68, 179.13]]}, {\"category\": \"car\", \"corners_3d\": [[595.15, 200.2], [618.13, 200.19], [619.3, 202.3], [594.55, 202.31], [595.15, 178.49], [618.13, 178.49], [619.3, 178.92], [594.55, 178.92]]}, {\"category\": \"car\", \"corners_3d\": [[347.68, 395.96], [144.82, 396.01], [325.02, 309.38], [449.11, 309.36], [347.68, 211.06], [144.82, 211.07], [325.02, 196.23], [449.11, 196.23]]}, {\"category\": \"car\", \"corners_3d\": [[908.96, 430.26], [1160.01, 429.44], [3850.68, 1668.69], [2415.3, 1697.26], [908.96, 177.42], [1160.01, 177.41], [3850.68, 199.41], [2415.3, 199.92]]}, {\"category\": \"car\", \"corners_3d\": [[787.46, 326.3], [910.23, 325.0], [1074.61, 399.39], [893.15, 402.29], [787.46, 173.72], [910.23, 173.72], [1074.61, 174.14], [893.15, 174.16]]}]\n```" - } - ] - }, - { - "id": 2783, - "data_source": "Mono3DRefer", - "scene_name": "003494", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[898.48, 290.86], [951.73, 290.86], [978.51, 300.12], [921.08, 300.11], [898.48, 150.45], [951.73, 150.45], [978.51, 148.7], [921.08, 148.7]]}]\n```" - } - ] - }, - { - "id": 2784, - "data_source": "Mono3DRefer", - "scene_name": "003494", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[681.74, 239.46], [743.91, 239.23], [786.15, 256.47], [707.95, 256.83], [681.74, 180.33], [743.91, 180.3], [786.15, 182.24], [707.95, 182.28]]}, {\"category\": \"car\", \"corners_3d\": [[647.93, 210.43], [681.5, 210.45], [690.29, 215.6], [652.12, 215.58], [647.93, 178.61], [681.5, 178.62], [690.29, 179.4], [652.12, 179.4]]}, {\"category\": \"car\", \"corners_3d\": [[668.02, 226.14], [716.94, 226.11], [738.66, 236.25], [680.43, 236.3], [668.02, 177.77], [716.94, 177.77], [738.66, 178.71], [680.43, 178.71]]}, {\"category\": \"car\", \"corners_3d\": [[410.85, 219.83], [403.05, 221.82], [322.7, 221.92], [333.79, 219.93], [410.85, 190.32], [403.05, 191.06], [322.7, 191.1], [333.79, 190.36]]}, {\"category\": \"car\", \"corners_3d\": [[304.15, 222.61], [290.44, 224.79], [204.9, 224.73], [222.19, 222.55], [304.15, 191.9], [290.44, 192.73], [204.9, 192.71], [222.19, 191.87]]}, {\"category\": \"car\", \"corners_3d\": [[635.82, 199.28], [657.22, 199.26], [662.81, 201.4], [639.68, 201.43], [635.82, 178.66], [657.22, 178.66], [662.81, 179.13], [639.68, 179.13]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[898.48, 290.86], [951.73, 290.86], [978.51, 300.12], [921.08, 300.11], [898.48, 150.45], [951.73, 150.45], [978.51, 148.7], [921.08, 148.7]]}, {\"category\": \"car\", \"corners_3d\": [[595.15, 200.2], [618.13, 200.19], [619.3, 202.3], [594.55, 202.31], [595.15, 178.49], [618.13, 178.49], [619.3, 178.92], [594.55, 178.92]]}, {\"category\": \"car\", \"corners_3d\": [[347.68, 395.96], [144.82, 396.01], [325.02, 309.38], [449.11, 309.36], [347.68, 211.06], [144.82, 211.07], [325.02, 196.23], [449.11, 196.23]]}, {\"category\": \"car\", \"corners_3d\": [[908.96, 430.26], [1160.01, 429.44], [3850.68, 1668.69], [2415.3, 1697.26], [908.96, 177.42], [1160.01, 177.41], [3850.68, 199.41], [2415.3, 199.92]]}, {\"category\": \"car\", \"corners_3d\": [[787.46, 326.3], [910.23, 325.0], [1074.61, 399.39], [893.15, 402.29], [787.46, 173.72], [910.23, 173.72], [1074.61, 174.14], [893.15, 174.16]]}]\n```" - } - ] - }, - { - "id": 2785, - "data_source": "Mono3DRefer", - "scene_name": "003516", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[607.18, 222.56], [624.57, 222.53], [627.75, 226.91], [608.84, 226.95], [607.18, 175.33], [624.57, 175.33], [627.75, 175.55], [608.84, 175.55]]}]\n```" - } - ] - }, - { - "id": 2786, - "data_source": "Mono3DRefer", - "scene_name": "003516", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-138.79, 341.68], [-25.16, 314.14], [238.51, 310.66], [178.1, 336.73], [-138.79, 218.08], [-25.16, 210.71], [238.51, 209.77], [178.1, 216.76]]}, {\"category\": \"car\", \"corners_3d\": [[299.21, 242.19], [318.97, 237.3], [439.04, 236.86], [428.46, 241.68], [299.21, 192.05], [318.97, 190.69], [439.04, 190.57], [428.46, 191.91]]}]\n```" - } - ] - }, - { - "id": 2787, - "data_source": "Mono3DRefer", - "scene_name": "003516", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-138.79, 341.68], [-25.16, 314.14], [238.51, 310.66], [178.1, 336.73], [-138.79, 218.08], [-25.16, 210.71], [238.51, 209.77], [178.1, 216.76]]}, {\"category\": \"car\", \"corners_3d\": [[299.21, 242.19], [318.97, 237.3], [439.04, 236.86], [428.46, 241.68], [299.21, 192.05], [318.97, 190.69], [439.04, 190.57], [428.46, 191.91]]}, {\"category\": \"cyclist\", \"corners_3d\": [[607.18, 222.56], [624.57, 222.53], [627.75, 226.91], [608.84, 226.95], [607.18, 175.33], [624.57, 175.33], [627.75, 175.55], [608.84, 175.55]]}]\n```" - } - ] - }, - { - "id": 2788, - "data_source": "Mono3DRefer", - "scene_name": "003517", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[476.03, 257.36], [403.03, 257.2], [445.37, 240.99], [504.39, 241.1], [476.03, 190.22], [403.03, 190.19], [445.37, 186.86], [504.39, 186.88]]}, {\"category\": \"car\", \"corners_3d\": [[551.56, 204.09], [525.56, 204.15], [529.86, 201.31], [553.5, 201.26], [551.56, 182.95], [525.56, 182.97], [529.86, 182.05], [553.5, 182.03]]}, {\"category\": \"car\", \"corners_3d\": [[-353.65, 338.13], [-267.6, 314.24], [127.41, 289.87], [111.89, 305.77], [-353.65, 202.15], [-267.6, 197.92], [127.41, 193.6], [111.89, 196.42]]}, {\"category\": \"car\", \"corners_3d\": [[427.19, 210.87], [423.37, 209.61], [499.22, 208.17], [505.37, 209.33], [427.19, 182.12], [423.37, 181.81], [499.22, 181.46], [505.37, 181.74]]}]\n```" - } - ] - }, - { - "id": 2789, - "data_source": "Mono3DRefer", - "scene_name": "003517", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[568.41, 187.58], [585.33, 187.4], [609.1, 188.33], [591.4, 188.54], [568.41, 162.22], [585.33, 162.35], [609.1, 161.68], [591.4, 161.53]]}]\n```" - } - ] - }, - { - "id": 2790, - "data_source": "Mono3DRefer", - "scene_name": "003517", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[476.03, 257.36], [403.03, 257.2], [445.37, 240.99], [504.39, 241.1], [476.03, 190.22], [403.03, 190.19], [445.37, 186.86], [504.39, 186.88]]}, {\"category\": \"car\", \"corners_3d\": [[551.56, 204.09], [525.56, 204.15], [529.86, 201.31], [553.5, 201.26], [551.56, 182.95], [525.56, 182.97], [529.86, 182.05], [553.5, 182.03]]}, {\"category\": \"van\", \"corners_3d\": [[568.41, 187.58], [585.33, 187.4], [609.1, 188.33], [591.4, 188.54], [568.41, 162.22], [585.33, 162.35], [609.1, 161.68], [591.4, 161.53]]}, {\"category\": \"car\", \"corners_3d\": [[-353.65, 338.13], [-267.6, 314.24], [127.41, 289.87], [111.89, 305.77], [-353.65, 202.15], [-267.6, 197.92], [127.41, 193.6], [111.89, 196.42]]}, {\"category\": \"car\", \"corners_3d\": [[427.19, 210.87], [423.37, 209.61], [499.22, 208.17], [505.37, 209.33], [427.19, 182.12], [423.37, 181.81], [499.22, 181.46], [505.37, 181.74]]}]\n```" - } - ] - }, - { - "id": 2791, - "data_source": "Mono3DRefer", - "scene_name": "003518", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[508.71, 221.83], [520.25, 220.74], [563.11, 222.49], [552.15, 223.67], [508.71, 170.5], [520.25, 170.55], [563.11, 170.46], [552.15, 170.41]]}]\n```" - } - ] - }, - { - "id": 2792, - "data_source": "Mono3DRefer", - "scene_name": "003518", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[340.87, 271.76], [384.5, 269.58], [390.89, 277.36], [343.81, 279.91], [340.87, 158.25], [384.5, 158.57], [390.89, 157.42], [343.81, 157.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[479.69, 256.04], [458.56, 259.01], [427.56, 256.27], [449.08, 253.48], [479.69, 161.02], [458.56, 160.6], [427.56, 160.99], [449.08, 161.39]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[496.96, 250.93], [468.18, 253.38], [445.98, 249.84], [474.19, 247.6], [496.96, 155.37], [468.18, 154.82], [445.98, 155.61], [474.19, 156.11]]}]\n```" - } - ] - }, - { - "id": 2793, - "data_source": "Mono3DRefer", - "scene_name": "003518", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[161.08, 348.97], [-38.26, 348.52], [236.99, 274.64], [352.77, 274.79], [161.08, 194.27], [-38.26, 194.21], [236.99, 185.23], [352.77, 185.25]]}, {\"category\": \"car\", \"corners_3d\": [[1218.94, 202.13], [1165.89, 200.17], [1271.89, 198.89], [1329.47, 200.65], [1218.94, 149.91], [1165.89, 151.44], [1271.89, 152.45], [1329.47, 151.06]]}]\n```" - } - ] - }, - { - "id": 2794, - "data_source": "Mono3DRefer", - "scene_name": "003518", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[340.87, 271.76], [384.5, 269.58], [390.89, 277.36], [343.81, 279.91], [340.87, 158.25], [384.5, 158.57], [390.89, 157.42], [343.81, 157.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[479.69, 256.04], [458.56, 259.01], [427.56, 256.27], [449.08, 253.48], [479.69, 161.02], [458.56, 160.6], [427.56, 160.99], [449.08, 161.39]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[496.96, 250.93], [468.18, 253.38], [445.98, 249.84], [474.19, 247.6], [496.96, 155.37], [468.18, 154.82], [445.98, 155.61], [474.19, 156.11]]}, {\"category\": \"cyclist\", \"corners_3d\": [[508.71, 221.83], [520.25, 220.74], [563.11, 222.49], [552.15, 223.67], [508.71, 170.5], [520.25, 170.55], [563.11, 170.46], [552.15, 170.41]]}]\n```" - } - ] - }, - { - "id": 2795, - "data_source": "Mono3DRefer", - "scene_name": "003518", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[161.08, 348.97], [-38.26, 348.52], [236.99, 274.64], [352.77, 274.79], [161.08, 194.27], [-38.26, 194.21], [236.99, 185.23], [352.77, 185.25]]}, {\"category\": \"cyclist\", \"corners_3d\": [[508.71, 221.83], [520.25, 220.74], [563.11, 222.49], [552.15, 223.67], [508.71, 170.5], [520.25, 170.55], [563.11, 170.46], [552.15, 170.41]]}, {\"category\": \"car\", \"corners_3d\": [[1218.94, 202.13], [1165.89, 200.17], [1271.89, 198.89], [1329.47, 200.65], [1218.94, 149.91], [1165.89, 151.44], [1271.89, 152.45], [1329.47, 151.06]]}]\n```" - } - ] - }, - { - "id": 2796, - "data_source": "Mono3DRefer", - "scene_name": "003518", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[161.08, 348.97], [-38.26, 348.52], [236.99, 274.64], [352.77, 274.79], [161.08, 194.27], [-38.26, 194.21], [236.99, 185.23], [352.77, 185.25]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[340.87, 271.76], [384.5, 269.58], [390.89, 277.36], [343.81, 279.91], [340.87, 158.25], [384.5, 158.57], [390.89, 157.42], [343.81, 157.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[479.69, 256.04], [458.56, 259.01], [427.56, 256.27], [449.08, 253.48], [479.69, 161.02], [458.56, 160.6], [427.56, 160.99], [449.08, 161.39]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[496.96, 250.93], [468.18, 253.38], [445.98, 249.84], [474.19, 247.6], [496.96, 155.37], [468.18, 154.82], [445.98, 155.61], [474.19, 156.11]]}, {\"category\": \"car\", \"corners_3d\": [[1218.94, 202.13], [1165.89, 200.17], [1271.89, 198.89], [1329.47, 200.65], [1218.94, 149.91], [1165.89, 151.44], [1271.89, 152.45], [1329.47, 151.06]]}]\n```" - } - ] - }, - { - "id": 2797, - "data_source": "Mono3DRefer", - "scene_name": "003518", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[161.08, 348.97], [-38.26, 348.52], [236.99, 274.64], [352.77, 274.79], [161.08, 194.27], [-38.26, 194.21], [236.99, 185.23], [352.77, 185.25]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[340.87, 271.76], [384.5, 269.58], [390.89, 277.36], [343.81, 279.91], [340.87, 158.25], [384.5, 158.57], [390.89, 157.42], [343.81, 157.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[479.69, 256.04], [458.56, 259.01], [427.56, 256.27], [449.08, 253.48], [479.69, 161.02], [458.56, 160.6], [427.56, 160.99], [449.08, 161.39]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[496.96, 250.93], [468.18, 253.38], [445.98, 249.84], [474.19, 247.6], [496.96, 155.37], [468.18, 154.82], [445.98, 155.61], [474.19, 156.11]]}, {\"category\": \"cyclist\", \"corners_3d\": [[508.71, 221.83], [520.25, 220.74], [563.11, 222.49], [552.15, 223.67], [508.71, 170.5], [520.25, 170.55], [563.11, 170.46], [552.15, 170.41]]}, {\"category\": \"car\", \"corners_3d\": [[1218.94, 202.13], [1165.89, 200.17], [1271.89, 198.89], [1329.47, 200.65], [1218.94, 149.91], [1165.89, 151.44], [1271.89, 152.45], [1329.47, 151.06]]}]\n```" - } - ] - }, - { - "id": 2798, - "data_source": "Mono3DRefer", - "scene_name": "003524", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[397.62, 304.83], [286.57, 305.65], [368.38, 268.64], [448.25, 268.21], [397.62, 202.32], [286.57, 202.51], [368.38, 194.24], [448.25, 194.15]]}, {\"category\": \"car\", \"corners_3d\": [[553.79, 211.39], [515.59, 211.43], [524.53, 207.06], [558.4, 207.03], [553.79, 179.6], [515.59, 179.6], [524.53, 178.84], [558.4, 178.83]]}]\n```" - } - ] - }, - { - "id": 2799, - "data_source": "Mono3DRefer", - "scene_name": "003524", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[475.73, 253.64], [398.85, 253.73], [440.65, 237.07], [501.67, 237.01], [475.73, 170.07], [398.85, 170.06], [440.65, 170.64], [501.67, 170.64]]}]\n```" - } - ] - }, - { - "id": 2800, - "data_source": "Mono3DRefer", - "scene_name": "003524", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[397.62, 304.83], [286.57, 305.65], [368.38, 268.64], [448.25, 268.21], [397.62, 202.32], [286.57, 202.51], [368.38, 194.24], [448.25, 194.15]]}, {\"category\": \"van\", \"corners_3d\": [[475.73, 253.64], [398.85, 253.73], [440.65, 237.07], [501.67, 237.01], [475.73, 170.07], [398.85, 170.06], [440.65, 170.64], [501.67, 170.64]]}, {\"category\": \"car\", \"corners_3d\": [[553.79, 211.39], [515.59, 211.43], [524.53, 207.06], [558.4, 207.03], [553.79, 179.6], [515.59, 179.6], [524.53, 178.84], [558.4, 178.83]]}]\n```" - } - ] - }, - { - "id": 2801, - "data_source": "Mono3DRefer", - "scene_name": "003553", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-62.32, 264.7], [-129.79, 264.88], [2.19, 248.11], [57.18, 247.99], [-62.32, 204.07], [-129.79, 204.13], [2.19, 198.43], [57.18, 198.39]]}, {\"category\": \"car\", \"corners_3d\": [[233.39, 234.25], [187.25, 234.25], [231.97, 227.74], [273.21, 227.74], [233.39, 194.05], [187.25, 194.05], [231.97, 191.81], [273.21, 191.81]]}, {\"category\": \"car\", \"corners_3d\": [[411.65, 205.01], [389.66, 204.99], [403.86, 203.04], [424.52, 203.05], [411.65, 187.9], [389.66, 187.89], [403.86, 186.98], [424.52, 186.99]]}]\n```" - } - ] - }, - { - "id": 2802, - "data_source": "Mono3DRefer", - "scene_name": "003565", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[792.71, 374.43], [973.22, 372.94], [1452.09, 621.99], [1051.18, 629.54], [792.71, 198.02], [973.22, 197.83], [1452.09, 228.92], [1051.18, 229.86]]}, {\"category\": \"car\", \"corners_3d\": [[318.22, 328.72], [451.37, 327.61], [356.75, 440.35], [124.21, 443.68], [318.22, 193.08], [451.37, 192.94], [356.75, 207.57], [124.21, 208.0]]}, {\"category\": \"car\", \"corners_3d\": [[477.52, 303.59], [357.33, 303.83], [419.17, 270.7], [508.91, 270.57], [477.52, 197.81], [357.33, 197.85], [419.17, 191.53], [508.91, 191.5]]}, {\"category\": \"car\", \"corners_3d\": [[712.91, 272.35], [798.46, 271.4], [881.84, 304.78], [768.02, 306.48], [712.91, 184.71], [798.46, 184.59], [881.84, 188.57], [768.02, 188.77]]}, {\"category\": \"car\", \"corners_3d\": [[670.59, 244.47], [734.03, 244.35], [770.5, 263.19], [690.37, 263.39], [670.59, 185.57], [734.03, 185.55], [770.5, 188.9], [690.37, 188.93]]}, {\"category\": \"car\", \"corners_3d\": [[552.53, 231.11], [508.1, 230.97], [524.54, 223.61], [563.36, 223.72], [552.53, 186.11], [508.1, 186.08], [524.54, 184.4], [563.36, 184.43]]}, {\"category\": \"car\", \"corners_3d\": [[578.97, 219.85], [539.49, 219.78], [550.63, 214.04], [585.29, 214.1], [578.97, 174.28], [539.49, 174.27], [550.63, 174.1], [585.29, 174.1]]}, {\"category\": \"car\", \"corners_3d\": [[598.09, 213.32], [564.62, 213.19], [574.61, 208.95], [604.58, 209.05], [598.09, 176.83], [564.62, 176.82], [574.61, 176.4], [604.58, 176.41]]}]\n```" - } - ] - }, - { - "id": 2803, - "data_source": "Mono3DRefer", - "scene_name": "003569", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-403.21, 1032.82], [-1335.34, 1104.47], [83.4, 406.53], [302.45, 401.74], [-403.21, 130.15], [-1335.34, 126.59], [83.4, 161.25], [302.45, 161.49]]}, {\"category\": \"car\", \"corners_3d\": [[409.51, 349.26], [242.13, 350.11], [372.08, 284.75], [477.53, 284.41], [409.51, 169.88], [242.13, 169.86], [372.08, 170.97], [477.53, 170.97]]}, {\"category\": \"car\", \"corners_3d\": [[565.6, 231.19], [516.95, 230.34], [552.46, 222.03], [594.44, 222.65], [565.6, 183.82], [516.95, 183.66], [552.46, 182.1], [594.44, 182.22]]}, {\"category\": \"car\", \"corners_3d\": [[602.87, 219.95], [569.69, 219.28], [596.82, 214.48], [626.88, 215.02], [602.87, 181.43], [569.69, 181.31], [596.82, 180.44], [626.88, 180.53]]}, {\"category\": \"car\", \"corners_3d\": [[650.17, 213.26], [620.5, 212.69], [641.33, 209.28], [668.71, 209.76], [650.17, 182.8], [620.5, 182.65], [641.33, 181.82], [668.71, 181.93]]}, {\"category\": \"car\", \"corners_3d\": [[807.5, 186.3], [827.08, 186.5], [809.72, 187.39], [789.17, 187.16], [807.5, 166.46], [827.08, 166.36], [809.72, 165.94], [789.17, 166.05]]}]\n```" - } - ] - }, - { - "id": 2804, - "data_source": "Mono3DRefer", - "scene_name": "003573", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[847.2, 145.6], [819.87, 146.06], [832.57, 148.13], [857.96, 147.74], [847.2, 120.84], [819.87, 121.71], [832.57, 125.66], [857.96, 124.92]]}]\n```" - } - ] - }, - { - "id": 2805, - "data_source": "Mono3DRefer", - "scene_name": "003573", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (motorcyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"motorcyclist\", \"corners_3d\": [[598.65, 290.81], [543.25, 288.1], [584.69, 271.11], [632.58, 273.07], [598.65, 148.21], [543.25, 148.77], [584.69, 152.32], [632.58, 151.91]]}]\n```" - } - ] - }, - { - "id": 2806, - "data_source": "Mono3DRefer", - "scene_name": "003573", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, motorcyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"motorcyclist\", \"corners_3d\": [[598.65, 290.81], [543.25, 288.1], [584.69, 271.11], [632.58, 273.07], [598.65, 148.21], [543.25, 148.77], [584.69, 152.32], [632.58, 151.91]]}, {\"category\": \"car\", \"corners_3d\": [[847.2, 145.6], [819.87, 146.06], [832.57, 148.13], [857.96, 147.74], [847.2, 120.84], [819.87, 121.71], [832.57, 125.66], [857.96, 124.92]]}]\n```" - } - ] - }, - { - "id": 2807, - "data_source": "Mono3DRefer", - "scene_name": "003575", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[406.91, 249.18], [457.88, 254.27], [286.95, 274.44], [237.43, 266.64], [406.91, 183.54], [457.88, 184.26], [286.95, 187.08], [237.43, 185.99]]}, {\"category\": \"car\", \"corners_3d\": [[330.87, 243.22], [383.81, 247.89], [181.57, 269.27], [130.63, 261.69], [330.87, 170.37], [383.81, 170.2], [181.57, 169.45], [130.63, 169.72]]}]\n```" - } - ] - }, - { - "id": 2808, - "data_source": "Mono3DRefer", - "scene_name": "003577", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[841.84, 204.54], [811.42, 203.2], [858.9, 201.54], [889.58, 202.74], [841.84, 163.94], [811.42, 164.32], [858.9, 164.78], [889.58, 164.44]]}, {\"category\": \"car\", \"corners_3d\": [[687.51, 212.42], [728.62, 212.44], [745.71, 218.52], [698.28, 218.49], [687.51, 173.82], [728.62, 173.82], [745.71, 173.97], [698.28, 173.97]]}, {\"category\": \"car\", \"corners_3d\": [[567.8, 210.72], [531.93, 210.47], [551.57, 206.16], [583.42, 206.36], [567.8, 173.08], [531.93, 173.08], [551.57, 173.06], [583.42, 173.06]]}, {\"category\": \"car\", \"corners_3d\": [[737.01, 189.81], [734.36, 189.22], [791.3, 189.31], [796.04, 189.91], [737.01, 162.95], [734.36, 163.3], [791.3, 163.25], [796.04, 162.9]]}, {\"category\": \"car\", \"corners_3d\": [[316.84, 423.77], [99.53, 412.55], [370.08, 306.43], [495.64, 309.85], [316.84, 195.08], [99.53, 194.08], [370.08, 184.68], [495.64, 184.99]]}, {\"category\": \"car\", \"corners_3d\": [[804.43, 316.03], [953.23, 316.93], [1186.08, 421.3], [928.15, 418.63], [804.43, 177.22], [953.23, 177.24], [1186.08, 180.42], [928.15, 180.34]]}, {\"category\": \"car\", \"corners_3d\": [[1187.1, 222.42], [1249.23, 220.83], [1476.93, 233.73], [1407.03, 236.31], [1187.1, 140.23], [1249.23, 141.28], [1476.93, 132.78], [1407.03, 131.09]]}, {\"category\": \"car\", \"corners_3d\": [[479.67, 257.49], [399.6, 256.93], [449.94, 240.19], [514.25, 240.55], [479.67, 180.95], [399.6, 180.89], [449.94, 179.29], [514.25, 179.33]]}, {\"category\": \"car\", \"corners_3d\": [[519.91, 236.64], [460.68, 236.12], [499.98, 224.85], [548.85, 225.21], [519.91, 179.68], [460.68, 179.62], [499.98, 178.42], [548.85, 178.45]]}, {\"category\": \"car\", \"corners_3d\": [[956.71, 213.62], [909.98, 211.45], [964.21, 208.4], [1009.85, 210.23], [956.71, 160.39], [909.98, 161.06], [964.21, 161.99], [1009.85, 161.43]]}, {\"category\": \"car\", \"corners_3d\": [[881.04, 209.01], [842.12, 207.23], [895.1, 204.88], [933.78, 206.42], [881.04, 164.68], [842.12, 165.08], [895.1, 165.61], [933.78, 165.26]]}]\n```" - } - ] - }, - { - "id": 2809, - "data_source": "Mono3DRefer", - "scene_name": "003577", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[652.83, 188.75], [685.1, 188.85], [678.32, 192.03], [639.75, 191.88], [652.83, 149.62], [685.1, 149.46], [678.32, 144.82], [639.75, 145.04]]}]\n```" - } - ] - }, - { - "id": 2810, - "data_source": "Mono3DRefer", - "scene_name": "003577", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[586.0, 202.74], [550.5, 202.61], [563.84, 199.39], [595.55, 199.5], [586.0, 163.02], [550.5, 163.06], [563.84, 164.12], [595.55, 164.08]]}, {\"category\": \"van\", \"corners_3d\": [[1053.04, 231.03], [1113.7, 229.63], [1303.19, 245.45], [1231.12, 247.76], [1053.04, 141.77], [1113.7, 142.52], [1303.19, 134.07], [1231.12, 132.84]]}, {\"category\": \"van\", \"corners_3d\": [[922.43, 222.87], [988.45, 222.03], [1140.22, 236.89], [1057.21, 238.33], [922.43, 144.67], [988.45, 145.14], [1140.22, 136.77], [1057.21, 135.96]]}]\n```" - } - ] - }, - { - "id": 2811, - "data_source": "Mono3DRefer", - "scene_name": "003577", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[841.84, 204.54], [811.42, 203.2], [858.9, 201.54], [889.58, 202.74], [841.84, 163.94], [811.42, 164.32], [858.9, 164.78], [889.58, 164.44]]}, {\"category\": \"car\", \"corners_3d\": [[687.51, 212.42], [728.62, 212.44], [745.71, 218.52], [698.28, 218.49], [687.51, 173.82], [728.62, 173.82], [745.71, 173.97], [698.28, 173.97]]}, {\"category\": \"car\", \"corners_3d\": [[567.8, 210.72], [531.93, 210.47], [551.57, 206.16], [583.42, 206.36], [567.8, 173.08], [531.93, 173.08], [551.57, 173.06], [583.42, 173.06]]}, {\"category\": \"car\", \"corners_3d\": [[737.01, 189.81], [734.36, 189.22], [791.3, 189.31], [796.04, 189.91], [737.01, 162.95], [734.36, 163.3], [791.3, 163.25], [796.04, 162.9]]}, {\"category\": \"truck\", \"corners_3d\": [[652.83, 188.75], [685.1, 188.85], [678.32, 192.03], [639.75, 191.88], [652.83, 149.62], [685.1, 149.46], [678.32, 144.82], [639.75, 145.04]]}, {\"category\": \"car\", \"corners_3d\": [[316.84, 423.77], [99.53, 412.55], [370.08, 306.43], [495.64, 309.85], [316.84, 195.08], [99.53, 194.08], [370.08, 184.68], [495.64, 184.99]]}, {\"category\": \"car\", \"corners_3d\": [[804.43, 316.03], [953.23, 316.93], [1186.08, 421.3], [928.15, 418.63], [804.43, 177.22], [953.23, 177.24], [1186.08, 180.42], [928.15, 180.34]]}, {\"category\": \"car\", \"corners_3d\": [[1187.1, 222.42], [1249.23, 220.83], [1476.93, 233.73], [1407.03, 236.31], [1187.1, 140.23], [1249.23, 141.28], [1476.93, 132.78], [1407.03, 131.09]]}, {\"category\": \"car\", \"corners_3d\": [[479.67, 257.49], [399.6, 256.93], [449.94, 240.19], [514.25, 240.55], [479.67, 180.95], [399.6, 180.89], [449.94, 179.29], [514.25, 179.33]]}, {\"category\": \"car\", \"corners_3d\": [[519.91, 236.64], [460.68, 236.12], [499.98, 224.85], [548.85, 225.21], [519.91, 179.68], [460.68, 179.62], [499.98, 178.42], [548.85, 178.45]]}, {\"category\": \"car\", \"corners_3d\": [[956.71, 213.62], [909.98, 211.45], [964.21, 208.4], [1009.85, 210.23], [956.71, 160.39], [909.98, 161.06], [964.21, 161.99], [1009.85, 161.43]]}, {\"category\": \"car\", \"corners_3d\": [[881.04, 209.01], [842.12, 207.23], [895.1, 204.88], [933.78, 206.42], [881.04, 164.68], [842.12, 165.08], [895.1, 165.61], [933.78, 165.26]]}]\n```" - } - ] - }, - { - "id": 2812, - "data_source": "Mono3DRefer", - "scene_name": "003577", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[841.84, 204.54], [811.42, 203.2], [858.9, 201.54], [889.58, 202.74], [841.84, 163.94], [811.42, 164.32], [858.9, 164.78], [889.58, 164.44]]}, {\"category\": \"car\", \"corners_3d\": [[687.51, 212.42], [728.62, 212.44], [745.71, 218.52], [698.28, 218.49], [687.51, 173.82], [728.62, 173.82], [745.71, 173.97], [698.28, 173.97]]}, {\"category\": \"car\", \"corners_3d\": [[567.8, 210.72], [531.93, 210.47], [551.57, 206.16], [583.42, 206.36], [567.8, 173.08], [531.93, 173.08], [551.57, 173.06], [583.42, 173.06]]}, {\"category\": \"van\", \"corners_3d\": [[586.0, 202.74], [550.5, 202.61], [563.84, 199.39], [595.55, 199.5], [586.0, 163.02], [550.5, 163.06], [563.84, 164.12], [595.55, 164.08]]}, {\"category\": \"car\", \"corners_3d\": [[737.01, 189.81], [734.36, 189.22], [791.3, 189.31], [796.04, 189.91], [737.01, 162.95], [734.36, 163.3], [791.3, 163.25], [796.04, 162.9]]}, {\"category\": \"car\", \"corners_3d\": [[316.84, 423.77], [99.53, 412.55], [370.08, 306.43], [495.64, 309.85], [316.84, 195.08], [99.53, 194.08], [370.08, 184.68], [495.64, 184.99]]}, {\"category\": \"car\", \"corners_3d\": [[804.43, 316.03], [953.23, 316.93], [1186.08, 421.3], [928.15, 418.63], [804.43, 177.22], [953.23, 177.24], [1186.08, 180.42], [928.15, 180.34]]}, {\"category\": \"car\", \"corners_3d\": [[1187.1, 222.42], [1249.23, 220.83], [1476.93, 233.73], [1407.03, 236.31], [1187.1, 140.23], [1249.23, 141.28], [1476.93, 132.78], [1407.03, 131.09]]}, {\"category\": \"van\", \"corners_3d\": [[1053.04, 231.03], [1113.7, 229.63], [1303.19, 245.45], [1231.12, 247.76], [1053.04, 141.77], [1113.7, 142.52], [1303.19, 134.07], [1231.12, 132.84]]}, {\"category\": \"van\", \"corners_3d\": [[922.43, 222.87], [988.45, 222.03], [1140.22, 236.89], [1057.21, 238.33], [922.43, 144.67], [988.45, 145.14], [1140.22, 136.77], [1057.21, 135.96]]}, {\"category\": \"car\", \"corners_3d\": [[479.67, 257.49], [399.6, 256.93], [449.94, 240.19], [514.25, 240.55], [479.67, 180.95], [399.6, 180.89], [449.94, 179.29], [514.25, 179.33]]}, {\"category\": \"car\", \"corners_3d\": [[519.91, 236.64], [460.68, 236.12], [499.98, 224.85], [548.85, 225.21], [519.91, 179.68], [460.68, 179.62], [499.98, 178.42], [548.85, 178.45]]}, {\"category\": \"car\", \"corners_3d\": [[956.71, 213.62], [909.98, 211.45], [964.21, 208.4], [1009.85, 210.23], [956.71, 160.39], [909.98, 161.06], [964.21, 161.99], [1009.85, 161.43]]}, {\"category\": \"car\", \"corners_3d\": [[881.04, 209.01], [842.12, 207.23], [895.1, 204.88], [933.78, 206.42], [881.04, 164.68], [842.12, 165.08], [895.1, 165.61], [933.78, 165.26]]}]\n```" - } - ] - }, - { - "id": 2813, - "data_source": "Mono3DRefer", - "scene_name": "003577", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[586.0, 202.74], [550.5, 202.61], [563.84, 199.39], [595.55, 199.5], [586.0, 163.02], [550.5, 163.06], [563.84, 164.12], [595.55, 164.08]]}, {\"category\": \"truck\", \"corners_3d\": [[652.83, 188.75], [685.1, 188.85], [678.32, 192.03], [639.75, 191.88], [652.83, 149.62], [685.1, 149.46], [678.32, 144.82], [639.75, 145.04]]}, {\"category\": \"van\", \"corners_3d\": [[1053.04, 231.03], [1113.7, 229.63], [1303.19, 245.45], [1231.12, 247.76], [1053.04, 141.77], [1113.7, 142.52], [1303.19, 134.07], [1231.12, 132.84]]}, {\"category\": \"van\", \"corners_3d\": [[922.43, 222.87], [988.45, 222.03], [1140.22, 236.89], [1057.21, 238.33], [922.43, 144.67], [988.45, 145.14], [1140.22, 136.77], [1057.21, 135.96]]}]\n```" - } - ] - }, - { - "id": 2814, - "data_source": "Mono3DRefer", - "scene_name": "003577", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[841.84, 204.54], [811.42, 203.2], [858.9, 201.54], [889.58, 202.74], [841.84, 163.94], [811.42, 164.32], [858.9, 164.78], [889.58, 164.44]]}, {\"category\": \"car\", \"corners_3d\": [[687.51, 212.42], [728.62, 212.44], [745.71, 218.52], [698.28, 218.49], [687.51, 173.82], [728.62, 173.82], [745.71, 173.97], [698.28, 173.97]]}, {\"category\": \"car\", \"corners_3d\": [[567.8, 210.72], [531.93, 210.47], [551.57, 206.16], [583.42, 206.36], [567.8, 173.08], [531.93, 173.08], [551.57, 173.06], [583.42, 173.06]]}, {\"category\": \"van\", \"corners_3d\": [[586.0, 202.74], [550.5, 202.61], [563.84, 199.39], [595.55, 199.5], [586.0, 163.02], [550.5, 163.06], [563.84, 164.12], [595.55, 164.08]]}, {\"category\": \"car\", \"corners_3d\": [[737.01, 189.81], [734.36, 189.22], [791.3, 189.31], [796.04, 189.91], [737.01, 162.95], [734.36, 163.3], [791.3, 163.25], [796.04, 162.9]]}, {\"category\": \"truck\", \"corners_3d\": [[652.83, 188.75], [685.1, 188.85], [678.32, 192.03], [639.75, 191.88], [652.83, 149.62], [685.1, 149.46], [678.32, 144.82], [639.75, 145.04]]}, {\"category\": \"car\", \"corners_3d\": [[316.84, 423.77], [99.53, 412.55], [370.08, 306.43], [495.64, 309.85], [316.84, 195.08], [99.53, 194.08], [370.08, 184.68], [495.64, 184.99]]}, {\"category\": \"car\", \"corners_3d\": [[804.43, 316.03], [953.23, 316.93], [1186.08, 421.3], [928.15, 418.63], [804.43, 177.22], [953.23, 177.24], [1186.08, 180.42], [928.15, 180.34]]}, {\"category\": \"car\", \"corners_3d\": [[1187.1, 222.42], [1249.23, 220.83], [1476.93, 233.73], [1407.03, 236.31], [1187.1, 140.23], [1249.23, 141.28], [1476.93, 132.78], [1407.03, 131.09]]}, {\"category\": \"van\", \"corners_3d\": [[1053.04, 231.03], [1113.7, 229.63], [1303.19, 245.45], [1231.12, 247.76], [1053.04, 141.77], [1113.7, 142.52], [1303.19, 134.07], [1231.12, 132.84]]}, {\"category\": \"van\", \"corners_3d\": [[922.43, 222.87], [988.45, 222.03], [1140.22, 236.89], [1057.21, 238.33], [922.43, 144.67], [988.45, 145.14], [1140.22, 136.77], [1057.21, 135.96]]}, {\"category\": \"car\", \"corners_3d\": [[479.67, 257.49], [399.6, 256.93], [449.94, 240.19], [514.25, 240.55], [479.67, 180.95], [399.6, 180.89], [449.94, 179.29], [514.25, 179.33]]}, {\"category\": \"car\", \"corners_3d\": [[519.91, 236.64], [460.68, 236.12], [499.98, 224.85], [548.85, 225.21], [519.91, 179.68], [460.68, 179.62], [499.98, 178.42], [548.85, 178.45]]}, {\"category\": \"car\", \"corners_3d\": [[956.71, 213.62], [909.98, 211.45], [964.21, 208.4], [1009.85, 210.23], [956.71, 160.39], [909.98, 161.06], [964.21, 161.99], [1009.85, 161.43]]}, {\"category\": \"car\", \"corners_3d\": [[881.04, 209.01], [842.12, 207.23], [895.1, 204.88], [933.78, 206.42], [881.04, 164.68], [842.12, 165.08], [895.1, 165.61], [933.78, 165.26]]}]\n```" - } - ] - }, - { - "id": 2815, - "data_source": "Mono3DRefer", - "scene_name": "003578", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[513.53, 228.04], [475.1, 228.19], [486.94, 221.49], [520.7, 221.38], [513.53, 189.18], [475.1, 189.23], [486.94, 187.24], [520.7, 187.21]]}, {\"category\": \"car\", \"corners_3d\": [[512.22, 242.8], [454.17, 243.52], [462.2, 234.14], [512.54, 233.59], [512.22, 189.11], [454.17, 189.28], [462.2, 187.1], [512.54, 186.97]]}]\n```" - } - ] - }, - { - "id": 2816, - "data_source": "Mono3DRefer", - "scene_name": "003584", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[504.23, 277.58], [412.54, 277.85], [455.21, 253.24], [525.36, 253.07], [504.23, 201.5], [412.54, 201.57], [455.21, 194.84], [525.36, 194.79]]}]\n```" - } - ] - }, - { - "id": 2817, - "data_source": "Mono3DRefer", - "scene_name": "003584", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[599.2, 234.27], [621.94, 234.18], [625.9, 239.24], [601.29, 239.35], [599.2, 208.59], [621.94, 208.53], [625.9, 211.48], [601.29, 211.54]]}]\n```" - } - ] - }, - { - "id": 2818, - "data_source": "Mono3DRefer", - "scene_name": "003584", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[599.2, 234.27], [621.94, 234.18], [625.9, 239.24], [601.29, 239.35], [599.2, 208.59], [621.94, 208.53], [625.9, 211.48], [601.29, 211.54]]}, {\"category\": \"car\", \"corners_3d\": [[504.23, 277.58], [412.54, 277.85], [455.21, 253.24], [525.36, 253.07], [504.23, 201.5], [412.54, 201.57], [455.21, 194.84], [525.36, 194.79]]}]\n```" - } - ] - }, - { - "id": 2819, - "data_source": "Mono3DRefer", - "scene_name": "003585", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[869.92, 225.14], [877.95, 225.11], [895.33, 228.03], [886.86, 228.06], [869.92, 185.44], [877.95, 185.43], [895.33, 186.13], [886.86, 186.14]]}]\n```" - } - ] - }, - { - "id": 2820, - "data_source": "Mono3DRefer", - "scene_name": "003585", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[551.83, 202.21], [532.0, 202.22], [535.93, 200.57], [554.64, 200.56], [551.83, 181.85], [532.0, 181.85], [535.93, 181.35], [554.64, 181.34]]}, {\"category\": \"car\", \"corners_3d\": [[256.2, 208.24], [231.57, 208.24], [264.69, 205.13], [287.16, 205.13], [256.2, 179.0], [231.57, 179.0], [264.69, 178.46], [287.16, 178.46]]}, {\"category\": \"car\", \"corners_3d\": [[301.44, 204.83], [276.94, 204.81], [302.46, 202.45], [325.17, 202.47], [301.44, 182.18], [276.94, 182.17], [302.46, 181.48], [325.17, 181.49]]}, {\"category\": \"car\", \"corners_3d\": [[1003.97, 226.96], [988.86, 225.07], [1063.75, 224.78], [1081.47, 226.65], [1003.97, 195.28], [988.86, 194.5], [1063.75, 194.38], [1081.47, 195.16]]}, {\"category\": \"car\", \"corners_3d\": [[337.74, 203.07], [318.07, 203.07], [338.97, 200.95], [357.27, 200.96], [337.74, 182.52], [318.07, 182.52], [338.97, 181.84], [357.27, 181.84]]}, {\"category\": \"car\", \"corners_3d\": [[565.9, 196.62], [548.57, 196.62], [553.05, 194.86], [569.09, 194.86], [565.9, 180.11], [548.57, 180.11], [553.05, 179.57], [569.09, 179.57]]}, {\"category\": \"car\", \"corners_3d\": [[572.83, 221.7], [616.11, 221.7], [617.29, 231.4], [565.42, 231.4], [572.83, 178.69], [616.11, 178.69], [617.29, 179.84], [565.42, 179.84]]}, {\"category\": \"car\", \"corners_3d\": [[188.93, 212.65], [163.48, 212.56], [214.19, 208.44], [237.1, 208.51], [188.93, 183.15], [163.48, 183.13], [214.19, 182.06], [237.1, 182.08]]}, {\"category\": \"car\", \"corners_3d\": [[1128.27, 224.03], [1154.74, 226.38], [1102.73, 226.71], [1078.41, 224.33], [1128.27, 196.32], [1154.74, 197.4], [1102.73, 197.55], [1078.41, 196.46]]}]\n```" - } - ] - }, - { - "id": 2821, - "data_source": "Mono3DRefer", - "scene_name": "003585", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[551.83, 202.21], [532.0, 202.22], [535.93, 200.57], [554.64, 200.56], [551.83, 181.85], [532.0, 181.85], [535.93, 181.35], [554.64, 181.34]]}, {\"category\": \"car\", \"corners_3d\": [[256.2, 208.24], [231.57, 208.24], [264.69, 205.13], [287.16, 205.13], [256.2, 179.0], [231.57, 179.0], [264.69, 178.46], [287.16, 178.46]]}, {\"category\": \"car\", \"corners_3d\": [[301.44, 204.83], [276.94, 204.81], [302.46, 202.45], [325.17, 202.47], [301.44, 182.18], [276.94, 182.17], [302.46, 181.48], [325.17, 181.49]]}, {\"category\": \"car\", \"corners_3d\": [[1003.97, 226.96], [988.86, 225.07], [1063.75, 224.78], [1081.47, 226.65], [1003.97, 195.28], [988.86, 194.5], [1063.75, 194.38], [1081.47, 195.16]]}, {\"category\": \"car\", \"corners_3d\": [[337.74, 203.07], [318.07, 203.07], [338.97, 200.95], [357.27, 200.96], [337.74, 182.52], [318.07, 182.52], [338.97, 181.84], [357.27, 181.84]]}, {\"category\": \"car\", \"corners_3d\": [[565.9, 196.62], [548.57, 196.62], [553.05, 194.86], [569.09, 194.86], [565.9, 180.11], [548.57, 180.11], [553.05, 179.57], [569.09, 179.57]]}, {\"category\": \"car\", \"corners_3d\": [[572.83, 221.7], [616.11, 221.7], [617.29, 231.4], [565.42, 231.4], [572.83, 178.69], [616.11, 178.69], [617.29, 179.84], [565.42, 179.84]]}, {\"category\": \"car\", \"corners_3d\": [[188.93, 212.65], [163.48, 212.56], [214.19, 208.44], [237.1, 208.51], [188.93, 183.15], [163.48, 183.13], [214.19, 182.06], [237.1, 182.08]]}, {\"category\": \"cyclist\", \"corners_3d\": [[869.92, 225.14], [877.95, 225.11], [895.33, 228.03], [886.86, 228.06], [869.92, 185.44], [877.95, 185.43], [895.33, 186.13], [886.86, 186.14]]}, {\"category\": \"car\", \"corners_3d\": [[1128.27, 224.03], [1154.74, 226.38], [1102.73, 226.71], [1078.41, 224.33], [1128.27, 196.32], [1154.74, 197.4], [1102.73, 197.55], [1078.41, 196.46]]}]\n```" - } - ] - }, - { - "id": 2822, - "data_source": "Mono3DRefer", - "scene_name": "003610", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[516.73, 244.73], [455.03, 245.58], [462.34, 235.41], [515.42, 234.77], [516.73, 187.7], [455.03, 187.88], [462.34, 185.78], [515.42, 185.65]]}, {\"category\": \"car\", \"corners_3d\": [[516.97, 230.7], [476.81, 230.8], [490.76, 223.49], [525.84, 223.41], [516.97, 189.97], [476.81, 190.0], [490.76, 187.83], [525.84, 187.81]]}]\n```" - } - ] - }, - { - "id": 2823, - "data_source": "Mono3DRefer", - "scene_name": "003623", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[559.76, 241.53], [519.05, 241.61], [528.65, 232.86], [564.18, 232.8], [559.76, 203.93], [519.05, 203.96], [528.65, 200.01], [564.18, 199.98]]}]\n```" - } - ] - }, - { - "id": 2824, - "data_source": "Mono3DRefer", - "scene_name": "003630", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[390.74, 213.09], [360.87, 213.04], [382.34, 209.84], [409.85, 209.89], [390.74, 186.97], [360.87, 186.95], [382.34, 185.83], [409.85, 185.84]]}]\n```" - } - ] - }, - { - "id": 2825, - "data_source": "Mono3DRefer", - "scene_name": "003630", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[657.33, 209.17], [715.24, 209.17], [778.71, 231.11], [685.83, 231.1], [657.33, 146.43], [715.24, 146.43], [778.71, 130.46], [685.83, 130.47]]}]\n```" - } - ] - }, - { - "id": 2826, - "data_source": "Mono3DRefer", - "scene_name": "003630", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[657.33, 209.17], [715.24, 209.17], [778.71, 231.11], [685.83, 231.1], [657.33, 146.43], [715.24, 146.43], [778.71, 130.46], [685.83, 130.47]]}, {\"category\": \"car\", \"corners_3d\": [[390.74, 213.09], [360.87, 213.04], [382.34, 209.84], [409.85, 209.89], [390.74, 186.97], [360.87, 186.95], [382.34, 185.83], [409.85, 185.84]]}]\n```" - } - ] - }, - { - "id": 2827, - "data_source": "Mono3DRefer", - "scene_name": "003631", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[271.71, 222.0], [237.81, 222.88], [241.18, 218.69], [272.23, 217.95], [271.71, 195.0], [237.81, 195.4], [241.18, 193.51], [272.23, 193.18]]}]\n```" - } - ] - }, - { - "id": 2828, - "data_source": "Mono3DRefer", - "scene_name": "003632", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[832.39, 329.67], [999.76, 328.65], [1440.26, 482.85], [1101.28, 487.15], [832.39, 146.53], [999.76, 146.77], [1440.26, 110.36], [1101.28, 109.34]]}]\n```" - } - ] - }, - { - "id": 2829, - "data_source": "Mono3DRefer", - "scene_name": "003632", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[454.24, 272.36], [418.08, 272.32], [427.45, 267.81], [461.83, 267.85], [454.24, 189.05], [418.08, 189.05], [427.45, 188.67], [461.83, 188.68]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[402.52, 303.03], [459.28, 303.04], [446.17, 314.28], [384.17, 314.28], [402.52, 195.99], [459.28, 195.99], [446.17, 197.33], [384.17, 197.33]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[350.23, 301.21], [405.27, 301.22], [386.94, 312.43], [326.75, 312.42], [350.23, 195.96], [405.27, 195.97], [386.94, 197.32], [326.75, 197.32]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[417.98, 297.81], [372.33, 297.59], [388.5, 290.05], [431.21, 290.24], [417.98, 197.02], [372.33, 196.99], [388.5, 195.99], [431.21, 196.01]]}]\n```" - } - ] - }, - { - "id": 2830, - "data_source": "Mono3DRefer", - "scene_name": "003632", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[200.56, 310.02], [256.56, 308.2], [243.08, 316.8], [183.01, 318.87], [200.56, 161.85], [256.56, 162.13], [243.08, 160.82], [183.01, 160.5]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[235.05, 302.44], [191.5, 302.22], [214.31, 295.82], [255.59, 296.01], [235.05, 158.45], [191.5, 158.49], [214.31, 159.71], [255.59, 159.68]]}]\n```" - } - ] - }, - { - "id": 2831, - "data_source": "Mono3DRefer", - "scene_name": "003632", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[544.55, 193.85], [556.88, 194.51], [358.8, 195.91], [356.27, 195.1], [544.55, 142.17], [556.88, 140.07], [358.8, 135.61], [356.27, 138.17]]}]\n```" - } - ] - }, - { - "id": 2832, - "data_source": "Mono3DRefer", - "scene_name": "003632", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[454.24, 272.36], [418.08, 272.32], [427.45, 267.81], [461.83, 267.85], [454.24, 189.05], [418.08, 189.05], [427.45, 188.67], [461.83, 188.68]]}, {\"category\": \"van\", \"corners_3d\": [[832.39, 329.67], [999.76, 328.65], [1440.26, 482.85], [1101.28, 487.15], [832.39, 146.53], [999.76, 146.77], [1440.26, 110.36], [1101.28, 109.34]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[402.52, 303.03], [459.28, 303.04], [446.17, 314.28], [384.17, 314.28], [402.52, 195.99], [459.28, 195.99], [446.17, 197.33], [384.17, 197.33]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[350.23, 301.21], [405.27, 301.22], [386.94, 312.43], [326.75, 312.42], [350.23, 195.96], [405.27, 195.97], [386.94, 197.32], [326.75, 197.32]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[417.98, 297.81], [372.33, 297.59], [388.5, 290.05], [431.21, 290.24], [417.98, 197.02], [372.33, 196.99], [388.5, 195.99], [431.21, 196.01]]}]\n```" - } - ] - }, - { - "id": 2833, - "data_source": "Mono3DRefer", - "scene_name": "003632", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[200.56, 310.02], [256.56, 308.2], [243.08, 316.8], [183.01, 318.87], [200.56, 161.85], [256.56, 162.13], [243.08, 160.82], [183.01, 160.5]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[235.05, 302.44], [191.5, 302.22], [214.31, 295.82], [255.59, 296.01], [235.05, 158.45], [191.5, 158.49], [214.31, 159.71], [255.59, 159.68]]}, {\"category\": \"van\", \"corners_3d\": [[832.39, 329.67], [999.76, 328.65], [1440.26, 482.85], [1101.28, 487.15], [832.39, 146.53], [999.76, 146.77], [1440.26, 110.36], [1101.28, 109.34]]}]\n```" - } - ] - }, - { - "id": 2834, - "data_source": "Mono3DRefer", - "scene_name": "003632", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[832.39, 329.67], [999.76, 328.65], [1440.26, 482.85], [1101.28, 487.15], [832.39, 146.53], [999.76, 146.77], [1440.26, 110.36], [1101.28, 109.34]]}, {\"category\": \"tram\", \"corners_3d\": [[544.55, 193.85], [556.88, 194.51], [358.8, 195.91], [356.27, 195.1], [544.55, 142.17], [556.88, 140.07], [358.8, 135.61], [356.27, 138.17]]}]\n```" - } - ] - }, - { - "id": 2835, - "data_source": "Mono3DRefer", - "scene_name": "003632", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[454.24, 272.36], [418.08, 272.32], [427.45, 267.81], [461.83, 267.85], [454.24, 189.05], [418.08, 189.05], [427.45, 188.67], [461.83, 188.68]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[200.56, 310.02], [256.56, 308.2], [243.08, 316.8], [183.01, 318.87], [200.56, 161.85], [256.56, 162.13], [243.08, 160.82], [183.01, 160.5]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[235.05, 302.44], [191.5, 302.22], [214.31, 295.82], [255.59, 296.01], [235.05, 158.45], [191.5, 158.49], [214.31, 159.71], [255.59, 159.68]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[402.52, 303.03], [459.28, 303.04], [446.17, 314.28], [384.17, 314.28], [402.52, 195.99], [459.28, 195.99], [446.17, 197.33], [384.17, 197.33]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[350.23, 301.21], [405.27, 301.22], [386.94, 312.43], [326.75, 312.42], [350.23, 195.96], [405.27, 195.97], [386.94, 197.32], [326.75, 197.32]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[417.98, 297.81], [372.33, 297.59], [388.5, 290.05], [431.21, 290.24], [417.98, 197.02], [372.33, 196.99], [388.5, 195.99], [431.21, 196.01]]}]\n```" - } - ] - }, - { - "id": 2836, - "data_source": "Mono3DRefer", - "scene_name": "003632", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[454.24, 272.36], [418.08, 272.32], [427.45, 267.81], [461.83, 267.85], [454.24, 189.05], [418.08, 189.05], [427.45, 188.67], [461.83, 188.68]]}, {\"category\": \"tram\", \"corners_3d\": [[544.55, 193.85], [556.88, 194.51], [358.8, 195.91], [356.27, 195.1], [544.55, 142.17], [556.88, 140.07], [358.8, 135.61], [356.27, 138.17]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[402.52, 303.03], [459.28, 303.04], [446.17, 314.28], [384.17, 314.28], [402.52, 195.99], [459.28, 195.99], [446.17, 197.33], [384.17, 197.33]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[350.23, 301.21], [405.27, 301.22], [386.94, 312.43], [326.75, 312.42], [350.23, 195.96], [405.27, 195.97], [386.94, 197.32], [326.75, 197.32]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[417.98, 297.81], [372.33, 297.59], [388.5, 290.05], [431.21, 290.24], [417.98, 197.02], [372.33, 196.99], [388.5, 195.99], [431.21, 196.01]]}]\n```" - } - ] - }, - { - "id": 2837, - "data_source": "Mono3DRefer", - "scene_name": "003632", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[200.56, 310.02], [256.56, 308.2], [243.08, 316.8], [183.01, 318.87], [200.56, 161.85], [256.56, 162.13], [243.08, 160.82], [183.01, 160.5]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[235.05, 302.44], [191.5, 302.22], [214.31, 295.82], [255.59, 296.01], [235.05, 158.45], [191.5, 158.49], [214.31, 159.71], [255.59, 159.68]]}, {\"category\": \"tram\", \"corners_3d\": [[544.55, 193.85], [556.88, 194.51], [358.8, 195.91], [356.27, 195.1], [544.55, 142.17], [556.88, 140.07], [358.8, 135.61], [356.27, 138.17]]}]\n```" - } - ] - }, - { - "id": 2838, - "data_source": "Mono3DRefer", - "scene_name": "003632", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, person_sitting, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[454.24, 272.36], [418.08, 272.32], [427.45, 267.81], [461.83, 267.85], [454.24, 189.05], [418.08, 189.05], [427.45, 188.67], [461.83, 188.68]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[200.56, 310.02], [256.56, 308.2], [243.08, 316.8], [183.01, 318.87], [200.56, 161.85], [256.56, 162.13], [243.08, 160.82], [183.01, 160.5]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[235.05, 302.44], [191.5, 302.22], [214.31, 295.82], [255.59, 296.01], [235.05, 158.45], [191.5, 158.49], [214.31, 159.71], [255.59, 159.68]]}, {\"category\": \"van\", \"corners_3d\": [[832.39, 329.67], [999.76, 328.65], [1440.26, 482.85], [1101.28, 487.15], [832.39, 146.53], [999.76, 146.77], [1440.26, 110.36], [1101.28, 109.34]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[402.52, 303.03], [459.28, 303.04], [446.17, 314.28], [384.17, 314.28], [402.52, 195.99], [459.28, 195.99], [446.17, 197.33], [384.17, 197.33]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[350.23, 301.21], [405.27, 301.22], [386.94, 312.43], [326.75, 312.42], [350.23, 195.96], [405.27, 195.97], [386.94, 197.32], [326.75, 197.32]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[417.98, 297.81], [372.33, 297.59], [388.5, 290.05], [431.21, 290.24], [417.98, 197.02], [372.33, 196.99], [388.5, 195.99], [431.21, 196.01]]}]\n```" - } - ] - }, - { - "id": 2839, - "data_source": "Mono3DRefer", - "scene_name": "003632", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, person_sitting, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[454.24, 272.36], [418.08, 272.32], [427.45, 267.81], [461.83, 267.85], [454.24, 189.05], [418.08, 189.05], [427.45, 188.67], [461.83, 188.68]]}, {\"category\": \"van\", \"corners_3d\": [[832.39, 329.67], [999.76, 328.65], [1440.26, 482.85], [1101.28, 487.15], [832.39, 146.53], [999.76, 146.77], [1440.26, 110.36], [1101.28, 109.34]]}, {\"category\": \"tram\", \"corners_3d\": [[544.55, 193.85], [556.88, 194.51], [358.8, 195.91], [356.27, 195.1], [544.55, 142.17], [556.88, 140.07], [358.8, 135.61], [356.27, 138.17]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[402.52, 303.03], [459.28, 303.04], [446.17, 314.28], [384.17, 314.28], [402.52, 195.99], [459.28, 195.99], [446.17, 197.33], [384.17, 197.33]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[350.23, 301.21], [405.27, 301.22], [386.94, 312.43], [326.75, 312.42], [350.23, 195.96], [405.27, 195.97], [386.94, 197.32], [326.75, 197.32]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[417.98, 297.81], [372.33, 297.59], [388.5, 290.05], [431.21, 290.24], [417.98, 197.02], [372.33, 196.99], [388.5, 195.99], [431.21, 196.01]]}]\n```" - } - ] - }, - { - "id": 2840, - "data_source": "Mono3DRefer", - "scene_name": "003632", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[200.56, 310.02], [256.56, 308.2], [243.08, 316.8], [183.01, 318.87], [200.56, 161.85], [256.56, 162.13], [243.08, 160.82], [183.01, 160.5]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[235.05, 302.44], [191.5, 302.22], [214.31, 295.82], [255.59, 296.01], [235.05, 158.45], [191.5, 158.49], [214.31, 159.71], [255.59, 159.68]]}, {\"category\": \"van\", \"corners_3d\": [[832.39, 329.67], [999.76, 328.65], [1440.26, 482.85], [1101.28, 487.15], [832.39, 146.53], [999.76, 146.77], [1440.26, 110.36], [1101.28, 109.34]]}, {\"category\": \"tram\", \"corners_3d\": [[544.55, 193.85], [556.88, 194.51], [358.8, 195.91], [356.27, 195.1], [544.55, 142.17], [556.88, 140.07], [358.8, 135.61], [356.27, 138.17]]}]\n```" - } - ] - }, - { - "id": 2841, - "data_source": "Mono3DRefer", - "scene_name": "003632", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, pedestrian, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[454.24, 272.36], [418.08, 272.32], [427.45, 267.81], [461.83, 267.85], [454.24, 189.05], [418.08, 189.05], [427.45, 188.67], [461.83, 188.68]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[200.56, 310.02], [256.56, 308.2], [243.08, 316.8], [183.01, 318.87], [200.56, 161.85], [256.56, 162.13], [243.08, 160.82], [183.01, 160.5]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[235.05, 302.44], [191.5, 302.22], [214.31, 295.82], [255.59, 296.01], [235.05, 158.45], [191.5, 158.49], [214.31, 159.71], [255.59, 159.68]]}, {\"category\": \"tram\", \"corners_3d\": [[544.55, 193.85], [556.88, 194.51], [358.8, 195.91], [356.27, 195.1], [544.55, 142.17], [556.88, 140.07], [358.8, 135.61], [356.27, 138.17]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[402.52, 303.03], [459.28, 303.04], [446.17, 314.28], [384.17, 314.28], [402.52, 195.99], [459.28, 195.99], [446.17, 197.33], [384.17, 197.33]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[350.23, 301.21], [405.27, 301.22], [386.94, 312.43], [326.75, 312.42], [350.23, 195.96], [405.27, 195.97], [386.94, 197.32], [326.75, 197.32]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[417.98, 297.81], [372.33, 297.59], [388.5, 290.05], [431.21, 290.24], [417.98, 197.02], [372.33, 196.99], [388.5, 195.99], [431.21, 196.01]]}]\n```" - } - ] - }, - { - "id": 2842, - "data_source": "Mono3DRefer", - "scene_name": "003632", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, person_sitting, pedestrian, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[454.24, 272.36], [418.08, 272.32], [427.45, 267.81], [461.83, 267.85], [454.24, 189.05], [418.08, 189.05], [427.45, 188.67], [461.83, 188.68]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[200.56, 310.02], [256.56, 308.2], [243.08, 316.8], [183.01, 318.87], [200.56, 161.85], [256.56, 162.13], [243.08, 160.82], [183.01, 160.5]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[235.05, 302.44], [191.5, 302.22], [214.31, 295.82], [255.59, 296.01], [235.05, 158.45], [191.5, 158.49], [214.31, 159.71], [255.59, 159.68]]}, {\"category\": \"van\", \"corners_3d\": [[832.39, 329.67], [999.76, 328.65], [1440.26, 482.85], [1101.28, 487.15], [832.39, 146.53], [999.76, 146.77], [1440.26, 110.36], [1101.28, 109.34]]}, {\"category\": \"tram\", \"corners_3d\": [[544.55, 193.85], [556.88, 194.51], [358.8, 195.91], [356.27, 195.1], [544.55, 142.17], [556.88, 140.07], [358.8, 135.61], [356.27, 138.17]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[402.52, 303.03], [459.28, 303.04], [446.17, 314.28], [384.17, 314.28], [402.52, 195.99], [459.28, 195.99], [446.17, 197.33], [384.17, 197.33]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[350.23, 301.21], [405.27, 301.22], [386.94, 312.43], [326.75, 312.42], [350.23, 195.96], [405.27, 195.97], [386.94, 197.32], [326.75, 197.32]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[417.98, 297.81], [372.33, 297.59], [388.5, 290.05], [431.21, 290.24], [417.98, 197.02], [372.33, 196.99], [388.5, 195.99], [431.21, 196.01]]}]\n```" - } - ] - }, - { - "id": 2843, - "data_source": "Mono3DRefer", - "scene_name": "003635", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[614.52, 244.77], [584.0, 244.83], [585.04, 237.73], [612.55, 237.68], [614.52, 216.53], [584.0, 216.56], [585.04, 212.25], [612.55, 212.22]]}]\n```" - } - ] - }, - { - "id": 2844, - "data_source": "Mono3DRefer", - "scene_name": "003670", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[411.41, 328.11], [341.66, 334.76], [330.01, 317.87], [392.86, 312.57], [411.41, 146.38], [341.66, 144.84], [330.01, 148.75], [392.86, 149.97]]}]\n```" - } - ] - }, - { - "id": 2845, - "data_source": "Mono3DRefer", - "scene_name": "003671", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[752.4, 208.75], [760.97, 210.47], [670.0, 210.71], [665.56, 208.97], [752.4, 178.83], [760.97, 179.12], [670.0, 179.16], [665.56, 178.87]]}, {\"category\": \"car\", \"corners_3d\": [[650.35, 196.68], [650.25, 195.96], [700.9, 196.04], [702.58, 196.77], [650.35, 177.41], [650.25, 177.27], [700.9, 177.29], [702.58, 177.42]]}]\n```" - } - ] - }, - { - "id": 2846, - "data_source": "Mono3DRefer", - "scene_name": "003674", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[857.09, 440.4], [1096.82, 437.77], [2531.32, 1185.35], [1644.6, 1224.76], [857.09, 204.91], [1096.82, 204.6], [2531.32, 294.17], [1644.6, 298.89]]}, {\"category\": \"car\", \"corners_3d\": [[254.18, 363.55], [418.86, 362.3], [233.62, 569.17], [-115.98, 574.7], [254.18, 195.7], [418.86, 195.55], [233.62, 220.34], [-115.98, 221.0]]}, {\"category\": \"car\", \"corners_3d\": [[455.9, 327.59], [312.75, 327.93], [395.79, 283.37], [497.75, 283.2], [455.9, 201.68], [312.75, 201.74], [395.79, 193.44], [497.75, 193.41]]}, {\"category\": \"car\", \"corners_3d\": [[731.77, 284.48], [829.45, 283.42], [939.17, 328.36], [802.72, 330.47], [731.77, 184.45], [829.45, 184.34], [939.17, 189.01], [802.72, 189.22]]}, {\"category\": \"car\", \"corners_3d\": [[678.72, 251.45], [748.32, 251.3], [793.63, 274.59], [703.42, 274.84], [678.72, 186.81], [748.32, 186.79], [793.63, 190.92], [703.42, 190.97]]}, {\"category\": \"car\", \"corners_3d\": [[551.15, 233.5], [503.73, 233.34], [521.71, 225.23], [562.8, 225.35], [551.15, 185.47], [503.73, 185.43], [521.71, 183.75], [562.8, 183.77]]}, {\"category\": \"car\", \"corners_3d\": [[579.35, 220.65], [537.67, 220.6], [548.65, 214.48], [585.0, 214.52], [579.35, 172.59], [537.67, 172.59], [548.65, 172.63], [585.0, 172.63]]}, {\"category\": \"car\", \"corners_3d\": [[600.28, 214.14], [565.27, 214.0], [575.6, 209.49], [606.79, 209.6], [600.28, 175.97], [565.27, 175.96], [575.6, 175.62], [606.79, 175.63]]}, {\"category\": \"car\", \"corners_3d\": [[636.9, 203.1], [614.29, 202.84], [628.68, 200.7], [649.79, 200.93], [636.9, 181.35], [614.29, 181.27], [628.68, 180.67], [649.79, 180.74]]}]\n```" - } - ] - }, - { - "id": 2847, - "data_source": "Mono3DRefer", - "scene_name": "003680", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[663.55, 192.23], [646.25, 192.11], [655.64, 190.99], [671.99, 191.1], [663.55, 177.06], [646.25, 177.04], [655.64, 176.79], [671.99, 176.82]]}]\n```" - } - ] - }, - { - "id": 2848, - "data_source": "Mono3DRefer", - "scene_name": "003701", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[278.92, 372.08], [78.07, 372.68], [301.61, 287.39], [416.49, 287.19], [278.92, 174.0], [78.07, 174.01], [301.61, 173.52], [416.49, 173.51]]}, {\"category\": \"car\", \"corners_3d\": [[185.99, 211.56], [220.66, 211.56], [183.85, 215.22], [145.91, 215.21], [185.99, 180.1], [220.66, 180.11], [183.85, 180.79], [145.91, 180.79]]}, {\"category\": \"car\", \"corners_3d\": [[519.38, 219.41], [475.12, 219.39], [494.48, 213.0], [532.67, 213.02], [519.38, 178.7], [475.12, 178.7], [494.48, 177.89], [532.67, 177.9]]}, {\"category\": \"car\", \"corners_3d\": [[329.72, 197.16], [350.44, 197.18], [332.01, 198.77], [309.95, 198.75], [329.72, 176.19], [350.44, 176.19], [332.01, 176.41], [309.95, 176.41]]}, {\"category\": \"car\", \"corners_3d\": [[352.64, 196.33], [336.97, 196.31], [363.38, 194.25], [377.69, 194.27], [352.64, 175.5], [336.97, 175.5], [363.38, 175.27], [377.69, 175.27]]}, {\"category\": \"car\", \"corners_3d\": [[574.17, 192.24], [556.16, 192.24], [558.92, 191.07], [575.84, 191.07], [574.17, 177.22], [556.16, 177.22], [558.92, 176.95], [575.84, 176.95]]}, {\"category\": \"car\", \"corners_3d\": [[557.05, 201.56], [530.1, 201.57], [536.44, 199.03], [561.01, 199.02], [557.05, 173.66], [530.1, 173.66], [536.44, 173.59], [561.01, 173.59]]}, {\"category\": \"car\", \"corners_3d\": [[589.93, 207.22], [622.31, 207.25], [622.0, 212.12], [585.03, 212.07], [589.93, 175.03], [622.31, 175.03], [622.0, 175.34], [585.03, 175.34]]}]\n```" - } - ] - }, - { - "id": 2849, - "data_source": "Mono3DRefer", - "scene_name": "003705", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[415.88, 293.04], [314.92, 293.71], [383.91, 262.27], [458.43, 261.9], [415.88, 199.69], [314.92, 199.84], [383.91, 192.82], [458.43, 192.74]]}, {\"category\": \"car\", \"corners_3d\": [[554.69, 209.76], [517.68, 209.8], [526.15, 205.73], [559.07, 205.7], [554.69, 178.96], [517.68, 178.97], [526.15, 178.3], [559.07, 178.29]]}]\n```" - } - ] - }, - { - "id": 2850, - "data_source": "Mono3DRefer", - "scene_name": "003705", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[483.27, 248.83], [410.98, 248.91], [448.4, 233.99], [506.49, 233.94], [483.27, 170.23], [410.98, 170.23], [448.4, 170.74], [506.49, 170.75]]}]\n```" - } - ] - }, - { - "id": 2851, - "data_source": "Mono3DRefer", - "scene_name": "003705", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[415.88, 293.04], [314.92, 293.71], [383.91, 262.27], [458.43, 261.9], [415.88, 199.69], [314.92, 199.84], [383.91, 192.82], [458.43, 192.74]]}, {\"category\": \"van\", \"corners_3d\": [[483.27, 248.83], [410.98, 248.91], [448.4, 233.99], [506.49, 233.94], [483.27, 170.23], [410.98, 170.23], [448.4, 170.74], [506.49, 170.75]]}, {\"category\": \"car\", \"corners_3d\": [[554.69, 209.76], [517.68, 209.8], [526.15, 205.73], [559.07, 205.7], [554.69, 178.96], [517.68, 178.97], [526.15, 178.3], [559.07, 178.29]]}]\n```" - } - ] - }, - { - "id": 2852, - "data_source": "Mono3DRefer", - "scene_name": "003710", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[706.47, 246.8], [772.92, 246.47], [839.84, 271.23], [751.3, 271.84], [706.47, 185.31], [772.92, 185.26], [839.84, 189.43], [751.3, 189.53]]}, {\"category\": \"car\", \"corners_3d\": [[979.1, 365.34], [1183.42, 364.28], [2108.9, 665.71], [1591.49, 672.78], [979.1, 165.31], [1183.42, 165.35], [2108.9, 153.53], [1591.49, 153.26]]}, {\"category\": \"car\", \"corners_3d\": [[-53.42, 235.54], [-105.99, 235.58], [-3.36, 226.49], [41.53, 226.45], [-53.42, 190.07], [-105.99, 190.08], [-3.36, 187.58], [41.53, 187.57]]}]\n```" - } - ] - }, - { - "id": 2853, - "data_source": "Mono3DRefer", - "scene_name": "003724", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[548.71, 207.34], [518.33, 207.39], [524.62, 204.24], [552.23, 204.2], [548.71, 177.42], [518.33, 177.43], [524.62, 177.01], [552.23, 177.01]]}]\n```" - } - ] - }, - { - "id": 2854, - "data_source": "Mono3DRefer", - "scene_name": "003727", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[565.54, 224.14], [517.16, 223.8], [541.77, 216.16], [583.0, 216.41], [565.54, 181.93], [517.16, 181.87], [541.77, 180.52], [583.0, 180.56]]}]\n```" - } - ] - }, - { - "id": 2855, - "data_source": "Mono3DRefer", - "scene_name": "003733", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[588.22, 190.74], [567.29, 190.75], [569.08, 189.58], [588.64, 189.57], [588.22, 168.71], [567.29, 168.71], [569.08, 168.98], [588.64, 168.98]]}, {\"category\": \"car\", \"corners_3d\": [[532.14, 243.21], [469.43, 243.33], [491.55, 230.88], [543.15, 230.79], [532.14, 189.77], [469.43, 189.8], [491.55, 186.81], [543.15, 186.79]]}]\n```" - } - ] - }, - { - "id": 2856, - "data_source": "Mono3DRefer", - "scene_name": "003738", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[436.94, 331.17], [280.69, 331.88], [388.18, 277.42], [490.76, 277.11], [436.94, 202.02], [280.69, 202.15], [388.18, 192.11], [490.76, 192.06]]}]\n```" - } - ] - }, - { - "id": 2857, - "data_source": "Mono3DRefer", - "scene_name": "003738", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[598.57, 233.54], [623.18, 233.37], [629.56, 238.82], [602.75, 239.01], [598.57, 205.67], [623.18, 205.58], [629.56, 208.52], [602.75, 208.63]]}]\n```" - } - ] - }, - { - "id": 2858, - "data_source": "Mono3DRefer", - "scene_name": "003738", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[436.94, 331.17], [280.69, 331.88], [388.18, 277.42], [490.76, 277.11], [436.94, 202.02], [280.69, 202.15], [388.18, 192.11], [490.76, 192.06]]}, {\"category\": \"van\", \"corners_3d\": [[598.57, 233.54], [623.18, 233.37], [629.56, 238.82], [602.75, 239.01], [598.57, 205.67], [623.18, 205.58], [629.56, 208.52], [602.75, 208.63]]}]\n```" - } - ] - }, - { - "id": 2859, - "data_source": "Mono3DRefer", - "scene_name": "003745", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[286.54, 203.73], [266.69, 203.69], [307.14, 200.33], [324.87, 200.36], [286.54, 177.24], [266.69, 177.23], [307.14, 176.75], [324.87, 176.76]]}, {\"category\": \"car\", \"corners_3d\": [[562.11, 201.08], [535.28, 201.09], [541.09, 198.62], [565.58, 198.61], [562.11, 178.72], [535.28, 178.72], [541.09, 178.21], [565.58, 178.2]]}, {\"category\": \"car\", \"corners_3d\": [[516.28, 227.53], [463.08, 227.57], [485.38, 218.77], [530.01, 218.74], [516.28, 172.52], [463.08, 172.52], [485.38, 172.58], [530.01, 172.58]]}, {\"category\": \"car\", \"corners_3d\": [[591.65, 207.42], [624.42, 207.44], [625.44, 212.4], [587.97, 212.38], [591.65, 174.86], [624.42, 174.86], [625.44, 175.15], [587.97, 175.14]]}, {\"category\": \"car\", \"corners_3d\": [[114.36, 449.34], [-141.8, 448.43], [222.59, 315.95], [356.02, 316.2], [114.36, 211.84], [-141.8, 211.71], [222.59, 193.03], [356.02, 193.06]]}, {\"category\": \"car\", \"corners_3d\": [[250.84, 205.21], [277.6, 205.23], [247.73, 208.02], [218.7, 207.99], [250.84, 178.18], [277.6, 178.19], [247.73, 178.65], [218.7, 178.64]]}]\n```" - } - ] - }, - { - "id": 2860, - "data_source": "Mono3DRefer", - "scene_name": "003752", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[581.11, 217.66], [539.13, 217.35], [559.91, 211.45], [596.43, 211.69], [581.11, 180.99], [539.13, 180.94], [559.91, 179.87], [596.43, 179.91]]}]\n```" - } - ] - }, - { - "id": 2861, - "data_source": "Mono3DRefer", - "scene_name": "003756", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[615.29, 220.97], [652.32, 220.8], [665.51, 227.61], [623.24, 227.83], [615.29, 178.95], [652.32, 178.93], [665.51, 179.79], [623.24, 179.82]]}, {\"category\": \"van\", \"corners_3d\": [[537.11, 201.73], [557.89, 201.61], [561.35, 203.63], [539.13, 203.76], [537.11, 176.09], [557.89, 176.08], [561.35, 176.31], [539.13, 176.32]]}]\n```" - } - ] - }, - { - "id": 2862, - "data_source": "Mono3DRefer", - "scene_name": "003756", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[988.16, 239.61], [987.33, 234.44], [1181.64, 241.12], [1200.8, 247.53], [988.16, 177.29], [987.33, 176.95], [1181.64, 177.39], [1200.8, 177.82]]}, {\"category\": \"car\", \"corners_3d\": [[766.11, 226.44], [790.58, 228.42], [726.99, 232.58], [703.19, 230.29], [766.11, 198.84], [790.58, 199.8], [726.99, 201.81], [703.19, 200.7]]}]\n```" - } - ] - }, - { - "id": 2863, - "data_source": "Mono3DRefer", - "scene_name": "003756", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[988.16, 239.61], [987.33, 234.44], [1181.64, 241.12], [1200.8, 247.53], [988.16, 177.29], [987.33, 176.95], [1181.64, 177.39], [1200.8, 177.82]]}, {\"category\": \"van\", \"corners_3d\": [[615.29, 220.97], [652.32, 220.8], [665.51, 227.61], [623.24, 227.83], [615.29, 178.95], [652.32, 178.93], [665.51, 179.79], [623.24, 179.82]]}, {\"category\": \"car\", \"corners_3d\": [[766.11, 226.44], [790.58, 228.42], [726.99, 232.58], [703.19, 230.29], [766.11, 198.84], [790.58, 199.8], [726.99, 201.81], [703.19, 200.7]]}, {\"category\": \"van\", \"corners_3d\": [[537.11, 201.73], [557.89, 201.61], [561.35, 203.63], [539.13, 203.76], [537.11, 176.09], [557.89, 176.08], [561.35, 176.31], [539.13, 176.32]]}]\n```" - } - ] - }, - { - "id": 2864, - "data_source": "Mono3DRefer", - "scene_name": "003762", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[331.29, 245.42], [297.91, 245.36], [307.87, 243.14], [340.23, 243.2], [331.29, 179.85], [297.91, 179.85], [307.87, 179.63], [340.23, 179.64]]}]\n```" - } - ] - }, - { - "id": 2865, - "data_source": "Mono3DRefer", - "scene_name": "003763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[295.13, 382.28], [188.22, 385.46], [271.52, 334.32], [350.97, 332.53], [295.13, 151.88], [188.22, 151.41], [271.52, 158.96], [350.97, 159.22]]}]\n```" - } - ] - }, - { - "id": 2866, - "data_source": "Mono3DRefer", - "scene_name": "003763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[696.99, 235.4], [652.37, 237.09], [635.3, 234.66], [678.52, 233.12], [696.99, 148.19], [652.37, 147.15], [635.3, 148.65], [678.52, 149.61]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[466.16, 213.91], [489.31, 213.8], [488.23, 214.87], [464.3, 214.99], [466.16, 166.75], [489.31, 166.8], [488.23, 166.31], [464.3, 166.26]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[482.47, 213.94], [507.03, 213.8], [507.03, 214.99], [481.56, 215.14], [482.47, 164.97], [507.03, 165.04], [507.03, 164.43], [481.56, 164.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[338.67, 220.59], [366.7, 220.48], [358.68, 222.18], [329.41, 222.3], [338.67, 167.43], [366.7, 167.47], [358.68, 166.86], [329.41, 166.82]]}]\n```" - } - ] - }, - { - "id": 2867, - "data_source": "Mono3DRefer", - "scene_name": "003763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[197.81, 359.34], [-2.82, 364.54], [203.66, 292.32], [322.92, 290.39], [197.81, 185.77], [-2.82, 185.89], [203.66, 184.17], [322.92, 184.12]]}]\n```" - } - ] - }, - { - "id": 2868, - "data_source": "Mono3DRefer", - "scene_name": "003763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[295.13, 382.28], [188.22, 385.46], [271.52, 334.32], [350.97, 332.53], [295.13, 151.88], [188.22, 151.41], [271.52, 158.96], [350.97, 159.22]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[696.99, 235.4], [652.37, 237.09], [635.3, 234.66], [678.52, 233.12], [696.99, 148.19], [652.37, 147.15], [635.3, 148.65], [678.52, 149.61]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[466.16, 213.91], [489.31, 213.8], [488.23, 214.87], [464.3, 214.99], [466.16, 166.75], [489.31, 166.8], [488.23, 166.31], [464.3, 166.26]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[482.47, 213.94], [507.03, 213.8], [507.03, 214.99], [481.56, 215.14], [482.47, 164.97], [507.03, 165.04], [507.03, 164.43], [481.56, 164.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[338.67, 220.59], [366.7, 220.48], [358.68, 222.18], [329.41, 222.3], [338.67, 167.43], [366.7, 167.47], [358.68, 166.86], [329.41, 166.82]]}]\n```" - } - ] - }, - { - "id": 2869, - "data_source": "Mono3DRefer", - "scene_name": "003763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[197.81, 359.34], [-2.82, 364.54], [203.66, 292.32], [322.92, 290.39], [197.81, 185.77], [-2.82, 185.89], [203.66, 184.17], [322.92, 184.12]]}, {\"category\": \"cyclist\", \"corners_3d\": [[295.13, 382.28], [188.22, 385.46], [271.52, 334.32], [350.97, 332.53], [295.13, 151.88], [188.22, 151.41], [271.52, 158.96], [350.97, 159.22]]}]\n```" - } - ] - }, - { - "id": 2870, - "data_source": "Mono3DRefer", - "scene_name": "003763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[197.81, 359.34], [-2.82, 364.54], [203.66, 292.32], [322.92, 290.39], [197.81, 185.77], [-2.82, 185.89], [203.66, 184.17], [322.92, 184.12]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[696.99, 235.4], [652.37, 237.09], [635.3, 234.66], [678.52, 233.12], [696.99, 148.19], [652.37, 147.15], [635.3, 148.65], [678.52, 149.61]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[466.16, 213.91], [489.31, 213.8], [488.23, 214.87], [464.3, 214.99], [466.16, 166.75], [489.31, 166.8], [488.23, 166.31], [464.3, 166.26]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[482.47, 213.94], [507.03, 213.8], [507.03, 214.99], [481.56, 215.14], [482.47, 164.97], [507.03, 165.04], [507.03, 164.43], [481.56, 164.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[338.67, 220.59], [366.7, 220.48], [358.68, 222.18], [329.41, 222.3], [338.67, 167.43], [366.7, 167.47], [358.68, 166.86], [329.41, 166.82]]}]\n```" - } - ] - }, - { - "id": 2871, - "data_source": "Mono3DRefer", - "scene_name": "003763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[197.81, 359.34], [-2.82, 364.54], [203.66, 292.32], [322.92, 290.39], [197.81, 185.77], [-2.82, 185.89], [203.66, 184.17], [322.92, 184.12]]}, {\"category\": \"cyclist\", \"corners_3d\": [[295.13, 382.28], [188.22, 385.46], [271.52, 334.32], [350.97, 332.53], [295.13, 151.88], [188.22, 151.41], [271.52, 158.96], [350.97, 159.22]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[696.99, 235.4], [652.37, 237.09], [635.3, 234.66], [678.52, 233.12], [696.99, 148.19], [652.37, 147.15], [635.3, 148.65], [678.52, 149.61]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[466.16, 213.91], [489.31, 213.8], [488.23, 214.87], [464.3, 214.99], [466.16, 166.75], [489.31, 166.8], [488.23, 166.31], [464.3, 166.26]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[482.47, 213.94], [507.03, 213.8], [507.03, 214.99], [481.56, 215.14], [482.47, 164.97], [507.03, 165.04], [507.03, 164.43], [481.56, 164.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[338.67, 220.59], [366.7, 220.48], [358.68, 222.18], [329.41, 222.3], [338.67, 167.43], [366.7, 167.47], [358.68, 166.86], [329.41, 166.82]]}]\n```" - } - ] - }, - { - "id": 2872, - "data_source": "Mono3DRefer", - "scene_name": "003768", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[500.5, 234.83], [451.1, 235.37], [458.41, 228.18], [502.11, 227.75], [500.5, 189.17], [451.1, 189.31], [458.41, 187.42], [502.11, 187.31]]}]\n```" - } - ] - }, - { - "id": 2873, - "data_source": "Mono3DRefer", - "scene_name": "003774", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1094.13, 295.93], [1135.43, 305.81], [1021.04, 305.6], [988.8, 295.76], [1094.13, 147.93], [1135.43, 145.15], [1021.04, 145.2], [988.8, 147.98]]}]\n```" - } - ] - }, - { - "id": 2874, - "data_source": "Mono3DRefer", - "scene_name": "003776", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[562.99, 222.26], [612.02, 222.06], [619.67, 230.39], [562.35, 230.66], [562.99, 174.91], [612.02, 174.91], [619.67, 175.25], [562.35, 175.26]]}]\n```" - } - ] - }, - { - "id": 2875, - "data_source": "Mono3DRefer", - "scene_name": "003777", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[205.58, 272.74], [157.49, 272.61], [177.81, 268.25], [223.83, 268.37], [205.58, 177.91], [157.49, 177.9], [177.81, 177.68], [223.83, 177.69]]}]\n```" - } - ] - }, - { - "id": 2876, - "data_source": "Mono3DRefer", - "scene_name": "003789", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[501.37, 233.15], [451.76, 233.23], [471.85, 224.74], [514.46, 224.67], [501.37, 184.33], [451.76, 184.35], [471.85, 182.73], [514.46, 182.72]]}]\n```" - } - ] - }, - { - "id": 2877, - "data_source": "Mono3DRefer", - "scene_name": "003794", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[333.15, 219.68], [292.38, 219.68], [330.75, 214.0], [366.59, 214.0], [333.15, 186.07], [292.38, 186.07], [330.75, 184.47], [366.59, 184.47]]}]\n```" - } - ] - }, - { - "id": 2878, - "data_source": "Mono3DRefer", - "scene_name": "003794", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[267.21, 218.18], [213.44, 218.19], [284.69, 210.02], [328.77, 210.02], [267.21, 141.59], [213.44, 141.58], [284.69, 147.22], [328.77, 147.22]]}, {\"category\": \"truck\", \"corners_3d\": [[428.27, 198.47], [403.34, 198.47], [422.03, 196.14], [444.69, 196.14], [428.27, 165.18], [403.34, 165.18], [422.03, 165.88], [444.69, 165.88]]}, {\"category\": \"truck\", \"corners_3d\": [[151.95, 233.84], [78.33, 233.9], [199.83, 219.73], [256.27, 219.69], [151.95, 132.02], [78.33, 131.98], [199.83, 141.46], [256.27, 141.49]]}]\n```" - } - ] - }, - { - "id": 2879, - "data_source": "Mono3DRefer", - "scene_name": "003794", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[267.21, 218.18], [213.44, 218.19], [284.69, 210.02], [328.77, 210.02], [267.21, 141.59], [213.44, 141.58], [284.69, 147.22], [328.77, 147.22]]}, {\"category\": \"truck\", \"corners_3d\": [[428.27, 198.47], [403.34, 198.47], [422.03, 196.14], [444.69, 196.14], [428.27, 165.18], [403.34, 165.18], [422.03, 165.88], [444.69, 165.88]]}, {\"category\": \"car\", \"corners_3d\": [[333.15, 219.68], [292.38, 219.68], [330.75, 214.0], [366.59, 214.0], [333.15, 186.07], [292.38, 186.07], [330.75, 184.47], [366.59, 184.47]]}, {\"category\": \"truck\", \"corners_3d\": [[151.95, 233.84], [78.33, 233.9], [199.83, 219.73], [256.27, 219.69], [151.95, 132.02], [78.33, 131.98], [199.83, 141.46], [256.27, 141.49]]}]\n```" - } - ] - }, - { - "id": 2880, - "data_source": "Mono3DRefer", - "scene_name": "003803", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1242.04, 240.39], [1345.79, 251.59], [995.01, 251.21], [941.41, 240.11], [1242.04, 141.08], [1345.79, 135.81], [995.01, 135.98], [941.41, 141.21]]}, {\"category\": \"car\", \"corners_3d\": [[605.74, 203.98], [565.41, 203.88], [575.49, 200.28], [611.16, 200.36], [605.74, 165.36], [565.41, 165.39], [575.49, 166.25], [611.16, 166.23]]}, {\"category\": \"car\", \"corners_3d\": [[638.64, 181.22], [618.28, 181.2], [621.91, 180.65], [640.96, 180.67], [638.64, 163.74], [618.28, 163.77], [621.91, 164.36], [640.96, 164.34]]}, {\"category\": \"car\", \"corners_3d\": [[529.77, 222.02], [583.09, 222.4], [558.32, 234.74], [492.07, 234.16], [529.77, 167.62], [583.09, 167.58], [558.32, 166.26], [492.07, 166.33]]}, {\"category\": \"car\", \"corners_3d\": [[-98.08, 326.96], [-15.79, 315.47], [52.77, 329.63], [-32.3, 343.63], [-98.08, 244.89], [-15.79, 239.52], [52.77, 246.14], [-32.3, 252.68]]}, {\"category\": \"car\", \"corners_3d\": [[1602.67, 275.66], [1874.83, 303.83], [1392.01, 303.83], [1223.7, 275.66], [1602.67, 134.73], [1874.83, 124.29], [1392.01, 124.29], [1223.7, 134.73]]}]\n```" - } - ] - }, - { - "id": 2881, - "data_source": "Mono3DRefer", - "scene_name": "003817", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[566.71, 218.38], [534.26, 218.61], [533.78, 214.19], [563.11, 214.01], [566.71, 185.12], [534.26, 185.18], [533.78, 183.99], [563.11, 183.94]]}, {\"category\": \"car\", \"corners_3d\": [[530.79, 204.12], [509.77, 204.36], [502.93, 202.22], [522.58, 202.0], [530.79, 185.95], [509.77, 186.05], [502.93, 185.15], [522.58, 185.06]]}]\n```" - } - ] - }, - { - "id": 2882, - "data_source": "Mono3DRefer", - "scene_name": "003824", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[691.46, 226.81], [749.13, 227.37], [758.67, 239.02], [688.71, 238.2], [691.46, 173.19], [749.13, 173.19], [758.67, 173.26], [688.71, 173.26]]}, {\"category\": \"car\", \"corners_3d\": [[453.36, 259.5], [381.97, 257.78], [452.86, 241.32], [511.34, 242.43], [453.36, 173.33], [381.97, 173.32], [452.86, 173.23], [511.34, 173.23]]}, {\"category\": \"car\", \"corners_3d\": [[62.02, 300.41], [-27.69, 296.85], [184.3, 262.95], [253.36, 264.81], [62.02, 194.54], [-27.69, 193.94], [184.3, 188.18], [253.36, 188.49]]}, {\"category\": \"car\", \"corners_3d\": [[440.55, 230.7], [395.51, 229.45], [464.28, 220.7], [503.5, 221.59], [440.55, 186.29], [395.51, 186.0], [464.28, 183.97], [503.5, 184.17]]}, {\"category\": \"car\", \"corners_3d\": [[857.59, 192.47], [877.76, 193.18], [825.0, 193.93], [806.03, 193.16], [857.59, 166.42], [877.76, 166.19], [825.0, 165.94], [806.03, 166.19]]}]\n```" - } - ] - }, - { - "id": 2883, - "data_source": "Mono3DRefer", - "scene_name": "003825", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[488.48, 239.35], [421.73, 240.32], [436.04, 226.69], [489.29, 226.07], [488.48, 181.82], [421.73, 181.95], [436.04, 180.11], [489.29, 180.03]]}]\n```" - } - ] - }, - { - "id": 2884, - "data_source": "Mono3DRefer", - "scene_name": "003832", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[209.1, 213.78], [177.42, 213.77], [220.83, 209.72], [249.39, 209.73], [209.1, 175.07], [177.42, 175.06], [220.83, 174.85], [249.39, 174.85]]}, {\"category\": \"car\", \"corners_3d\": [[590.38, 210.45], [624.07, 210.49], [623.99, 216.06], [585.32, 216.01], [590.38, 176.98], [624.07, 176.98], [623.99, 177.59], [585.32, 177.59]]}, {\"category\": \"car\", \"corners_3d\": [[488.7, 249.09], [424.35, 248.96], [460.54, 235.12], [513.23, 235.21], [488.7, 182.69], [424.35, 182.67], [460.54, 180.89], [513.23, 180.9]]}]\n```" - } - ] - }, - { - "id": 2885, - "data_source": "Mono3DRefer", - "scene_name": "003832", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[570.07, 202.06], [536.31, 202.07], [545.82, 197.83], [574.68, 197.82], [570.07, 154.93], [536.31, 154.92], [545.82, 157.52], [574.68, 157.53]]}]\n```" - } - ] - }, - { - "id": 2886, - "data_source": "Mono3DRefer", - "scene_name": "003832", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[209.1, 213.78], [177.42, 213.77], [220.83, 209.72], [249.39, 209.73], [209.1, 175.07], [177.42, 175.06], [220.83, 174.85], [249.39, 174.85]]}, {\"category\": \"truck\", \"corners_3d\": [[570.07, 202.06], [536.31, 202.07], [545.82, 197.83], [574.68, 197.82], [570.07, 154.93], [536.31, 154.92], [545.82, 157.52], [574.68, 157.53]]}, {\"category\": \"car\", \"corners_3d\": [[590.38, 210.45], [624.07, 210.49], [623.99, 216.06], [585.32, 216.01], [590.38, 176.98], [624.07, 176.98], [623.99, 177.59], [585.32, 177.59]]}, {\"category\": \"car\", \"corners_3d\": [[488.7, 249.09], [424.35, 248.96], [460.54, 235.12], [513.23, 235.21], [488.7, 182.69], [424.35, 182.67], [460.54, 180.89], [513.23, 180.9]]}]\n```" - } - ] - }, - { - "id": 2887, - "data_source": "Mono3DRefer", - "scene_name": "003841", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[569.93, 190.37], [592.86, 190.35], [594.89, 193.41], [567.95, 193.44], [569.93, 165.53], [592.86, 165.54], [594.89, 164.26], [567.95, 164.25]]}]\n```" - } - ] - }, - { - "id": 2888, - "data_source": "Mono3DRefer", - "scene_name": "003841", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[531.87, 194.29], [550.94, 194.27], [549.47, 195.31], [529.47, 195.34], [531.87, 176.04], [550.94, 176.04], [549.47, 176.19], [529.47, 176.19]]}]\n```" - } - ] - }, - { - "id": 2889, - "data_source": "Mono3DRefer", - "scene_name": "003841", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[485.44, 192.78], [509.31, 192.76], [490.22, 198.04], [459.97, 198.08], [485.44, 164.4], [509.31, 164.41], [490.22, 162.17], [459.97, 162.15]]}]\n```" - } - ] - }, - { - "id": 2890, - "data_source": "Mono3DRefer", - "scene_name": "003841", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[531.87, 194.29], [550.94, 194.27], [549.47, 195.31], [529.47, 195.34], [531.87, 176.04], [550.94, 176.04], [549.47, 176.19], [529.47, 176.19]]}, {\"category\": \"bus\", \"corners_3d\": [[569.93, 190.37], [592.86, 190.35], [594.89, 193.41], [567.95, 193.44], [569.93, 165.53], [592.86, 165.54], [594.89, 164.26], [567.95, 164.25]]}]\n```" - } - ] - }, - { - "id": 2891, - "data_source": "Mono3DRefer", - "scene_name": "003841", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[569.93, 190.37], [592.86, 190.35], [594.89, 193.41], [567.95, 193.44], [569.93, 165.53], [592.86, 165.54], [594.89, 164.26], [567.95, 164.25]]}, {\"category\": \"tram\", \"corners_3d\": [[485.44, 192.78], [509.31, 192.76], [490.22, 198.04], [459.97, 198.08], [485.44, 164.4], [509.31, 164.41], [490.22, 162.17], [459.97, 162.15]]}]\n```" - } - ] - }, - { - "id": 2892, - "data_source": "Mono3DRefer", - "scene_name": "003841", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[531.87, 194.29], [550.94, 194.27], [549.47, 195.31], [529.47, 195.34], [531.87, 176.04], [550.94, 176.04], [549.47, 176.19], [529.47, 176.19]]}, {\"category\": \"tram\", \"corners_3d\": [[485.44, 192.78], [509.31, 192.76], [490.22, 198.04], [459.97, 198.08], [485.44, 164.4], [509.31, 164.41], [490.22, 162.17], [459.97, 162.15]]}]\n```" - } - ] - }, - { - "id": 2893, - "data_source": "Mono3DRefer", - "scene_name": "003841", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[531.87, 194.29], [550.94, 194.27], [549.47, 195.31], [529.47, 195.34], [531.87, 176.04], [550.94, 176.04], [549.47, 176.19], [529.47, 176.19]]}, {\"category\": \"bus\", \"corners_3d\": [[569.93, 190.37], [592.86, 190.35], [594.89, 193.41], [567.95, 193.44], [569.93, 165.53], [592.86, 165.54], [594.89, 164.26], [567.95, 164.25]]}, {\"category\": \"tram\", \"corners_3d\": [[485.44, 192.78], [509.31, 192.76], [490.22, 198.04], [459.97, 198.08], [485.44, 164.4], [509.31, 164.41], [490.22, 162.17], [459.97, 162.15]]}]\n```" - } - ] - }, - { - "id": 2894, - "data_source": "Mono3DRefer", - "scene_name": "003847", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-12.12, 571.07], [-414.73, 575.61], [175.27, 340.17], [340.86, 339.38], [-12.12, 225.24], [-414.73, 225.84], [175.27, 194.86], [340.86, 194.76]]}, {\"category\": \"car\", \"corners_3d\": [[796.25, 263.55], [712.5, 263.98], [682.83, 243.56], [747.99, 243.3], [796.25, 177.1], [712.5, 177.12], [682.83, 176.17], [747.99, 176.15]]}, {\"category\": \"car\", \"corners_3d\": [[516.95, 237.91], [455.89, 237.91], [483.96, 225.99], [533.82, 225.98], [516.95, 184.9], [455.89, 184.9], [483.96, 182.7], [533.82, 182.7]]}, {\"category\": \"car\", \"corners_3d\": [[655.64, 226.49], [703.11, 226.28], [727.23, 235.73], [671.44, 236.02], [655.64, 182.19], [703.11, 182.15], [727.23, 183.79], [671.44, 183.84]]}, {\"category\": \"car\", \"corners_3d\": [[551.03, 212.19], [520.18, 212.22], [526.42, 208.92], [554.68, 208.89], [551.03, 182.43], [520.18, 182.44], [526.42, 181.63], [554.68, 181.62]]}]\n```" - } - ] - }, - { - "id": 2895, - "data_source": "Mono3DRefer", - "scene_name": "003850", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[783.67, 281.51], [889.9, 280.87], [1098.77, 353.97], [921.99, 355.76], [783.67, 182.84], [889.9, 182.78], [1098.77, 189.5], [921.99, 189.66]]}, {\"category\": \"car\", \"corners_3d\": [[222.59, 205.6], [193.25, 205.65], [227.65, 202.74], [254.33, 202.69], [222.59, 178.27], [193.25, 178.28], [227.65, 177.8], [254.33, 177.79]]}, {\"category\": \"car\", \"corners_3d\": [[257.78, 199.2], [231.02, 199.23], [259.2, 197.14], [283.82, 197.11], [257.78, 176.16], [231.02, 176.17], [259.2, 175.9], [283.82, 175.9]]}]\n```" - } - ] - }, - { - "id": 2896, - "data_source": "Mono3DRefer", - "scene_name": "003852", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[848.23, 353.24], [1007.32, 353.28], [1470.5, 563.68], [1125.78, 563.53], [848.23, 199.32], [1007.32, 199.33], [1470.5, 230.2], [1125.78, 230.17]]}, {\"category\": \"car\", \"corners_3d\": [[740.87, 276.32], [839.03, 276.05], [948.99, 322.68], [806.72, 323.25], [740.87, 184.0], [839.03, 183.98], [948.99, 189.0], [806.72, 189.06]]}, {\"category\": \"car\", \"corners_3d\": [[530.9, 229.77], [478.29, 230.06], [490.84, 221.1], [535.19, 220.9], [530.9, 177.01], [478.29, 177.03], [490.84, 176.37], [535.19, 176.36]]}, {\"category\": \"car\", \"corners_3d\": [[664.35, 213.27], [701.89, 213.25], [715.45, 218.78], [672.77, 218.8], [664.35, 179.46], [701.89, 179.46], [715.45, 180.37], [672.77, 180.37]]}, {\"category\": \"car\", \"corners_3d\": [[555.28, 208.14], [520.95, 208.19], [528.22, 204.32], [558.79, 204.28], [555.28, 180.38], [520.95, 180.39], [528.22, 179.56], [558.79, 179.56]]}, {\"category\": \"car\", \"corners_3d\": [[569.57, 199.13], [541.65, 199.17], [544.7, 197.08], [570.41, 197.05], [569.57, 172.85], [541.65, 172.85], [544.7, 172.85], [570.41, 172.85]]}, {\"category\": \"car\", \"corners_3d\": [[676.09, 202.21], [647.95, 202.22], [643.7, 199.52], [669.26, 199.51], [676.09, 176.43], [647.95, 176.43], [643.7, 176.1], [669.26, 176.1]]}]\n```" - } - ] - }, - { - "id": 2897, - "data_source": "Mono3DRefer", - "scene_name": "003852", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[681.38, 236.76], [751.17, 236.7], [807.34, 260.89], [711.15, 261.0], [681.38, 152.71], [751.17, 152.73], [807.34, 145.11], [711.15, 145.07]]}]\n```" - } - ] - }, - { - "id": 2898, - "data_source": "Mono3DRefer", - "scene_name": "003852", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[848.23, 353.24], [1007.32, 353.28], [1470.5, 563.68], [1125.78, 563.53], [848.23, 199.32], [1007.32, 199.33], [1470.5, 230.2], [1125.78, 230.17]]}, {\"category\": \"car\", \"corners_3d\": [[740.87, 276.32], [839.03, 276.05], [948.99, 322.68], [806.72, 323.25], [740.87, 184.0], [839.03, 183.98], [948.99, 189.0], [806.72, 189.06]]}, {\"category\": \"van\", \"corners_3d\": [[681.38, 236.76], [751.17, 236.7], [807.34, 260.89], [711.15, 261.0], [681.38, 152.71], [751.17, 152.73], [807.34, 145.11], [711.15, 145.07]]}, {\"category\": \"car\", \"corners_3d\": [[530.9, 229.77], [478.29, 230.06], [490.84, 221.1], [535.19, 220.9], [530.9, 177.01], [478.29, 177.03], [490.84, 176.37], [535.19, 176.36]]}, {\"category\": \"car\", \"corners_3d\": [[664.35, 213.27], [701.89, 213.25], [715.45, 218.78], [672.77, 218.8], [664.35, 179.46], [701.89, 179.46], [715.45, 180.37], [672.77, 180.37]]}, {\"category\": \"car\", \"corners_3d\": [[555.28, 208.14], [520.95, 208.19], [528.22, 204.32], [558.79, 204.28], [555.28, 180.38], [520.95, 180.39], [528.22, 179.56], [558.79, 179.56]]}, {\"category\": \"car\", \"corners_3d\": [[569.57, 199.13], [541.65, 199.17], [544.7, 197.08], [570.41, 197.05], [569.57, 172.85], [541.65, 172.85], [544.7, 172.85], [570.41, 172.85]]}, {\"category\": \"car\", \"corners_3d\": [[676.09, 202.21], [647.95, 202.22], [643.7, 199.52], [669.26, 199.51], [676.09, 176.43], [647.95, 176.43], [643.7, 176.1], [669.26, 176.1]]}]\n```" - } - ] - }, - { - "id": 2899, - "data_source": "Mono3DRefer", - "scene_name": "003859", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[162.06, 378.43], [372.16, 377.9], [-404.05, 1065.74], [-1368.45, 1076.52], [162.06, 197.44], [372.16, 197.4], [-404.05, 253.05], [-1368.45, 253.92]]}, {\"category\": \"car\", \"corners_3d\": [[533.43, 198.66], [563.72, 198.95], [524.91, 201.1], [491.69, 200.73], [533.43, 170.1], [563.72, 169.9], [524.91, 168.47], [491.69, 168.72]]}, {\"category\": \"car\", \"corners_3d\": [[574.2, 190.46], [600.51, 190.5], [594.34, 191.26], [565.82, 191.22], [574.2, 167.22], [600.51, 167.17], [594.34, 165.95], [565.82, 166.01]]}, {\"category\": \"car\", \"corners_3d\": [[106.35, 226.58], [90.52, 229.66], [-61.59, 232.62], [-34.58, 229.16], [106.35, 180.21], [90.52, 180.12], [-61.59, 180.04], [-34.58, 180.14]]}, {\"category\": \"car\", \"corners_3d\": [[385.56, 193.45], [357.44, 193.37], [392.48, 192.23], [418.11, 192.3], [385.56, 162.99], [357.44, 163.12], [392.48, 164.88], [418.11, 164.78]]}]\n```" - } - ] - }, - { - "id": 2900, - "data_source": "Mono3DRefer", - "scene_name": "003864", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[640.23, 196.38], [672.56, 196.38], [680.85, 199.51], [644.23, 199.51], [640.23, 165.35], [672.56, 165.35], [680.85, 164.35], [644.23, 164.35]]}]\n```" - } - ] - }, - { - "id": 2901, - "data_source": "Mono3DRefer", - "scene_name": "003864", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[657.99, 205.72], [689.01, 205.74], [698.01, 209.86], [663.08, 209.84], [657.99, 179.23], [689.01, 179.24], [698.01, 180.04], [663.08, 180.03]]}, {\"category\": \"car\", \"corners_3d\": [[537.43, 212.23], [533.06, 214.61], [425.24, 214.61], [435.77, 212.23], [537.43, 179.26], [533.06, 179.64], [425.24, 179.64], [435.77, 179.26]]}, {\"category\": \"car\", \"corners_3d\": [[505.99, 225.42], [492.29, 229.69], [359.29, 228.73], [382.6, 224.6], [505.99, 179.19], [492.29, 179.7], [359.29, 179.59], [382.6, 179.09]]}, {\"category\": \"car\", \"corners_3d\": [[492.18, 232.85], [478.28, 237.25], [340.61, 235.99], [363.49, 231.76], [492.18, 182.0], [478.28, 182.67], [340.61, 182.47], [363.49, 181.83]]}, {\"category\": \"car\", \"corners_3d\": [[460.83, 238.86], [446.21, 245.06], [298.3, 244.91], [325.55, 238.74], [460.83, 181.14], [446.21, 181.92], [298.3, 181.9], [325.55, 181.13]]}, {\"category\": \"car\", \"corners_3d\": [[132.2, 273.68], [185.01, 262.72], [359.9, 262.97], [328.33, 274.0], [132.2, 184.49], [185.01, 183.23], [359.9, 183.26], [328.33, 184.53]]}, {\"category\": \"car\", \"corners_3d\": [[25.34, 298.38], [106.22, 280.52], [344.21, 279.77], [303.13, 297.35], [25.34, 194.85], [106.22, 191.72], [344.21, 191.59], [303.13, 194.67]]}, {\"category\": \"car\", \"corners_3d\": [[-599.22, 416.55], [-325.29, 360.26], [142.66, 356.57], [12.67, 410.34], [-599.22, 208.23], [-325.29, 200.06], [142.66, 199.52], [12.67, 207.33]]}]\n```" - } - ] - }, - { - "id": 2902, - "data_source": "Mono3DRefer", - "scene_name": "003864", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[544.34, 208.55], [541.91, 210.55], [429.8, 210.73], [438.22, 208.71], [544.34, 166.26], [541.91, 165.89], [429.8, 165.86], [438.22, 166.23]]}]\n```" - } - ] - }, - { - "id": 2903, - "data_source": "Mono3DRefer", - "scene_name": "003864", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[657.99, 205.72], [689.01, 205.74], [698.01, 209.86], [663.08, 209.84], [657.99, 179.23], [689.01, 179.24], [698.01, 180.04], [663.08, 180.03]]}, {\"category\": \"car\", \"corners_3d\": [[537.43, 212.23], [533.06, 214.61], [425.24, 214.61], [435.77, 212.23], [537.43, 179.26], [533.06, 179.64], [425.24, 179.64], [435.77, 179.26]]}, {\"category\": \"car\", \"corners_3d\": [[505.99, 225.42], [492.29, 229.69], [359.29, 228.73], [382.6, 224.6], [505.99, 179.19], [492.29, 179.7], [359.29, 179.59], [382.6, 179.09]]}, {\"category\": \"car\", \"corners_3d\": [[492.18, 232.85], [478.28, 237.25], [340.61, 235.99], [363.49, 231.76], [492.18, 182.0], [478.28, 182.67], [340.61, 182.47], [363.49, 181.83]]}, {\"category\": \"car\", \"corners_3d\": [[460.83, 238.86], [446.21, 245.06], [298.3, 244.91], [325.55, 238.74], [460.83, 181.14], [446.21, 181.92], [298.3, 181.9], [325.55, 181.13]]}, {\"category\": \"car\", \"corners_3d\": [[132.2, 273.68], [185.01, 262.72], [359.9, 262.97], [328.33, 274.0], [132.2, 184.49], [185.01, 183.23], [359.9, 183.26], [328.33, 184.53]]}, {\"category\": \"car\", \"corners_3d\": [[25.34, 298.38], [106.22, 280.52], [344.21, 279.77], [303.13, 297.35], [25.34, 194.85], [106.22, 191.72], [344.21, 191.59], [303.13, 194.67]]}, {\"category\": \"car\", \"corners_3d\": [[-599.22, 416.55], [-325.29, 360.26], [142.66, 356.57], [12.67, 410.34], [-599.22, 208.23], [-325.29, 200.06], [142.66, 199.52], [12.67, 207.33]]}, {\"category\": \"truck\", \"corners_3d\": [[640.23, 196.38], [672.56, 196.38], [680.85, 199.51], [644.23, 199.51], [640.23, 165.35], [672.56, 165.35], [680.85, 164.35], [644.23, 164.35]]}]\n```" - } - ] - }, - { - "id": 2904, - "data_source": "Mono3DRefer", - "scene_name": "003864", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[544.34, 208.55], [541.91, 210.55], [429.8, 210.73], [438.22, 208.71], [544.34, 166.26], [541.91, 165.89], [429.8, 165.86], [438.22, 166.23]]}, {\"category\": \"truck\", \"corners_3d\": [[640.23, 196.38], [672.56, 196.38], [680.85, 199.51], [644.23, 199.51], [640.23, 165.35], [672.56, 165.35], [680.85, 164.35], [644.23, 164.35]]}]\n```" - } - ] - }, - { - "id": 2905, - "data_source": "Mono3DRefer", - "scene_name": "003864", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[657.99, 205.72], [689.01, 205.74], [698.01, 209.86], [663.08, 209.84], [657.99, 179.23], [689.01, 179.24], [698.01, 180.04], [663.08, 180.03]]}, {\"category\": \"van\", \"corners_3d\": [[544.34, 208.55], [541.91, 210.55], [429.8, 210.73], [438.22, 208.71], [544.34, 166.26], [541.91, 165.89], [429.8, 165.86], [438.22, 166.23]]}, {\"category\": \"car\", \"corners_3d\": [[537.43, 212.23], [533.06, 214.61], [425.24, 214.61], [435.77, 212.23], [537.43, 179.26], [533.06, 179.64], [425.24, 179.64], [435.77, 179.26]]}, {\"category\": \"car\", \"corners_3d\": [[505.99, 225.42], [492.29, 229.69], [359.29, 228.73], [382.6, 224.6], [505.99, 179.19], [492.29, 179.7], [359.29, 179.59], [382.6, 179.09]]}, {\"category\": \"car\", \"corners_3d\": [[492.18, 232.85], [478.28, 237.25], [340.61, 235.99], [363.49, 231.76], [492.18, 182.0], [478.28, 182.67], [340.61, 182.47], [363.49, 181.83]]}, {\"category\": \"car\", \"corners_3d\": [[460.83, 238.86], [446.21, 245.06], [298.3, 244.91], [325.55, 238.74], [460.83, 181.14], [446.21, 181.92], [298.3, 181.9], [325.55, 181.13]]}, {\"category\": \"car\", \"corners_3d\": [[132.2, 273.68], [185.01, 262.72], [359.9, 262.97], [328.33, 274.0], [132.2, 184.49], [185.01, 183.23], [359.9, 183.26], [328.33, 184.53]]}, {\"category\": \"car\", \"corners_3d\": [[25.34, 298.38], [106.22, 280.52], [344.21, 279.77], [303.13, 297.35], [25.34, 194.85], [106.22, 191.72], [344.21, 191.59], [303.13, 194.67]]}, {\"category\": \"car\", \"corners_3d\": [[-599.22, 416.55], [-325.29, 360.26], [142.66, 356.57], [12.67, 410.34], [-599.22, 208.23], [-325.29, 200.06], [142.66, 199.52], [12.67, 207.33]]}]\n```" - } - ] - }, - { - "id": 2906, - "data_source": "Mono3DRefer", - "scene_name": "003864", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[657.99, 205.72], [689.01, 205.74], [698.01, 209.86], [663.08, 209.84], [657.99, 179.23], [689.01, 179.24], [698.01, 180.04], [663.08, 180.03]]}, {\"category\": \"van\", \"corners_3d\": [[544.34, 208.55], [541.91, 210.55], [429.8, 210.73], [438.22, 208.71], [544.34, 166.26], [541.91, 165.89], [429.8, 165.86], [438.22, 166.23]]}, {\"category\": \"car\", \"corners_3d\": [[537.43, 212.23], [533.06, 214.61], [425.24, 214.61], [435.77, 212.23], [537.43, 179.26], [533.06, 179.64], [425.24, 179.64], [435.77, 179.26]]}, {\"category\": \"car\", \"corners_3d\": [[505.99, 225.42], [492.29, 229.69], [359.29, 228.73], [382.6, 224.6], [505.99, 179.19], [492.29, 179.7], [359.29, 179.59], [382.6, 179.09]]}, {\"category\": \"car\", \"corners_3d\": [[492.18, 232.85], [478.28, 237.25], [340.61, 235.99], [363.49, 231.76], [492.18, 182.0], [478.28, 182.67], [340.61, 182.47], [363.49, 181.83]]}, {\"category\": \"car\", \"corners_3d\": [[460.83, 238.86], [446.21, 245.06], [298.3, 244.91], [325.55, 238.74], [460.83, 181.14], [446.21, 181.92], [298.3, 181.9], [325.55, 181.13]]}, {\"category\": \"car\", \"corners_3d\": [[132.2, 273.68], [185.01, 262.72], [359.9, 262.97], [328.33, 274.0], [132.2, 184.49], [185.01, 183.23], [359.9, 183.26], [328.33, 184.53]]}, {\"category\": \"car\", \"corners_3d\": [[25.34, 298.38], [106.22, 280.52], [344.21, 279.77], [303.13, 297.35], [25.34, 194.85], [106.22, 191.72], [344.21, 191.59], [303.13, 194.67]]}, {\"category\": \"car\", \"corners_3d\": [[-599.22, 416.55], [-325.29, 360.26], [142.66, 356.57], [12.67, 410.34], [-599.22, 208.23], [-325.29, 200.06], [142.66, 199.52], [12.67, 207.33]]}, {\"category\": \"truck\", \"corners_3d\": [[640.23, 196.38], [672.56, 196.38], [680.85, 199.51], [644.23, 199.51], [640.23, 165.35], [672.56, 165.35], [680.85, 164.35], [644.23, 164.35]]}]\n```" - } - ] - }, - { - "id": 2907, - "data_source": "Mono3DRefer", - "scene_name": "003867", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[485.55, 202.88], [453.46, 203.28], [446.03, 200.3], [475.1, 199.97], [485.55, 175.63], [453.46, 175.67], [446.03, 175.39], [475.1, 175.36]]}, {\"category\": \"car\", \"corners_3d\": [[499.72, 221.4], [450.95, 221.95], [455.6, 215.26], [497.75, 214.85], [499.72, 172.26], [450.95, 172.25], [455.6, 172.33], [497.75, 172.34]]}]\n```" - } - ] - }, - { - "id": 2908, - "data_source": "Mono3DRefer", - "scene_name": "003874", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[617.96, 230.12], [649.75, 229.95], [660.53, 236.78], [624.97, 236.99], [617.96, 194.07], [649.75, 194.01], [660.53, 196.54], [624.97, 196.62]]}, {\"category\": \"van\", \"corners_3d\": [[547.64, 215.24], [566.37, 215.13], [568.17, 217.8], [548.27, 217.92], [547.64, 192.14], [566.37, 192.09], [568.17, 193.31], [548.27, 193.36]]}]\n```" - } - ] - }, - { - "id": 2909, - "data_source": "Mono3DRefer", - "scene_name": "003874", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[741.6, 230.38], [760.96, 232.16], [706.49, 235.74], [687.73, 233.74], [741.6, 207.34], [760.96, 208.4], [706.49, 210.55], [687.73, 209.35]]}]\n```" - } - ] - }, - { - "id": 2910, - "data_source": "Mono3DRefer", - "scene_name": "003874", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[617.96, 230.12], [649.75, 229.95], [660.53, 236.78], [624.97, 236.99], [617.96, 194.07], [649.75, 194.01], [660.53, 196.54], [624.97, 196.62]]}, {\"category\": \"car\", \"corners_3d\": [[741.6, 230.38], [760.96, 232.16], [706.49, 235.74], [687.73, 233.74], [741.6, 207.34], [760.96, 208.4], [706.49, 210.55], [687.73, 209.35]]}, {\"category\": \"van\", \"corners_3d\": [[547.64, 215.24], [566.37, 215.13], [568.17, 217.8], [548.27, 217.92], [547.64, 192.14], [566.37, 192.09], [568.17, 193.31], [548.27, 193.36]]}]\n```" - } - ] - }, - { - "id": 2911, - "data_source": "Mono3DRefer", - "scene_name": "003880", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[115.79, 520.8], [-177.8, 509.18], [254.02, 339.76], [400.53, 342.43], [115.79, 217.88], [-177.8, 216.64], [254.02, 198.48], [400.53, 198.76]]}]\n```" - } - ] - }, - { - "id": 2912, - "data_source": "Mono3DRefer", - "scene_name": "003880", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[926.51, 277.21], [856.02, 278.85], [830.72, 272.24], [896.89, 270.81], [926.51, 147.8], [856.02, 147.22], [830.72, 149.55], [896.89, 150.06]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[601.51, 224.08], [625.06, 224.08], [625.9, 225.56], [601.53, 225.56], [601.51, 175.58], [625.06, 175.58], [625.9, 175.38], [601.53, 175.38]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[620.67, 223.73], [645.74, 223.79], [646.35, 225.41], [620.33, 225.34], [620.67, 173.41], [645.74, 173.4], [646.35, 173.09], [620.33, 173.1]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[476.49, 227.49], [503.76, 227.42], [500.8, 229.4], [472.35, 229.48], [476.49, 174.98], [503.76, 174.99], [500.8, 174.71], [472.35, 174.7]]}]\n```" - } - ] - }, - { - "id": 2913, - "data_source": "Mono3DRefer", - "scene_name": "003880", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[115.79, 520.8], [-177.8, 509.18], [254.02, 339.76], [400.53, 342.43], [115.79, 217.88], [-177.8, 216.64], [254.02, 198.48], [400.53, 198.76]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[926.51, 277.21], [856.02, 278.85], [830.72, 272.24], [896.89, 270.81], [926.51, 147.8], [856.02, 147.22], [830.72, 149.55], [896.89, 150.06]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[601.51, 224.08], [625.06, 224.08], [625.9, 225.56], [601.53, 225.56], [601.51, 175.58], [625.06, 175.58], [625.9, 175.38], [601.53, 175.38]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[620.67, 223.73], [645.74, 223.79], [646.35, 225.41], [620.33, 225.34], [620.67, 173.41], [645.74, 173.4], [646.35, 173.09], [620.33, 173.1]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[476.49, 227.49], [503.76, 227.42], [500.8, 229.4], [472.35, 229.48], [476.49, 174.98], [503.76, 174.99], [500.8, 174.71], [472.35, 174.7]]}]\n```" - } - ] - }, - { - "id": 2914, - "data_source": "Mono3DRefer", - "scene_name": "003888", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[524.06, 216.73], [485.82, 216.78], [498.01, 211.85], [531.96, 211.81], [524.06, 179.05], [485.82, 179.06], [498.01, 178.36], [531.96, 178.36]]}]\n```" - } - ] - }, - { - "id": 2915, - "data_source": "Mono3DRefer", - "scene_name": "003893", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[498.44, 281.95], [409.59, 282.23], [452.7, 256.74], [520.8, 256.57], [498.44, 206.37], [409.59, 206.45], [452.7, 198.62], [520.8, 198.57]]}, {\"category\": \"car\", \"corners_3d\": [[583.51, 199.41], [560.6, 199.42], [563.02, 197.53], [584.3, 197.51], [583.51, 175.3], [560.6, 175.3], [563.02, 175.12], [584.3, 175.12]]}]\n```" - } - ] - }, - { - "id": 2916, - "data_source": "Mono3DRefer", - "scene_name": "003905", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[207.2, 306.54], [243.8, 297.13], [328.19, 299.11], [297.85, 308.85], [207.2, 156.32], [243.8, 158.13], [328.19, 157.75], [297.85, 155.88]]}]\n```" - } - ] - }, - { - "id": 2917, - "data_source": "Mono3DRefer", - "scene_name": "003906", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[459.72, 230.03], [404.22, 229.99], [435.53, 221.55], [482.84, 221.58], [459.72, 183.6], [404.22, 183.6], [435.53, 182.01], [482.84, 182.01]]}]\n```" - } - ] - }, - { - "id": 2918, - "data_source": "Mono3DRefer", - "scene_name": "003909", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[582.69, 225.26], [528.3, 225.07], [547.27, 216.59], [592.86, 216.73], [582.69, 176.85], [528.3, 176.83], [547.27, 176.19], [592.86, 176.2]]}, {\"category\": \"car\", \"corners_3d\": [[751.21, 214.63], [714.79, 214.43], [711.16, 209.91], [743.59, 210.07], [751.21, 177.64], [714.79, 177.62], [711.16, 177.1], [743.59, 177.12]]}, {\"category\": \"car\", \"corners_3d\": [[623.72, 204.53], [592.72, 204.37], [603.27, 201.15], [631.14, 201.28], [623.72, 177.35], [592.72, 177.32], [603.27, 176.87], [631.14, 176.89]]}, {\"category\": \"car\", \"corners_3d\": [[910.59, 391.6], [1119.62, 398.92], [1765.37, 744.45], [1211.38, 699.85], [910.59, 185.24], [1119.62, 185.66], [1765.37, 205.22], [1211.38, 202.7]]}]\n```" - } - ] - }, - { - "id": 2919, - "data_source": "Mono3DRefer", - "scene_name": "003913", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[570.24, 216.15], [542.55, 216.34], [541.44, 212.7], [566.83, 212.55], [570.24, 187.73], [542.55, 187.8], [541.44, 186.55], [566.83, 186.49]]}]\n```" - } - ] - }, - { - "id": 2920, - "data_source": "Mono3DRefer", - "scene_name": "003932", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-730.45, 1035.35], [-1584.8, 1036.17], [65.31, 386.8], [276.81, 386.75], [-730.45, 281.8], [-1584.8, 281.91], [65.31, 199.88], [276.81, 199.87]]}, {\"category\": \"car\", \"corners_3d\": [[829.46, 277.52], [732.34, 277.68], [699.76, 251.52], [772.71, 251.44], [829.46, 177.76], [732.34, 177.77], [699.76, 176.54], [772.71, 176.54]]}, {\"category\": \"car\", \"corners_3d\": [[515.62, 244.2], [448.65, 244.2], [480.32, 230.11], [534.06, 230.11], [515.62, 186.07], [448.65, 186.07], [480.32, 183.46], [534.06, 183.46]]}, {\"category\": \"car\", \"corners_3d\": [[666.33, 229.39], [717.56, 229.19], [745.37, 240.11], [684.29, 240.39], [666.33, 181.62], [717.56, 181.59], [745.37, 183.28], [684.29, 183.32]]}, {\"category\": \"car\", \"corners_3d\": [[554.8, 213.2], [522.31, 213.23], [528.67, 209.68], [558.29, 209.65], [554.8, 181.86], [522.31, 181.87], [528.67, 181.08], [558.29, 181.07]]}]\n```" - } - ] - }, - { - "id": 2921, - "data_source": "Mono3DRefer", - "scene_name": "003933", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[556.88, 197.37], [536.19, 197.18], [559.18, 195.12], [578.29, 195.29], [556.88, 169.98], [536.19, 170.0], [559.18, 170.24], [578.29, 170.22]]}]\n```" - } - ] - }, - { - "id": 2922, - "data_source": "Mono3DRefer", - "scene_name": "003933", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[326.06, 230.12], [366.03, 231.9], [253.93, 243.41], [210.46, 240.88], [326.06, 186.19], [366.03, 186.6], [253.93, 189.28], [210.46, 188.69]]}, {\"category\": \"car\", \"corners_3d\": [[506.72, 212.25], [537.08, 213.14], [482.51, 218.24], [449.74, 217.12], [506.72, 177.51], [537.08, 177.61], [482.51, 178.22], [449.74, 178.08]]}, {\"category\": \"car\", \"corners_3d\": [[67.78, 241.53], [37.37, 239.82], [177.89, 228.98], [206.27, 230.18], [67.78, 187.38], [37.37, 187.02], [177.89, 184.72], [206.27, 184.98]]}, {\"category\": \"car\", \"corners_3d\": [[425.54, 209.71], [398.29, 209.17], [449.47, 205.18], [474.37, 205.61], [425.54, 173.48], [398.29, 173.47], [449.47, 173.4], [474.37, 173.41]]}, {\"category\": \"car\", \"corners_3d\": [[606.52, 200.18], [631.33, 200.6], [604.93, 203.03], [578.42, 202.53], [606.52, 174.57], [631.33, 174.59], [604.93, 174.75], [578.42, 174.71]]}, {\"category\": \"car\", \"corners_3d\": [[714.03, 187.83], [729.89, 187.96], [721.19, 188.67], [704.68, 188.53], [714.03, 171.97], [729.89, 171.96], [721.19, 171.92], [704.68, 171.93]]}]\n```" - } - ] - }, - { - "id": 2923, - "data_source": "Mono3DRefer", - "scene_name": "003933", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[326.06, 230.12], [366.03, 231.9], [253.93, 243.41], [210.46, 240.88], [326.06, 186.19], [366.03, 186.6], [253.93, 189.28], [210.46, 188.69]]}, {\"category\": \"car\", \"corners_3d\": [[506.72, 212.25], [537.08, 213.14], [482.51, 218.24], [449.74, 217.12], [506.72, 177.51], [537.08, 177.61], [482.51, 178.22], [449.74, 178.08]]}, {\"category\": \"car\", \"corners_3d\": [[67.78, 241.53], [37.37, 239.82], [177.89, 228.98], [206.27, 230.18], [67.78, 187.38], [37.37, 187.02], [177.89, 184.72], [206.27, 184.98]]}, {\"category\": \"car\", \"corners_3d\": [[425.54, 209.71], [398.29, 209.17], [449.47, 205.18], [474.37, 205.61], [425.54, 173.48], [398.29, 173.47], [449.47, 173.4], [474.37, 173.41]]}, {\"category\": \"van\", \"corners_3d\": [[556.88, 197.37], [536.19, 197.18], [559.18, 195.12], [578.29, 195.29], [556.88, 169.98], [536.19, 170.0], [559.18, 170.24], [578.29, 170.22]]}, {\"category\": \"car\", \"corners_3d\": [[606.52, 200.18], [631.33, 200.6], [604.93, 203.03], [578.42, 202.53], [606.52, 174.57], [631.33, 174.59], [604.93, 174.75], [578.42, 174.71]]}, {\"category\": \"car\", \"corners_3d\": [[714.03, 187.83], [729.89, 187.96], [721.19, 188.67], [704.68, 188.53], [714.03, 171.97], [729.89, 171.96], [721.19, 171.92], [704.68, 171.93]]}]\n```" - } - ] - }, - { - "id": 2924, - "data_source": "Mono3DRefer", - "scene_name": "003947", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[370.27, 305.46], [358.46, 312.89], [250.63, 313.89], [268.62, 306.34], [370.27, 148.88], [358.46, 147.0], [250.63, 146.75], [268.62, 148.66]]}]\n```" - } - ] - }, - { - "id": 2925, - "data_source": "Mono3DRefer", - "scene_name": "003952", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[579.53, 214.69], [617.91, 214.69], [618.91, 220.07], [575.6, 220.06], [579.53, 177.7], [617.91, 177.7], [618.91, 178.32], [575.6, 178.32]]}, {\"category\": \"car\", \"corners_3d\": [[182.16, 206.4], [149.35, 206.41], [191.77, 203.31], [221.56, 203.31], [182.16, 179.29], [149.35, 179.29], [191.77, 178.7], [221.56, 178.7]]}, {\"category\": \"car\", \"corners_3d\": [[285.97, 200.18], [260.08, 200.25], [284.06, 198.04], [307.82, 197.99], [285.97, 179.16], [260.08, 179.17], [284.06, 178.66], [307.82, 178.65]]}, {\"category\": \"car\", \"corners_3d\": [[240.96, 203.61], [214.16, 203.64], [245.41, 201.06], [269.94, 201.04], [240.96, 179.2], [214.16, 179.2], [245.41, 178.67], [269.94, 178.66]]}]\n```" - } - ] - }, - { - "id": 2926, - "data_source": "Mono3DRefer", - "scene_name": "003953", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[380.2, 211.97], [391.01, 210.06], [470.77, 210.01], [464.08, 211.92], [380.2, 178.27], [391.01, 178.01], [470.77, 178.0], [464.08, 178.27]]}]\n```" - } - ] - }, - { - "id": 2927, - "data_source": "Mono3DRefer", - "scene_name": "003961", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[418.24, 289.52], [293.09, 289.34], [389.48, 254.59], [477.37, 254.68], [418.24, 172.14], [293.09, 172.14], [389.48, 172.35], [477.37, 172.35]]}, {\"category\": \"car\", \"corners_3d\": [[664.51, 201.82], [695.41, 202.13], [687.89, 204.35], [654.76, 203.99], [664.51, 170.66], [695.41, 170.64], [687.89, 170.47], [654.76, 170.5]]}, {\"category\": \"car\", \"corners_3d\": [[642.51, 189.06], [619.28, 188.93], [632.07, 187.75], [653.68, 187.86], [642.51, 168.72], [619.28, 168.76], [632.07, 169.06], [653.68, 169.03]]}]\n```" - } - ] - }, - { - "id": 2928, - "data_source": "Mono3DRefer", - "scene_name": "003961", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[265.75, 399.28], [20.37, 406.71], [254.53, 300.63], [386.44, 298.38], [265.75, 151.36], [20.37, 150.66], [254.53, 160.73], [386.44, 160.94]]}]\n```" - } - ] - }, - { - "id": 2929, - "data_source": "Mono3DRefer", - "scene_name": "003961", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[265.75, 399.28], [20.37, 406.71], [254.53, 300.63], [386.44, 298.38], [265.75, 151.36], [20.37, 150.66], [254.53, 160.73], [386.44, 160.94]]}, {\"category\": \"car\", \"corners_3d\": [[418.24, 289.52], [293.09, 289.34], [389.48, 254.59], [477.37, 254.68], [418.24, 172.14], [293.09, 172.14], [389.48, 172.35], [477.37, 172.35]]}, {\"category\": \"car\", \"corners_3d\": [[664.51, 201.82], [695.41, 202.13], [687.89, 204.35], [654.76, 203.99], [664.51, 170.66], [695.41, 170.64], [687.89, 170.47], [654.76, 170.5]]}, {\"category\": \"car\", \"corners_3d\": [[642.51, 189.06], [619.28, 188.93], [632.07, 187.75], [653.68, 187.86], [642.51, 168.72], [619.28, 168.76], [632.07, 169.06], [653.68, 169.03]]}]\n```" - } - ] - }, - { - "id": 2930, - "data_source": "Mono3DRefer", - "scene_name": "003965", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[440.92, 253.93], [323.55, 253.81], [425.49, 225.61], [502.04, 225.66], [440.92, 75.85], [323.55, 75.99], [425.49, 109.73], [502.04, 109.67]]}]\n```" - } - ] - }, - { - "id": 2931, - "data_source": "Mono3DRefer", - "scene_name": "003965", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[616.34, 221.35], [672.47, 221.28], [687.22, 230.59], [620.32, 230.69], [616.34, 167.7], [672.47, 167.7], [687.22, 166.71], [620.32, 166.7]]}]\n```" - } - ] - }, - { - "id": 2932, - "data_source": "Mono3DRefer", - "scene_name": "003965", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[596.51, 226.04], [612.39, 225.16], [622.62, 226.13], [606.61, 227.04], [596.51, 158.84], [612.39, 159.07], [622.62, 158.81], [606.61, 158.58]]}]\n```" - } - ] - }, - { - "id": 2933, - "data_source": "Mono3DRefer", - "scene_name": "003965", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[440.92, 253.93], [323.55, 253.81], [425.49, 225.61], [502.04, 225.66], [440.92, 75.85], [323.55, 75.99], [425.49, 109.73], [502.04, 109.67]]}, {\"category\": \"car\", \"corners_3d\": [[616.34, 221.35], [672.47, 221.28], [687.22, 230.59], [620.32, 230.69], [616.34, 167.7], [672.47, 167.7], [687.22, 166.71], [620.32, 166.7]]}]\n```" - } - ] - }, - { - "id": 2934, - "data_source": "Mono3DRefer", - "scene_name": "003965", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[440.92, 253.93], [323.55, 253.81], [425.49, 225.61], [502.04, 225.66], [440.92, 75.85], [323.55, 75.99], [425.49, 109.73], [502.04, 109.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[596.51, 226.04], [612.39, 225.16], [622.62, 226.13], [606.61, 227.04], [596.51, 158.84], [612.39, 159.07], [622.62, 158.81], [606.61, 158.58]]}]\n```" - } - ] - }, - { - "id": 2935, - "data_source": "Mono3DRefer", - "scene_name": "003965", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[616.34, 221.35], [672.47, 221.28], [687.22, 230.59], [620.32, 230.69], [616.34, 167.7], [672.47, 167.7], [687.22, 166.71], [620.32, 166.7]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[596.51, 226.04], [612.39, 225.16], [622.62, 226.13], [606.61, 227.04], [596.51, 158.84], [612.39, 159.07], [622.62, 158.81], [606.61, 158.58]]}]\n```" - } - ] - }, - { - "id": 2936, - "data_source": "Mono3DRefer", - "scene_name": "003965", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[440.92, 253.93], [323.55, 253.81], [425.49, 225.61], [502.04, 225.66], [440.92, 75.85], [323.55, 75.99], [425.49, 109.73], [502.04, 109.67]]}, {\"category\": \"car\", \"corners_3d\": [[616.34, 221.35], [672.47, 221.28], [687.22, 230.59], [620.32, 230.69], [616.34, 167.7], [672.47, 167.7], [687.22, 166.71], [620.32, 166.7]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[596.51, 226.04], [612.39, 225.16], [622.62, 226.13], [606.61, 227.04], [596.51, 158.84], [612.39, 159.07], [622.62, 158.81], [606.61, 158.58]]}]\n```" - } - ] - }, - { - "id": 2937, - "data_source": "Mono3DRefer", - "scene_name": "003966", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[476.04, 223.22], [426.74, 223.22], [451.01, 216.51], [493.74, 216.51], [476.04, 182.1], [426.74, 182.1], [451.01, 180.87], [493.74, 180.87]]}]\n```" - } - ] - }, - { - "id": 2938, - "data_source": "Mono3DRefer", - "scene_name": "003967", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[601.99, 298.19], [602.92, 307.3], [503.53, 307.66], [509.76, 298.5], [601.99, 151.31], [602.92, 149.05], [503.53, 148.96], [509.76, 151.24]]}]\n```" - } - ] - }, - { - "id": 2939, - "data_source": "Mono3DRefer", - "scene_name": "003969", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[908.4, 276.43], [941.27, 281.0], [866.48, 288.85], [834.82, 283.56], [908.4, 176.2], [941.27, 175.99], [866.48, 175.64], [834.82, 175.88]]}]\n```" - } - ] - }, - { - "id": 2940, - "data_source": "Mono3DRefer", - "scene_name": "003969", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[240.18, 263.73], [210.35, 266.29], [194.49, 263.12], [223.71, 260.74], [240.18, 168.3], [210.35, 167.92], [194.49, 168.39], [223.71, 168.74]]}]\n```" - } - ] - }, - { - "id": 2941, - "data_source": "Mono3DRefer", - "scene_name": "003969", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[908.4, 276.43], [941.27, 281.0], [866.48, 288.85], [834.82, 283.56], [908.4, 176.2], [941.27, 175.99], [866.48, 175.64], [834.82, 175.88]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[240.18, 263.73], [210.35, 266.29], [194.49, 263.12], [223.71, 260.74], [240.18, 168.3], [210.35, 167.92], [194.49, 168.39], [223.71, 168.74]]}]\n```" - } - ] - }, - { - "id": 2942, - "data_source": "Mono3DRefer", - "scene_name": "003972", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[588.16, 188.42], [611.28, 188.41], [614.04, 191.16], [586.82, 191.17], [588.16, 163.35], [611.28, 163.36], [614.04, 161.68], [586.82, 161.67]]}]\n```" - } - ] - }, - { - "id": 2943, - "data_source": "Mono3DRefer", - "scene_name": "003972", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[437.04, 201.25], [477.8, 201.25], [403.17, 217.27], [339.42, 217.26], [437.04, 152.54], [477.8, 152.54], [403.17, 141.08], [339.42, 141.08]]}]\n```" - } - ] - }, - { - "id": 2944, - "data_source": "Mono3DRefer", - "scene_name": "003972", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[555.96, 190.39], [573.0, 190.4], [571.07, 191.16], [553.29, 191.16], [555.96, 174.03], [573.0, 174.03], [571.07, 174.08], [553.29, 174.08]]}]\n```" - } - ] - }, - { - "id": 2945, - "data_source": "Mono3DRefer", - "scene_name": "003972", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[588.16, 188.42], [611.28, 188.41], [614.04, 191.16], [586.82, 191.17], [588.16, 163.35], [611.28, 163.36], [614.04, 161.68], [586.82, 161.67]]}, {\"category\": \"tram\", \"corners_3d\": [[437.04, 201.25], [477.8, 201.25], [403.17, 217.27], [339.42, 217.26], [437.04, 152.54], [477.8, 152.54], [403.17, 141.08], [339.42, 141.08]]}]\n```" - } - ] - }, - { - "id": 2946, - "data_source": "Mono3DRefer", - "scene_name": "003972", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[555.96, 190.39], [573.0, 190.4], [571.07, 191.16], [553.29, 191.16], [555.96, 174.03], [573.0, 174.03], [571.07, 174.08], [553.29, 174.08]]}, {\"category\": \"bus\", \"corners_3d\": [[588.16, 188.42], [611.28, 188.41], [614.04, 191.16], [586.82, 191.17], [588.16, 163.35], [611.28, 163.36], [614.04, 161.68], [586.82, 161.67]]}]\n```" - } - ] - }, - { - "id": 2947, - "data_source": "Mono3DRefer", - "scene_name": "003972", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[555.96, 190.39], [573.0, 190.4], [571.07, 191.16], [553.29, 191.16], [555.96, 174.03], [573.0, 174.03], [571.07, 174.08], [553.29, 174.08]]}, {\"category\": \"tram\", \"corners_3d\": [[437.04, 201.25], [477.8, 201.25], [403.17, 217.27], [339.42, 217.26], [437.04, 152.54], [477.8, 152.54], [403.17, 141.08], [339.42, 141.08]]}]\n```" - } - ] - }, - { - "id": 2948, - "data_source": "Mono3DRefer", - "scene_name": "003972", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus, tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[555.96, 190.39], [573.0, 190.4], [571.07, 191.16], [553.29, 191.16], [555.96, 174.03], [573.0, 174.03], [571.07, 174.08], [553.29, 174.08]]}, {\"category\": \"bus\", \"corners_3d\": [[588.16, 188.42], [611.28, 188.41], [614.04, 191.16], [586.82, 191.17], [588.16, 163.35], [611.28, 163.36], [614.04, 161.68], [586.82, 161.67]]}, {\"category\": \"tram\", \"corners_3d\": [[437.04, 201.25], [477.8, 201.25], [403.17, 217.27], [339.42, 217.26], [437.04, 152.54], [477.8, 152.54], [403.17, 141.08], [339.42, 141.08]]}]\n```" - } - ] - }, - { - "id": 2949, - "data_source": "Mono3DRefer", - "scene_name": "003989", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[788.59, 295.58], [779.84, 290.08], [855.07, 290.1], [867.6, 295.6], [788.59, 155.39], [779.84, 156.59], [855.07, 156.58], [867.6, 155.38]]}]\n```" - } - ] - }, - { - "id": 2950, - "data_source": "Mono3DRefer", - "scene_name": "003990", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[592.23, 237.16], [660.03, 237.16], [673.55, 254.59], [587.38, 254.58], [592.23, 177.62], [660.03, 177.62], [673.55, 178.91], [587.38, 178.91]]}, {\"category\": \"car\", \"corners_3d\": [[510.26, 224.13], [564.98, 224.25], [549.11, 236.47], [481.46, 236.28], [510.26, 171.87], [564.98, 171.87], [549.11, 171.63], [481.46, 171.64]]}, {\"category\": \"car\", \"corners_3d\": [[616.77, 214.16], [662.08, 214.29], [665.4, 222.98], [610.62, 222.78], [616.77, 171.73], [662.08, 171.73], [665.4, 171.5], [610.62, 171.5]]}, {\"category\": \"car\", \"corners_3d\": [[1192.49, 197.64], [1260.17, 199.78], [1148.24, 201.32], [1086.39, 198.94], [1192.49, 140.27], [1260.17, 137.45], [1148.24, 135.43], [1086.39, 138.56]]}]\n```" - } - ] - }, - { - "id": 2951, - "data_source": "Mono3DRefer", - "scene_name": "004005", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[564.53, 194.07], [546.77, 194.08], [549.64, 192.8], [566.34, 192.79], [564.53, 177.26], [546.77, 177.26], [549.64, 176.99], [566.34, 176.99]]}, {\"category\": \"car\", \"corners_3d\": [[543.1, 207.87], [513.92, 207.86], [523.49, 204.58], [549.94, 204.59], [543.1, 176.97], [513.92, 176.97], [523.49, 176.58], [549.94, 176.58]]}]\n```" - } - ] - }, - { - "id": 2952, - "data_source": "Mono3DRefer", - "scene_name": "004038", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[2585.15, 800.49], [1947.74, 840.99], [941.1, 354.36], [1139.61, 351.23], [2585.15, 51.33], [1947.74, 43.49], [941.1, 137.71], [1139.61, 138.32]]}, {\"category\": \"car\", \"corners_3d\": [[-96.13, 599.01], [-454.85, 605.35], [137.79, 359.41], [290.05, 358.22], [-96.13, 264.93], [-454.85, 266.3], [137.79, 213.16], [290.05, 212.91]]}, {\"category\": \"car\", \"corners_3d\": [[421.59, 287.82], [323.87, 288.45], [384.73, 260.99], [459.08, 260.63], [421.59, 208.45], [323.87, 208.65], [384.73, 200.15], [459.08, 200.03]]}, {\"category\": \"car\", \"corners_3d\": [[481.05, 249.48], [412.89, 249.71], [444.65, 235.71], [500.36, 235.56], [481.05, 182.68], [412.89, 182.7], [444.65, 180.91], [500.36, 180.89]]}, {\"category\": \"car\", \"corners_3d\": [[606.28, 206.3], [636.32, 206.25], [641.52, 209.4], [608.65, 209.46], [606.28, 177.57], [636.32, 177.56], [641.52, 178.01], [608.65, 178.02]]}]\n```" - } - ] - }, - { - "id": 2953, - "data_source": "Mono3DRefer", - "scene_name": "004038", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[558.67, 206.53], [522.85, 206.42], [542.07, 201.7], [572.9, 201.78], [558.67, 151.8], [522.85, 151.87], [542.07, 154.82], [572.9, 154.77]]}]\n```" - } - ] - }, - { - "id": 2954, - "data_source": "Mono3DRefer", - "scene_name": "004038", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[597.15, 208.44], [605.21, 208.13], [610.51, 208.46], [602.42, 208.78], [597.15, 174.07], [605.21, 174.06], [610.51, 174.07], [602.42, 174.09]]}]\n```" - } - ] - }, - { - "id": 2955, - "data_source": "Mono3DRefer", - "scene_name": "004038", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[2585.15, 800.49], [1947.74, 840.99], [941.1, 354.36], [1139.61, 351.23], [2585.15, 51.33], [1947.74, 43.49], [941.1, 137.71], [1139.61, 138.32]]}, {\"category\": \"car\", \"corners_3d\": [[-96.13, 599.01], [-454.85, 605.35], [137.79, 359.41], [290.05, 358.22], [-96.13, 264.93], [-454.85, 266.3], [137.79, 213.16], [290.05, 212.91]]}, {\"category\": \"car\", \"corners_3d\": [[421.59, 287.82], [323.87, 288.45], [384.73, 260.99], [459.08, 260.63], [421.59, 208.45], [323.87, 208.65], [384.73, 200.15], [459.08, 200.03]]}, {\"category\": \"car\", \"corners_3d\": [[481.05, 249.48], [412.89, 249.71], [444.65, 235.71], [500.36, 235.56], [481.05, 182.68], [412.89, 182.7], [444.65, 180.91], [500.36, 180.89]]}, {\"category\": \"truck\", \"corners_3d\": [[558.67, 206.53], [522.85, 206.42], [542.07, 201.7], [572.9, 201.78], [558.67, 151.8], [522.85, 151.87], [542.07, 154.82], [572.9, 154.77]]}, {\"category\": \"car\", \"corners_3d\": [[606.28, 206.3], [636.32, 206.25], [641.52, 209.4], [608.65, 209.46], [606.28, 177.57], [636.32, 177.56], [641.52, 178.01], [608.65, 178.02]]}]\n```" - } - ] - }, - { - "id": 2956, - "data_source": "Mono3DRefer", - "scene_name": "004038", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[2585.15, 800.49], [1947.74, 840.99], [941.1, 354.36], [1139.61, 351.23], [2585.15, 51.33], [1947.74, 43.49], [941.1, 137.71], [1139.61, 138.32]]}, {\"category\": \"car\", \"corners_3d\": [[-96.13, 599.01], [-454.85, 605.35], [137.79, 359.41], [290.05, 358.22], [-96.13, 264.93], [-454.85, 266.3], [137.79, 213.16], [290.05, 212.91]]}, {\"category\": \"car\", \"corners_3d\": [[421.59, 287.82], [323.87, 288.45], [384.73, 260.99], [459.08, 260.63], [421.59, 208.45], [323.87, 208.65], [384.73, 200.15], [459.08, 200.03]]}, {\"category\": \"car\", \"corners_3d\": [[481.05, 249.48], [412.89, 249.71], [444.65, 235.71], [500.36, 235.56], [481.05, 182.68], [412.89, 182.7], [444.65, 180.91], [500.36, 180.89]]}, {\"category\": \"car\", \"corners_3d\": [[606.28, 206.3], [636.32, 206.25], [641.52, 209.4], [608.65, 209.46], [606.28, 177.57], [636.32, 177.56], [641.52, 178.01], [608.65, 178.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[597.15, 208.44], [605.21, 208.13], [610.51, 208.46], [602.42, 208.78], [597.15, 174.07], [605.21, 174.06], [610.51, 174.07], [602.42, 174.09]]}]\n```" - } - ] - }, - { - "id": 2957, - "data_source": "Mono3DRefer", - "scene_name": "004038", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[558.67, 206.53], [522.85, 206.42], [542.07, 201.7], [572.9, 201.78], [558.67, 151.8], [522.85, 151.87], [542.07, 154.82], [572.9, 154.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[597.15, 208.44], [605.21, 208.13], [610.51, 208.46], [602.42, 208.78], [597.15, 174.07], [605.21, 174.06], [610.51, 174.07], [602.42, 174.09]]}]\n```" - } - ] - }, - { - "id": 2958, - "data_source": "Mono3DRefer", - "scene_name": "004038", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[2585.15, 800.49], [1947.74, 840.99], [941.1, 354.36], [1139.61, 351.23], [2585.15, 51.33], [1947.74, 43.49], [941.1, 137.71], [1139.61, 138.32]]}, {\"category\": \"car\", \"corners_3d\": [[-96.13, 599.01], [-454.85, 605.35], [137.79, 359.41], [290.05, 358.22], [-96.13, 264.93], [-454.85, 266.3], [137.79, 213.16], [290.05, 212.91]]}, {\"category\": \"car\", \"corners_3d\": [[421.59, 287.82], [323.87, 288.45], [384.73, 260.99], [459.08, 260.63], [421.59, 208.45], [323.87, 208.65], [384.73, 200.15], [459.08, 200.03]]}, {\"category\": \"car\", \"corners_3d\": [[481.05, 249.48], [412.89, 249.71], [444.65, 235.71], [500.36, 235.56], [481.05, 182.68], [412.89, 182.7], [444.65, 180.91], [500.36, 180.89]]}, {\"category\": \"truck\", \"corners_3d\": [[558.67, 206.53], [522.85, 206.42], [542.07, 201.7], [572.9, 201.78], [558.67, 151.8], [522.85, 151.87], [542.07, 154.82], [572.9, 154.77]]}, {\"category\": \"car\", \"corners_3d\": [[606.28, 206.3], [636.32, 206.25], [641.52, 209.4], [608.65, 209.46], [606.28, 177.57], [636.32, 177.56], [641.52, 178.01], [608.65, 178.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[597.15, 208.44], [605.21, 208.13], [610.51, 208.46], [602.42, 208.78], [597.15, 174.07], [605.21, 174.06], [610.51, 174.07], [602.42, 174.09]]}]\n```" - } - ] - }, - { - "id": 2959, - "data_source": "Mono3DRefer", - "scene_name": "004040", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[423.64, 203.27], [468.17, 203.29], [370.77, 223.16], [297.31, 223.1], [423.64, 149.95], [468.17, 149.93], [370.77, 134.97], [297.31, 135.01]]}]\n```" - } - ] - }, - { - "id": 2960, - "data_source": "Mono3DRefer", - "scene_name": "004040", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[557.79, 189.71], [574.64, 189.71], [573.12, 190.43], [555.54, 190.43], [557.79, 173.53], [574.64, 173.53], [573.12, 173.56], [555.54, 173.56]]}]\n```" - } - ] - }, - { - "id": 2961, - "data_source": "Mono3DRefer", - "scene_name": "004040", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[590.0, 187.9], [613.13, 187.89], [616.21, 190.55], [588.99, 190.56], [590.0, 162.82], [613.13, 162.83], [616.21, 161.05], [588.99, 161.04]]}]\n```" - } - ] - }, - { - "id": 2962, - "data_source": "Mono3DRefer", - "scene_name": "004040", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[557.79, 189.71], [574.64, 189.71], [573.12, 190.43], [555.54, 190.43], [557.79, 173.53], [574.64, 173.53], [573.12, 173.56], [555.54, 173.56]]}, {\"category\": \"tram\", \"corners_3d\": [[423.64, 203.27], [468.17, 203.29], [370.77, 223.16], [297.31, 223.1], [423.64, 149.95], [468.17, 149.93], [370.77, 134.97], [297.31, 135.01]]}]\n```" - } - ] - }, - { - "id": 2963, - "data_source": "Mono3DRefer", - "scene_name": "004040", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[590.0, 187.9], [613.13, 187.89], [616.21, 190.55], [588.99, 190.56], [590.0, 162.82], [613.13, 162.83], [616.21, 161.05], [588.99, 161.04]]}, {\"category\": \"tram\", \"corners_3d\": [[423.64, 203.27], [468.17, 203.29], [370.77, 223.16], [297.31, 223.1], [423.64, 149.95], [468.17, 149.93], [370.77, 134.97], [297.31, 135.01]]}]\n```" - } - ] - }, - { - "id": 2964, - "data_source": "Mono3DRefer", - "scene_name": "004040", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[557.79, 189.71], [574.64, 189.71], [573.12, 190.43], [555.54, 190.43], [557.79, 173.53], [574.64, 173.53], [573.12, 173.56], [555.54, 173.56]]}, {\"category\": \"bus\", \"corners_3d\": [[590.0, 187.9], [613.13, 187.89], [616.21, 190.55], [588.99, 190.56], [590.0, 162.82], [613.13, 162.83], [616.21, 161.05], [588.99, 161.04]]}]\n```" - } - ] - }, - { - "id": 2965, - "data_source": "Mono3DRefer", - "scene_name": "004040", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, car, bus). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[557.79, 189.71], [574.64, 189.71], [573.12, 190.43], [555.54, 190.43], [557.79, 173.53], [574.64, 173.53], [573.12, 173.56], [555.54, 173.56]]}, {\"category\": \"bus\", \"corners_3d\": [[590.0, 187.9], [613.13, 187.89], [616.21, 190.55], [588.99, 190.56], [590.0, 162.82], [613.13, 162.83], [616.21, 161.05], [588.99, 161.04]]}, {\"category\": \"tram\", \"corners_3d\": [[423.64, 203.27], [468.17, 203.29], [370.77, 223.16], [297.31, 223.1], [423.64, 149.95], [468.17, 149.93], [370.77, 134.97], [297.31, 135.01]]}]\n```" - } - ] - }, - { - "id": 2966, - "data_source": "Mono3DRefer", - "scene_name": "004046", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[248.86, 200.67], [221.19, 200.69], [251.11, 198.5], [276.59, 198.49], [248.86, 177.94], [221.19, 177.94], [251.11, 177.54], [276.59, 177.54]]}, {\"category\": \"car\", \"corners_3d\": [[322.72, 195.93], [300.13, 195.98], [318.49, 194.33], [339.44, 194.29], [322.72, 177.51], [300.13, 177.52], [318.49, 177.19], [339.44, 177.18]]}, {\"category\": \"car\", \"corners_3d\": [[290.24, 199.38], [267.3, 199.41], [290.57, 197.47], [311.83, 197.45], [290.24, 178.43], [267.3, 178.44], [290.57, 178.03], [311.83, 178.03]]}, {\"category\": \"car\", \"corners_3d\": [[575.96, 217.85], [617.01, 217.85], [617.96, 224.09], [571.23, 224.08], [575.96, 178.29], [617.01, 178.29], [617.96, 179.04], [571.23, 179.04]]}]\n```" - } - ] - }, - { - "id": 2967, - "data_source": "Mono3DRefer", - "scene_name": "004050", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[406.58, 208.63], [415.13, 207.06], [486.78, 207.02], [481.53, 208.59], [406.58, 178.51], [415.13, 178.26], [486.78, 178.25], [481.53, 178.5]]}]\n```" - } - ] - }, - { - "id": 2968, - "data_source": "Mono3DRefer", - "scene_name": "004073", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[773.04, 296.3], [787.29, 305.31], [687.93, 305.66], [680.84, 296.6], [773.04, 148.75], [787.29, 146.27], [687.93, 146.18], [680.84, 148.66]]}]\n```" - } - ] - }, - { - "id": 2969, - "data_source": "Mono3DRefer", - "scene_name": "004135", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[266.38, 289.18], [182.78, 287.16], [300.84, 261.59], [367.09, 262.81], [266.38, 198.88], [182.78, 198.43], [300.84, 192.71], [367.09, 192.98]]}, {\"category\": \"car\", \"corners_3d\": [[383.97, 259.49], [320.83, 258.38], [393.93, 242.34], [445.87, 243.07], [383.97, 177.51], [320.83, 177.45], [393.93, 176.59], [445.87, 176.63]]}, {\"category\": \"car\", \"corners_3d\": [[714.02, 226.18], [763.07, 226.33], [785.9, 236.17], [727.77, 235.96], [714.02, 180.13], [763.07, 180.15], [785.9, 181.49], [727.77, 181.46]]}, {\"category\": \"car\", \"corners_3d\": [[680.59, 202.04], [659.35, 201.78], [669.84, 200.02], [689.87, 200.25], [680.59, 182.1], [659.35, 182.02], [669.84, 181.46], [689.87, 181.54]]}, {\"category\": \"car\", \"corners_3d\": [[722.16, 219.05], [759.59, 219.45], [763.37, 226.35], [720.42, 225.82], [722.16, 168.93], [759.59, 168.9], [763.37, 168.31], [720.42, 168.36]]}, {\"category\": \"car\", \"corners_3d\": [[692.57, 204.46], [717.1, 204.69], [713.31, 207.26], [686.85, 206.99], [692.57, 181.5], [717.1, 181.56], [713.31, 182.26], [686.85, 182.19]]}]\n```" - } - ] - }, - { - "id": 2970, - "data_source": "Mono3DRefer", - "scene_name": "004149", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[568.76, 191.29], [608.33, 191.27], [615.05, 193.09], [568.11, 193.12], [568.76, 142.05], [608.33, 142.14], [615.05, 134.82], [568.11, 134.7]]}]\n```" - } - ] - }, - { - "id": 2971, - "data_source": "Mono3DRefer", - "scene_name": "004149", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[476.81, 242.46], [436.48, 242.39], [455.32, 235.93], [491.39, 235.99], [476.81, 160.47], [436.48, 160.49], [455.32, 162.72], [491.39, 162.7]]}, {\"category\": \"cyclist\", \"corners_3d\": [[400.86, 258.53], [361.76, 258.61], [392.6, 247.98], [426.29, 247.92], [400.86, 157.29], [361.76, 157.27], [392.6, 160.61], [426.29, 160.63]]}]\n```" - } - ] - }, - { - "id": 2972, - "data_source": "Mono3DRefer", - "scene_name": "004149", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[742.98, 264.21], [800.57, 263.75], [822.69, 271.02], [760.11, 271.58], [742.98, 160.07], [800.57, 160.19], [822.69, 158.3], [760.11, 158.16]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[655.91, 217.74], [687.66, 217.67], [690.59, 218.59], [658.04, 218.66], [655.91, 157.88], [687.66, 157.92], [690.59, 157.32], [658.04, 157.28]]}]\n```" - } - ] - }, - { - "id": 2973, - "data_source": "Mono3DRefer", - "scene_name": "004149", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[476.81, 242.46], [436.48, 242.39], [455.32, 235.93], [491.39, 235.99], [476.81, 160.47], [436.48, 160.49], [455.32, 162.72], [491.39, 162.7]]}, {\"category\": \"cyclist\", \"corners_3d\": [[400.86, 258.53], [361.76, 258.61], [392.6, 247.98], [426.29, 247.92], [400.86, 157.29], [361.76, 157.27], [392.6, 160.61], [426.29, 160.63]]}, {\"category\": \"van\", \"corners_3d\": [[568.76, 191.29], [608.33, 191.27], [615.05, 193.09], [568.11, 193.12], [568.76, 142.05], [608.33, 142.14], [615.05, 134.82], [568.11, 134.7]]}]\n```" - } - ] - }, - { - "id": 2974, - "data_source": "Mono3DRefer", - "scene_name": "004149", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[742.98, 264.21], [800.57, 263.75], [822.69, 271.02], [760.11, 271.58], [742.98, 160.07], [800.57, 160.19], [822.69, 158.3], [760.11, 158.16]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[655.91, 217.74], [687.66, 217.67], [690.59, 218.59], [658.04, 218.66], [655.91, 157.88], [687.66, 157.92], [690.59, 157.32], [658.04, 157.28]]}, {\"category\": \"van\", \"corners_3d\": [[568.76, 191.29], [608.33, 191.27], [615.05, 193.09], [568.11, 193.12], [568.76, 142.05], [608.33, 142.14], [615.05, 134.82], [568.11, 134.7]]}]\n```" - } - ] - }, - { - "id": 2975, - "data_source": "Mono3DRefer", - "scene_name": "004149", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[742.98, 264.21], [800.57, 263.75], [822.69, 271.02], [760.11, 271.58], [742.98, 160.07], [800.57, 160.19], [822.69, 158.3], [760.11, 158.16]]}, {\"category\": \"cyclist\", \"corners_3d\": [[476.81, 242.46], [436.48, 242.39], [455.32, 235.93], [491.39, 235.99], [476.81, 160.47], [436.48, 160.49], [455.32, 162.72], [491.39, 162.7]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[655.91, 217.74], [687.66, 217.67], [690.59, 218.59], [658.04, 218.66], [655.91, 157.88], [687.66, 157.92], [690.59, 157.32], [658.04, 157.28]]}, {\"category\": \"cyclist\", \"corners_3d\": [[400.86, 258.53], [361.76, 258.61], [392.6, 247.98], [426.29, 247.92], [400.86, 157.29], [361.76, 157.27], [392.6, 160.61], [426.29, 160.63]]}]\n```" - } - ] - }, - { - "id": 2976, - "data_source": "Mono3DRefer", - "scene_name": "004149", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[742.98, 264.21], [800.57, 263.75], [822.69, 271.02], [760.11, 271.58], [742.98, 160.07], [800.57, 160.19], [822.69, 158.3], [760.11, 158.16]]}, {\"category\": \"cyclist\", \"corners_3d\": [[476.81, 242.46], [436.48, 242.39], [455.32, 235.93], [491.39, 235.99], [476.81, 160.47], [436.48, 160.49], [455.32, 162.72], [491.39, 162.7]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[655.91, 217.74], [687.66, 217.67], [690.59, 218.59], [658.04, 218.66], [655.91, 157.88], [687.66, 157.92], [690.59, 157.32], [658.04, 157.28]]}, {\"category\": \"cyclist\", \"corners_3d\": [[400.86, 258.53], [361.76, 258.61], [392.6, 247.98], [426.29, 247.92], [400.86, 157.29], [361.76, 157.27], [392.6, 160.61], [426.29, 160.63]]}, {\"category\": \"van\", \"corners_3d\": [[568.76, 191.29], [608.33, 191.27], [615.05, 193.09], [568.11, 193.12], [568.76, 142.05], [608.33, 142.14], [615.05, 134.82], [568.11, 134.7]]}]\n```" - } - ] - }, - { - "id": 2977, - "data_source": "Mono3DRefer", - "scene_name": "004151", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[574.46, 200.48], [553.02, 200.48], [556.86, 198.59], [576.82, 198.58], [574.46, 181.65], [553.02, 181.65], [556.86, 181.04], [576.82, 181.04]]}]\n```" - } - ] - }, - { - "id": 2978, - "data_source": "Mono3DRefer", - "scene_name": "004158", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[604.08, 209.98], [591.55, 209.97], [592.44, 208.63], [604.52, 208.63], [604.08, 168.75], [591.55, 168.75], [592.44, 168.89], [604.52, 168.89]]}]\n```" - } - ] - }, - { - "id": 2979, - "data_source": "Mono3DRefer", - "scene_name": "004158", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[222.29, 341.6], [389.46, 341.24], [151.29, 529.12], [-204.3, 530.73], [222.29, 188.45], [389.46, 188.41], [151.29, 205.77], [-204.3, 205.92]]}, {\"category\": \"car\", \"corners_3d\": [[533.23, 223.85], [482.48, 223.56], [510.28, 215.92], [553.48, 216.12], [533.23, 169.99], [482.48, 170.01], [510.28, 170.44], [553.48, 170.43]]}, {\"category\": \"car\", \"corners_3d\": [[694.49, 257.21], [777.68, 257.41], [810.99, 275.8], [709.66, 275.5], [694.49, 174.4], [777.68, 174.4], [810.99, 174.74], [709.66, 174.73]]}, {\"category\": \"car\", \"corners_3d\": [[676.47, 233.68], [737.15, 234.04], [752.72, 246.5], [679.65, 245.98], [676.47, 181.91], [737.15, 181.96], [752.72, 183.82], [679.65, 183.74]]}, {\"category\": \"car\", \"corners_3d\": [[534.9, 206.23], [569.63, 206.29], [561.26, 210.33], [522.36, 210.25], [534.9, 177.02], [569.63, 177.03], [561.26, 177.53], [522.36, 177.52]]}, {\"category\": \"car\", \"corners_3d\": [[657.52, 196.23], [633.22, 196.22], [632.51, 194.52], [655.04, 194.54], [657.52, 173.75], [633.22, 173.75], [632.51, 173.69], [655.04, 173.69]]}, {\"category\": \"car\", \"corners_3d\": [[228.69, 234.04], [260.74, 228.91], [425.17, 228.93], [408.17, 234.07], [228.69, 169.6], [260.74, 169.87], [425.17, 169.87], [408.17, 169.6]]}]\n```" - } - ] - }, - { - "id": 2980, - "data_source": "Mono3DRefer", - "scene_name": "004158", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[222.29, 341.6], [389.46, 341.24], [151.29, 529.12], [-204.3, 530.73], [222.29, 188.45], [389.46, 188.41], [151.29, 205.77], [-204.3, 205.92]]}, {\"category\": \"car\", \"corners_3d\": [[533.23, 223.85], [482.48, 223.56], [510.28, 215.92], [553.48, 216.12], [533.23, 169.99], [482.48, 170.01], [510.28, 170.44], [553.48, 170.43]]}, {\"category\": \"car\", \"corners_3d\": [[694.49, 257.21], [777.68, 257.41], [810.99, 275.8], [709.66, 275.5], [694.49, 174.4], [777.68, 174.4], [810.99, 174.74], [709.66, 174.73]]}, {\"category\": \"cyclist\", \"corners_3d\": [[604.08, 209.98], [591.55, 209.97], [592.44, 208.63], [604.52, 208.63], [604.08, 168.75], [591.55, 168.75], [592.44, 168.89], [604.52, 168.89]]}, {\"category\": \"car\", \"corners_3d\": [[676.47, 233.68], [737.15, 234.04], [752.72, 246.5], [679.65, 245.98], [676.47, 181.91], [737.15, 181.96], [752.72, 183.82], [679.65, 183.74]]}, {\"category\": \"car\", \"corners_3d\": [[534.9, 206.23], [569.63, 206.29], [561.26, 210.33], [522.36, 210.25], [534.9, 177.02], [569.63, 177.03], [561.26, 177.53], [522.36, 177.52]]}, {\"category\": \"car\", \"corners_3d\": [[657.52, 196.23], [633.22, 196.22], [632.51, 194.52], [655.04, 194.54], [657.52, 173.75], [633.22, 173.75], [632.51, 173.69], [655.04, 173.69]]}, {\"category\": \"car\", \"corners_3d\": [[228.69, 234.04], [260.74, 228.91], [425.17, 228.93], [408.17, 234.07], [228.69, 169.6], [260.74, 169.87], [425.17, 169.87], [408.17, 169.6]]}]\n```" - } - ] - }, - { - "id": 2981, - "data_source": "Mono3DRefer", - "scene_name": "004159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[409.11, 227.97], [372.74, 228.89], [370.53, 224.33], [404.01, 223.55], [409.11, 195.37], [372.74, 195.75], [370.53, 193.88], [404.01, 193.56]]}, {\"category\": \"car\", \"corners_3d\": [[106.12, 649.55], [-392.42, 668.78], [212.32, 355.75], [391.92, 353.07], [106.12, 251.32], [-392.42, 254.48], [212.32, 202.96], [391.92, 202.52]]}]\n```" - } - ] - }, - { - "id": 2982, - "data_source": "Mono3DRefer", - "scene_name": "004165", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[200.95, 299.19], [220.89, 293.34], [294.38, 293.36], [278.25, 299.22], [200.95, 159.63], [220.89, 160.66], [294.38, 160.66], [278.25, 159.63]]}]\n```" - } - ] - }, - { - "id": 2983, - "data_source": "Mono3DRefer", - "scene_name": "004174", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[520.95, 259.46], [434.31, 259.47], [476.24, 238.68], [542.09, 238.67], [520.95, 179.54], [434.31, 179.54], [476.24, 177.93], [542.09, 177.93]]}, {\"category\": \"car\", \"corners_3d\": [[671.87, 220.82], [714.35, 220.73], [738.54, 229.94], [687.94, 230.06], [671.87, 176.19], [714.35, 176.19], [738.54, 176.83], [687.94, 176.83]]}, {\"category\": \"car\", \"corners_3d\": [[658.42, 210.96], [694.22, 210.87], [711.58, 216.36], [670.65, 216.48], [658.42, 175.14], [694.22, 175.14], [711.58, 175.47], [670.65, 175.47]]}]\n```" - } - ] - }, - { - "id": 2984, - "data_source": "Mono3DRefer", - "scene_name": "004174", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[700.34, 240.92], [768.55, 240.67], [830.6, 263.13], [739.99, 263.58], [700.34, 162.67], [768.55, 162.71], [830.6, 159.34], [739.99, 159.28]]}]\n```" - } - ] - }, - { - "id": 2985, - "data_source": "Mono3DRefer", - "scene_name": "004174", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[520.95, 259.46], [434.31, 259.47], [476.24, 238.68], [542.09, 238.67], [520.95, 179.54], [434.31, 179.54], [476.24, 177.93], [542.09, 177.93]]}, {\"category\": \"van\", \"corners_3d\": [[700.34, 240.92], [768.55, 240.67], [830.6, 263.13], [739.99, 263.58], [700.34, 162.67], [768.55, 162.71], [830.6, 159.34], [739.99, 159.28]]}, {\"category\": \"car\", \"corners_3d\": [[671.87, 220.82], [714.35, 220.73], [738.54, 229.94], [687.94, 230.06], [671.87, 176.19], [714.35, 176.19], [738.54, 176.83], [687.94, 176.83]]}, {\"category\": \"car\", \"corners_3d\": [[658.42, 210.96], [694.22, 210.87], [711.58, 216.36], [670.65, 216.48], [658.42, 175.14], [694.22, 175.14], [711.58, 175.47], [670.65, 175.47]]}]\n```" - } - ] - }, - { - "id": 2986, - "data_source": "Mono3DRefer", - "scene_name": "004176", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[399.71, 216.75], [388.88, 216.7], [400.26, 214.9], [410.65, 214.95], [399.71, 183.77], [388.88, 183.76], [400.26, 183.31], [410.65, 183.32]]}]\n```" - } - ] - }, - { - "id": 2987, - "data_source": "Mono3DRefer", - "scene_name": "004176", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[546.83, 214.68], [513.96, 214.7], [523.26, 210.27], [552.65, 210.26], [546.83, 183.57], [513.96, 183.58], [523.26, 182.44], [552.65, 182.44]]}, {\"category\": \"car\", \"corners_3d\": [[401.78, 300.75], [294.26, 300.39], [385.3, 265.11], [463.21, 265.31], [401.78, 186.04], [294.26, 186.0], [385.3, 182.36], [463.21, 182.38]]}]\n```" - } - ] - }, - { - "id": 2988, - "data_source": "Mono3DRefer", - "scene_name": "004176", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[546.83, 214.68], [513.96, 214.7], [523.26, 210.27], [552.65, 210.26], [546.83, 183.57], [513.96, 183.58], [523.26, 182.44], [552.65, 182.44]]}, {\"category\": \"car\", \"corners_3d\": [[401.78, 300.75], [294.26, 300.39], [385.3, 265.11], [463.21, 265.31], [401.78, 186.04], [294.26, 186.0], [385.3, 182.36], [463.21, 182.38]]}, {\"category\": \"cyclist\", \"corners_3d\": [[399.71, 216.75], [388.88, 216.7], [400.26, 214.9], [410.65, 214.95], [399.71, 183.77], [388.88, 183.76], [400.26, 183.31], [410.65, 183.32]]}]\n```" - } - ] - }, - { - "id": 2989, - "data_source": "Mono3DRefer", - "scene_name": "004184", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[434.4, 200.49], [442.9, 199.21], [530.08, 199.24], [525.81, 200.53], [434.4, 165.64], [442.9, 165.97], [530.08, 165.96], [525.81, 165.63]]}, {\"category\": \"car\", \"corners_3d\": [[339.91, 229.73], [360.87, 225.55], [476.15, 225.74], [464.29, 229.94], [339.91, 166.75], [360.87, 167.2], [476.15, 167.18], [464.29, 166.72]]}, {\"category\": \"car\", \"corners_3d\": [[377.13, 217.85], [391.92, 215.09], [498.51, 215.16], [490.67, 217.93], [377.13, 170.89], [391.92, 171.01], [498.51, 171.0], [490.67, 170.88]]}, {\"category\": \"car\", \"corners_3d\": [[785.02, 187.88], [819.08, 187.92], [841.09, 189.83], [802.64, 189.79], [785.02, 153.42], [819.08, 153.38], [841.09, 150.9], [802.64, 150.96]]}, {\"category\": \"car\", \"corners_3d\": [[413.14, 208.06], [422.63, 206.37], [514.09, 206.38], [509.21, 208.07], [413.14, 173.55], [422.63, 173.51], [514.09, 173.51], [509.21, 173.55]]}]\n```" - } - ] - }, - { - "id": 2990, - "data_source": "Mono3DRefer", - "scene_name": "004201", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[421.48, 306.1], [295.48, 306.12], [388.24, 266.69], [476.95, 266.68], [421.48, 200.65], [295.48, 200.66], [388.24, 192.43], [476.95, 192.43]]}]\n```" - } - ] - }, - { - "id": 2991, - "data_source": "Mono3DRefer", - "scene_name": "004203", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[512.57, 268.34], [428.24, 269.79], [443.96, 252.15], [512.94, 251.18], [512.57, 190.72], [428.24, 190.99], [443.96, 187.69], [512.94, 187.5]]}, {\"category\": \"car\", \"corners_3d\": [[521.82, 243.85], [472.02, 243.9], [491.76, 233.09], [533.98, 233.05], [521.82, 193.18], [472.02, 193.19], [491.76, 190.1], [533.98, 190.08]]}]\n```" - } - ] - }, - { - "id": 2992, - "data_source": "Mono3DRefer", - "scene_name": "004208", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-224.84, 821.44], [-713.79, 821.79], [83.04, 430.88], [277.34, 430.82], [-224.84, 241.02], [-713.79, 241.05], [83.04, 199.97], [277.34, 199.97]]}, {\"category\": \"car\", \"corners_3d\": [[547.51, 206.03], [522.99, 206.2], [521.27, 203.72], [543.99, 203.57], [547.51, 180.92], [522.99, 180.96], [521.27, 180.36], [543.99, 180.32]]}]\n```" - } - ] - }, - { - "id": 2993, - "data_source": "Mono3DRefer", - "scene_name": "004225", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[262.64, 482.45], [-23.73, 480.07], [299.47, 326.64], [443.52, 327.23], [262.64, 213.1], [-23.73, 212.79], [299.47, 192.84], [443.52, 192.92]]}, {\"category\": \"car\", \"corners_3d\": [[678.8, 276.93], [581.08, 276.52], [593.21, 251.16], [666.98, 251.39], [678.8, 188.11], [581.08, 188.05], [593.21, 184.33], [666.98, 184.37]]}]\n```" - } - ] - }, - { - "id": 2994, - "data_source": "Mono3DRefer", - "scene_name": "004230", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[366.07, 347.19], [219.43, 346.5], [359.74, 286.19], [455.68, 286.48], [366.07, 206.72], [219.43, 206.59], [359.74, 194.87], [455.68, 194.93]]}]\n```" - } - ] - }, - { - "id": 2995, - "data_source": "Mono3DRefer", - "scene_name": "004230", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[164.68, 265.01], [140.68, 264.89], [183.68, 256.93], [205.65, 257.03], [164.68, 192.0], [140.68, 191.98], [183.68, 190.32], [205.65, 190.34]]}]\n```" - } - ] - }, - { - "id": 2996, - "data_source": "Mono3DRefer", - "scene_name": "004230", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[366.07, 347.19], [219.43, 346.5], [359.74, 286.19], [455.68, 286.48], [366.07, 206.72], [219.43, 206.59], [359.74, 194.87], [455.68, 194.93]]}, {\"category\": \"cyclist\", \"corners_3d\": [[164.68, 265.01], [140.68, 264.89], [183.68, 256.93], [205.65, 257.03], [164.68, 192.0], [140.68, 191.98], [183.68, 190.32], [205.65, 190.34]]}]\n```" - } - ] - }, - { - "id": 2997, - "data_source": "Mono3DRefer", - "scene_name": "004231", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1370.08, 315.75], [1554.13, 345.48], [1234.03, 362.44], [1094.82, 327.17], [1370.08, 189.64], [1554.13, 193.13], [1234.03, 195.12], [1094.82, 190.98]]}, {\"category\": \"car\", \"corners_3d\": [[852.49, 289.59], [927.44, 312.08], [546.07, 328.46], [531.41, 300.9], [852.49, 186.98], [927.44, 189.7], [546.07, 191.69], [531.41, 188.35]]}]\n```" - } - ] - }, - { - "id": 2998, - "data_source": "Mono3DRefer", - "scene_name": "004238", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[422.93, 355.71], [259.17, 354.91], [385.62, 291.71], [492.73, 292.05], [422.93, 187.28], [259.17, 187.22], [385.62, 182.23], [492.73, 182.26]]}, {\"category\": \"car\", \"corners_3d\": [[761.02, 320.11], [883.76, 319.12], [1072.7, 410.47], [874.55, 413.1], [761.02, 187.05], [883.76, 186.95], [1072.7, 195.76], [874.55, 196.02]]}, {\"category\": \"car\", \"corners_3d\": [[682.35, 268.08], [763.54, 267.66], [823.82, 299.02], [715.98, 299.77], [682.35, 189.22], [763.54, 189.15], [823.82, 194.54], [715.98, 194.67]]}, {\"category\": \"car\", \"corners_3d\": [[447.47, 252.34], [515.91, 252.05], [497.68, 274.14], [409.97, 274.62], [447.47, 182.38], [515.91, 182.34], [497.68, 184.99], [409.97, 185.05]]}, {\"category\": \"car\", \"corners_3d\": [[520.35, 246.78], [453.76, 246.78], [478.32, 235.08], [534.38, 235.08], [520.35, 187.95], [453.76, 187.95], [478.32, 185.57], [534.38, 185.56]]}, {\"category\": \"car\", \"corners_3d\": [[661.02, 236.81], [714.29, 236.38], [745.75, 248.19], [682.75, 248.79], [661.02, 182.63], [714.29, 182.56], [745.75, 184.37], [682.75, 184.46]]}, {\"category\": \"car\", \"corners_3d\": [[638.99, 223.84], [683.36, 223.71], [700.8, 232.4], [648.88, 232.57], [638.99, 182.6], [683.36, 182.57], [700.8, 184.23], [648.88, 184.27]]}, {\"category\": \"car\", \"corners_3d\": [[553.83, 216.85], [519.64, 216.81], [530.04, 212.41], [560.81, 212.44], [553.83, 182.36], [519.64, 182.35], [530.04, 181.39], [560.81, 181.4]]}, {\"category\": \"car\", \"corners_3d\": [[571.76, 209.52], [540.25, 209.5], [547.84, 205.84], [576.2, 205.85], [571.76, 173.23], [540.25, 173.23], [547.84, 173.19], [576.2, 173.19]]}, {\"category\": \"car\", \"corners_3d\": [[584.36, 206.74], [556.69, 206.68], [564.53, 203.68], [589.76, 203.73], [584.36, 176.63], [556.69, 176.63], [564.53, 176.29], [589.76, 176.3]]}]\n```" - } - ] - }, - { - "id": 2999, - "data_source": "Mono3DRefer", - "scene_name": "004247", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[416.17, 307.19], [402.01, 313.59], [319.97, 311.49], [337.8, 305.29], [416.17, 167.5], [402.01, 166.84], [319.97, 167.06], [337.8, 167.7]]}]\n```" - } - ] - }, - { - "id": 3000, - "data_source": "Mono3DRefer", - "scene_name": "004266", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[557.53, 223.36], [520.07, 223.44], [527.57, 217.8], [560.85, 217.74], [557.53, 184.03], [520.07, 184.05], [527.57, 182.8], [560.85, 182.78]]}, {\"category\": \"car\", \"corners_3d\": [[572.71, 207.18], [554.5, 207.16], [558.59, 205.12], [575.72, 205.14], [572.71, 186.91], [554.5, 186.91], [558.59, 186.07], [575.72, 186.08]]}]\n```" - } - ] - }, - { - "id": 3001, - "data_source": "Mono3DRefer", - "scene_name": "004294", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[845.79, 289.28], [739.71, 289.29], [702.13, 255.81], [777.72, 255.81], [845.79, 196.12], [739.71, 196.12], [702.13, 189.43], [777.72, 189.43]]}, {\"category\": \"car\", \"corners_3d\": [[531.27, 227.5], [479.26, 227.54], [495.65, 220.22], [540.69, 220.19], [531.27, 177.31], [479.26, 177.31], [495.65, 176.72], [540.69, 176.71]]}, {\"category\": \"car\", \"corners_3d\": [[724.9, 222.06], [679.59, 222.0], [671.64, 215.04], [710.52, 215.08], [724.9, 182.8], [679.59, 182.79], [671.64, 181.38], [710.52, 181.39]]}, {\"category\": \"car\", \"corners_3d\": [[652.01, 198.53], [674.81, 198.54], [679.01, 200.69], [654.3, 200.67], [652.01, 176.5], [674.81, 176.5], [679.01, 176.8], [654.3, 176.8]]}, {\"category\": \"car\", \"corners_3d\": [[337.68, 383.86], [128.55, 381.43], [351.8, 290.0], [470.12, 290.77], [337.68, 197.26], [128.55, 196.98], [351.8, 186.4], [470.12, 186.49]]}, {\"category\": \"car\", \"corners_3d\": [[763.41, 302.75], [880.11, 302.13], [1051.79, 378.24], [867.18, 379.8], [763.41, 193.62], [880.11, 193.52], [1051.79, 205.68], [867.18, 205.93]]}, {\"category\": \"car\", \"corners_3d\": [[465.88, 269.97], [367.04, 269.71], [426.51, 247.24], [502.5, 247.4], [465.88, 177.57], [367.04, 177.56], [426.51, 176.47], [502.5, 176.47]]}]\n```" - } - ] - }, - { - "id": 3002, - "data_source": "Mono3DRefer", - "scene_name": "004294", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[644.42, 193.17], [667.19, 193.19], [670.67, 195.18], [645.66, 195.16], [644.42, 165.85], [667.19, 165.84], [670.67, 165.15], [645.66, 165.16]]}]\n```" - } - ] - }, - { - "id": 3003, - "data_source": "Mono3DRefer", - "scene_name": "004294", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[845.79, 289.28], [739.71, 289.29], [702.13, 255.81], [777.72, 255.81], [845.79, 196.12], [739.71, 196.12], [702.13, 189.43], [777.72, 189.43]]}, {\"category\": \"car\", \"corners_3d\": [[531.27, 227.5], [479.26, 227.54], [495.65, 220.22], [540.69, 220.19], [531.27, 177.31], [479.26, 177.31], [495.65, 176.72], [540.69, 176.71]]}, {\"category\": \"car\", \"corners_3d\": [[724.9, 222.06], [679.59, 222.0], [671.64, 215.04], [710.52, 215.08], [724.9, 182.8], [679.59, 182.79], [671.64, 181.38], [710.52, 181.39]]}, {\"category\": \"car\", \"corners_3d\": [[652.01, 198.53], [674.81, 198.54], [679.01, 200.69], [654.3, 200.67], [652.01, 176.5], [674.81, 176.5], [679.01, 176.8], [654.3, 176.8]]}, {\"category\": \"van\", \"corners_3d\": [[644.42, 193.17], [667.19, 193.19], [670.67, 195.18], [645.66, 195.16], [644.42, 165.85], [667.19, 165.84], [670.67, 165.15], [645.66, 165.16]]}, {\"category\": \"car\", \"corners_3d\": [[337.68, 383.86], [128.55, 381.43], [351.8, 290.0], [470.12, 290.77], [337.68, 197.26], [128.55, 196.98], [351.8, 186.4], [470.12, 186.49]]}, {\"category\": \"car\", \"corners_3d\": [[763.41, 302.75], [880.11, 302.13], [1051.79, 378.24], [867.18, 379.8], [763.41, 193.62], [880.11, 193.52], [1051.79, 205.68], [867.18, 205.93]]}, {\"category\": \"car\", \"corners_3d\": [[465.88, 269.97], [367.04, 269.71], [426.51, 247.24], [502.5, 247.4], [465.88, 177.57], [367.04, 177.56], [426.51, 176.47], [502.5, 176.47]]}]\n```" - } - ] - }, - { - "id": 3004, - "data_source": "Mono3DRefer", - "scene_name": "004311", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[776.14, 294.99], [914.15, 295.71], [1177.81, 410.43], [909.43, 407.77], [776.14, 171.16], [914.15, 171.15], [1177.81, 169.56], [909.43, 169.6]]}, {\"category\": \"car\", \"corners_3d\": [[-126.79, 287.53], [-19.52, 271.92], [192.0, 274.14], [116.12, 290.52], [-126.79, 181.49], [-19.52, 180.31], [192.0, 180.48], [116.12, 181.71]]}, {\"category\": \"car\", \"corners_3d\": [[509.9, 220.23], [566.78, 220.08], [563.73, 230.05], [494.78, 230.26], [509.9, 168.3], [566.78, 168.31], [563.73, 167.35], [494.78, 167.33]]}, {\"category\": \"car\", \"corners_3d\": [[686.18, 224.22], [742.01, 223.98], [783.48, 236.08], [714.59, 236.44], [686.18, 169.39], [742.01, 169.41], [783.48, 168.59], [714.59, 168.57]]}, {\"category\": \"car\", \"corners_3d\": [[676.6, 208.29], [723.77, 208.38], [740.25, 215.29], [683.88, 215.16], [676.6, 169.2], [723.77, 169.19], [740.25, 168.47], [683.88, 168.49]]}, {\"category\": \"car\", \"corners_3d\": [[562.67, 199.18], [594.15, 199.26], [587.55, 201.88], [552.99, 201.78], [562.67, 169.61], [594.15, 169.6], [587.55, 169.27], [552.99, 169.28]]}, {\"category\": \"car\", \"corners_3d\": [[662.53, 196.45], [695.68, 196.49], [704.26, 199.66], [666.65, 199.62], [662.53, 161.89], [695.68, 161.88], [704.26, 160.4], [666.65, 160.42]]}]\n```" - } - ] - }, - { - "id": 3005, - "data_source": "Mono3DRefer", - "scene_name": "004316", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[585.53, 221.21], [539.05, 221.34], [545.34, 213.92], [584.7, 213.82], [585.53, 180.13], [539.05, 180.15], [545.34, 179.03], [584.7, 179.02]]}]\n```" - } - ] - }, - { - "id": 3006, - "data_source": "Mono3DRefer", - "scene_name": "004325", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-273.85, 372.05], [-117.69, 334.65], [187.3, 330.07], [104.61, 365.16], [-273.85, 225.46], [-117.69, 215.59], [187.3, 214.38], [104.61, 223.64]]}, {\"category\": \"car\", \"corners_3d\": [[279.86, 245.8], [302.32, 240.33], [430.06, 239.84], [418.05, 245.23], [279.86, 192.22], [302.32, 190.77], [430.06, 190.64], [418.05, 192.07]]}]\n```" - } - ] - }, - { - "id": 3007, - "data_source": "Mono3DRefer", - "scene_name": "004325", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[610.24, 224.15], [627.98, 224.12], [631.54, 228.74], [612.2, 228.79], [610.24, 175.95], [627.98, 175.94], [631.54, 176.22], [612.2, 176.22]]}]\n```" - } - ] - }, - { - "id": 3008, - "data_source": "Mono3DRefer", - "scene_name": "004325", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[610.24, 224.15], [627.98, 224.12], [631.54, 228.74], [612.2, 228.79], [610.24, 175.95], [627.98, 175.94], [631.54, 176.22], [612.2, 176.22]]}, {\"category\": \"car\", \"corners_3d\": [[-273.85, 372.05], [-117.69, 334.65], [187.3, 330.07], [104.61, 365.16], [-273.85, 225.46], [-117.69, 215.59], [187.3, 214.38], [104.61, 223.64]]}, {\"category\": \"car\", \"corners_3d\": [[279.86, 245.8], [302.32, 240.33], [430.06, 239.84], [418.05, 245.23], [279.86, 192.22], [302.32, 190.77], [430.06, 190.64], [418.05, 192.07]]}]\n```" - } - ] - }, - { - "id": 3009, - "data_source": "Mono3DRefer", - "scene_name": "004342", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[550.71, 208.77], [584.65, 208.58], [596.83, 211.22], [559.13, 211.46], [550.71, 177.59], [584.65, 177.66], [596.83, 176.8], [559.13, 176.72]]}, {\"category\": \"car\", \"corners_3d\": [[504.81, 201.98], [526.56, 201.81], [538.33, 202.94], [515.22, 203.13], [504.81, 182.21], [526.56, 182.24], [538.33, 182.04], [515.22, 182.01]]}, {\"category\": \"car\", \"corners_3d\": [[429.62, 195.79], [445.76, 195.64], [462.5, 196.11], [445.88, 196.27], [429.62, 181.16], [445.76, 181.21], [462.5, 181.03], [445.88, 180.97]]}]\n```" - } - ] - }, - { - "id": 3010, - "data_source": "Mono3DRefer", - "scene_name": "004345", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[592.88, 214.84], [559.38, 214.57], [573.55, 210.66], [603.97, 210.89], [592.88, 184.7], [559.38, 184.62], [573.55, 183.52], [603.97, 183.59]]}]\n```" - } - ] - }, - { - "id": 3011, - "data_source": "Mono3DRefer", - "scene_name": "004349", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[613.3, 287.73], [640.36, 305.04], [352.8, 315.71], [367.02, 295.52], [613.3, 189.49], [640.36, 190.79], [352.8, 191.59], [367.02, 190.07]]}]\n```" - } - ] - }, - { - "id": 3012, - "data_source": "Mono3DRefer", - "scene_name": "004349", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[1103.56, 206.16], [1181.85, 208.79], [867.74, 213.95], [811.7, 210.3], [1103.56, 109.75], [1181.85, 102.09], [867.74, 87.07], [811.7, 97.71]]}]\n```" - } - ] - }, - { - "id": 3013, - "data_source": "Mono3DRefer", - "scene_name": "004349", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[243.05, 450.9], [274.53, 403.87], [428.5, 393.31], [429.52, 435.55], [243.05, 147.98], [274.53, 153.83], [428.5, 155.15], [429.52, 149.89]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[-140.18, 505.24], [-47.51, 449.74], [132.54, 434.78], [80.81, 483.7], [-140.18, 188.94], [-47.51, 187.67], [132.54, 187.32], [80.81, 188.45]]}]\n```" - } - ] - }, - { - "id": 3014, - "data_source": "Mono3DRefer", - "scene_name": "004349", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[1103.56, 206.16], [1181.85, 208.79], [867.74, 213.95], [811.7, 210.3], [1103.56, 109.75], [1181.85, 102.09], [867.74, 87.07], [811.7, 97.71]]}, {\"category\": \"car\", \"corners_3d\": [[613.3, 287.73], [640.36, 305.04], [352.8, 315.71], [367.02, 295.52], [613.3, 189.49], [640.36, 190.79], [352.8, 191.59], [367.02, 190.07]]}]\n```" - } - ] - }, - { - "id": 3015, - "data_source": "Mono3DRefer", - "scene_name": "004349", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[243.05, 450.9], [274.53, 403.87], [428.5, 393.31], [429.52, 435.55], [243.05, 147.98], [274.53, 153.83], [428.5, 155.15], [429.52, 149.89]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[-140.18, 505.24], [-47.51, 449.74], [132.54, 434.78], [80.81, 483.7], [-140.18, 188.94], [-47.51, 187.67], [132.54, 187.32], [80.81, 188.45]]}, {\"category\": \"car\", \"corners_3d\": [[613.3, 287.73], [640.36, 305.04], [352.8, 315.71], [367.02, 295.52], [613.3, 189.49], [640.36, 190.79], [352.8, 191.59], [367.02, 190.07]]}]\n```" - } - ] - }, - { - "id": 3016, - "data_source": "Mono3DRefer", - "scene_name": "004349", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[1103.56, 206.16], [1181.85, 208.79], [867.74, 213.95], [811.7, 210.3], [1103.56, 109.75], [1181.85, 102.09], [867.74, 87.07], [811.7, 97.71]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[243.05, 450.9], [274.53, 403.87], [428.5, 393.31], [429.52, 435.55], [243.05, 147.98], [274.53, 153.83], [428.5, 155.15], [429.52, 149.89]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[-140.18, 505.24], [-47.51, 449.74], [132.54, 434.78], [80.81, 483.7], [-140.18, 188.94], [-47.51, 187.67], [132.54, 187.32], [80.81, 188.45]]}]\n```" - } - ] - }, - { - "id": 3017, - "data_source": "Mono3DRefer", - "scene_name": "004349", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[1103.56, 206.16], [1181.85, 208.79], [867.74, 213.95], [811.7, 210.3], [1103.56, 109.75], [1181.85, 102.09], [867.74, 87.07], [811.7, 97.71]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[243.05, 450.9], [274.53, 403.87], [428.5, 393.31], [429.52, 435.55], [243.05, 147.98], [274.53, 153.83], [428.5, 155.15], [429.52, 149.89]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[-140.18, 505.24], [-47.51, 449.74], [132.54, 434.78], [80.81, 483.7], [-140.18, 188.94], [-47.51, 187.67], [132.54, 187.32], [80.81, 188.45]]}, {\"category\": \"car\", \"corners_3d\": [[613.3, 287.73], [640.36, 305.04], [352.8, 315.71], [367.02, 295.52], [613.3, 189.49], [640.36, 190.79], [352.8, 191.59], [367.02, 190.07]]}]\n```" - } - ] - }, - { - "id": 3018, - "data_source": "Mono3DRefer", - "scene_name": "004352", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[652.97, 197.63], [674.26, 197.65], [677.55, 199.57], [654.61, 199.54], [652.97, 177.07], [674.26, 177.07], [677.55, 177.4], [654.61, 177.4]]}, {\"category\": \"car\", \"corners_3d\": [[1603.63, 624.62], [1183.08, 622.23], [850.74, 358.77], [1023.46, 359.18], [1603.63, 196.54], [1183.08, 196.42], [850.74, 182.6], [1023.46, 182.62]]}, {\"category\": \"car\", \"corners_3d\": [[444.91, 304.85], [315.1, 303.91], [420.01, 261.31], [507.93, 261.74], [444.91, 190.22], [315.1, 190.09], [420.01, 184.49], [507.93, 184.55]]}, {\"category\": \"car\", \"corners_3d\": [[724.62, 268.25], [810.09, 267.92], [892.44, 303.25], [775.46, 303.88], [724.62, 188.56], [810.09, 188.5], [892.44, 194.32], [775.46, 194.42]]}, {\"category\": \"car\", \"corners_3d\": [[501.57, 251.61], [424.79, 251.36], [463.93, 236.44], [526.17, 236.6], [501.57, 179.75], [424.79, 179.73], [463.93, 178.42], [526.17, 178.43]]}, {\"category\": \"car\", \"corners_3d\": [[790.2, 259.74], [709.58, 259.65], [687.71, 239.32], [749.41, 239.37], [790.2, 189.04], [709.58, 189.02], [687.71, 185.24], [749.41, 185.25]]}, {\"category\": \"car\", \"corners_3d\": [[548.2, 222.22], [503.76, 222.25], [515.19, 216.49], [554.45, 216.46], [548.2, 179.31], [503.76, 179.31], [515.19, 178.56], [554.45, 178.56]]}, {\"category\": \"car\", \"corners_3d\": [[715.4, 216.63], [675.78, 216.58], [669.18, 211.07], [703.79, 211.11], [715.4, 182.3], [675.78, 182.29], [669.18, 181.1], [703.79, 181.11]]}]\n```" - } - ] - }, - { - "id": 3019, - "data_source": "Mono3DRefer", - "scene_name": "004352", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[645.27, 194.59], [666.84, 194.62], [669.42, 196.63], [645.85, 196.6], [645.27, 168.72], [666.84, 168.71], [669.42, 168.33], [645.85, 168.34]]}]\n```" - } - ] - }, - { - "id": 3020, - "data_source": "Mono3DRefer", - "scene_name": "004352", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[652.97, 197.63], [674.26, 197.65], [677.55, 199.57], [654.61, 199.54], [652.97, 177.07], [674.26, 177.07], [677.55, 177.4], [654.61, 177.4]]}, {\"category\": \"van\", \"corners_3d\": [[645.27, 194.59], [666.84, 194.62], [669.42, 196.63], [645.85, 196.6], [645.27, 168.72], [666.84, 168.71], [669.42, 168.33], [645.85, 168.34]]}, {\"category\": \"car\", \"corners_3d\": [[1603.63, 624.62], [1183.08, 622.23], [850.74, 358.77], [1023.46, 359.18], [1603.63, 196.54], [1183.08, 196.42], [850.74, 182.6], [1023.46, 182.62]]}, {\"category\": \"car\", \"corners_3d\": [[444.91, 304.85], [315.1, 303.91], [420.01, 261.31], [507.93, 261.74], [444.91, 190.22], [315.1, 190.09], [420.01, 184.49], [507.93, 184.55]]}, {\"category\": \"car\", \"corners_3d\": [[724.62, 268.25], [810.09, 267.92], [892.44, 303.25], [775.46, 303.88], [724.62, 188.56], [810.09, 188.5], [892.44, 194.32], [775.46, 194.42]]}, {\"category\": \"car\", \"corners_3d\": [[501.57, 251.61], [424.79, 251.36], [463.93, 236.44], [526.17, 236.6], [501.57, 179.75], [424.79, 179.73], [463.93, 178.42], [526.17, 178.43]]}, {\"category\": \"car\", \"corners_3d\": [[790.2, 259.74], [709.58, 259.65], [687.71, 239.32], [749.41, 239.37], [790.2, 189.04], [709.58, 189.02], [687.71, 185.24], [749.41, 185.25]]}, {\"category\": \"car\", \"corners_3d\": [[548.2, 222.22], [503.76, 222.25], [515.19, 216.49], [554.45, 216.46], [548.2, 179.31], [503.76, 179.31], [515.19, 178.56], [554.45, 178.56]]}, {\"category\": \"car\", \"corners_3d\": [[715.4, 216.63], [675.78, 216.58], [669.18, 211.07], [703.79, 211.11], [715.4, 182.3], [675.78, 182.29], [669.18, 181.1], [703.79, 181.11]]}]\n```" - } - ] - }, - { - "id": 3021, - "data_source": "Mono3DRefer", - "scene_name": "004354", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[576.51, 195.89], [555.49, 195.89], [558.82, 194.45], [578.53, 194.45], [576.51, 178.67], [555.49, 178.67], [558.82, 178.31], [578.53, 178.31]]}]\n```" - } - ] - }, - { - "id": 3022, - "data_source": "Mono3DRefer", - "scene_name": "004355", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[620.13, 226.98], [676.19, 227.11], [685.68, 238.71], [617.64, 238.52], [620.13, 177.86], [676.19, 177.87], [685.68, 178.95], [617.64, 178.93]]}, {\"category\": \"car\", \"corners_3d\": [[551.38, 213.61], [597.76, 213.79], [585.5, 221.76], [530.21, 221.5], [551.38, 169.26], [597.76, 169.25], [585.5, 168.54], [530.21, 168.57]]}, {\"category\": \"car\", \"corners_3d\": [[638.77, 205.87], [678.5, 205.98], [682.97, 211.9], [636.15, 211.74], [638.77, 168.72], [678.5, 168.7], [682.97, 167.96], [636.15, 167.98]]}, {\"category\": \"car\", \"corners_3d\": [[1244.04, 194.78], [1314.64, 196.69], [1193.76, 197.97], [1129.83, 195.86], [1244.04, 139.75], [1314.64, 136.87], [1193.76, 134.93], [1129.83, 138.12]]}, {\"category\": \"car\", \"corners_3d\": [[1142.31, 191.07], [1194.71, 192.43], [1091.97, 193.24], [1044.98, 191.77], [1142.31, 141.5], [1194.71, 139.16], [1091.97, 137.77], [1044.98, 140.3]]}]\n```" - } - ] - }, - { - "id": 3023, - "data_source": "Mono3DRefer", - "scene_name": "004356", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[482.37, 304.96], [348.82, 307.63], [395.77, 272.56], [494.39, 271.1], [482.37, 183.16], [348.82, 183.37], [395.77, 180.63], [494.39, 180.52]]}, {\"category\": \"car\", \"corners_3d\": [[513.95, 252.64], [447.87, 252.64], [478.89, 237.28], [532.25, 237.28], [513.95, 185.28], [447.87, 185.28], [478.89, 182.89], [532.25, 182.89]]}]\n```" - } - ] - }, - { - "id": 3024, - "data_source": "Mono3DRefer", - "scene_name": "004357", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[552.37, 205.63], [523.04, 205.68], [528.08, 202.78], [554.82, 202.74], [552.37, 176.75], [523.04, 176.76], [528.08, 176.42], [554.82, 176.41]]}]\n```" - } - ] - }, - { - "id": 3025, - "data_source": "Mono3DRefer", - "scene_name": "004371", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[594.32, 262.33], [634.63, 261.43], [656.18, 273.68], [610.47, 274.85], [594.32, 174.85], [634.63, 174.83], [656.18, 175.1], [610.47, 175.13]]}]\n```" - } - ] - }, - { - "id": 3026, - "data_source": "Mono3DRefer", - "scene_name": "004371", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[459.88, 209.5], [493.07, 209.11], [500.16, 213.44], [463.06, 213.93], [459.88, 173.39], [493.07, 173.38], [500.16, 173.45], [463.06, 173.45]]}, {\"category\": \"van\", \"corners_3d\": [[380.7, 236.63], [321.74, 237.62], [344.59, 227.12], [393.82, 226.42], [380.7, 179.03], [321.74, 179.12], [344.59, 178.11], [393.82, 178.04]]}]\n```" - } - ] - }, - { - "id": 3027, - "data_source": "Mono3DRefer", - "scene_name": "004371", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[768.7, 254.14], [851.02, 258.01], [821.35, 286.03], [715.33, 279.3], [768.7, 183.99], [851.02, 184.52], [821.35, 188.36], [715.33, 187.43]]}, {\"category\": \"car\", \"corners_3d\": [[616.98, 239.24], [673.64, 237.97], [732.09, 253.82], [662.73, 255.8], [616.98, 184.16], [673.64, 183.94], [732.09, 186.64], [662.73, 186.98]]}, {\"category\": \"car\", \"corners_3d\": [[352.3, 249.51], [276.19, 250.97], [313.33, 236.02], [374.52, 235.06], [352.3, 185.97], [276.19, 186.22], [313.33, 183.66], [374.52, 183.5]]}, {\"category\": \"car\", \"corners_3d\": [[568.44, 223.73], [614.91, 222.98], [641.58, 230.67], [588.32, 231.68], [568.44, 178.75], [614.91, 178.66], [641.58, 179.55], [588.32, 179.67]]}, {\"category\": \"car\", \"corners_3d\": [[541.91, 216.73], [581.58, 216.05], [606.06, 221.97], [561.3, 222.84], [541.91, 180.87], [581.58, 180.74], [606.06, 181.82], [561.3, 181.98]]}, {\"category\": \"car\", \"corners_3d\": [[523.79, 210.7], [553.94, 210.29], [567.96, 214.16], [534.83, 214.65], [523.79, 182.97], [553.94, 182.86], [567.96, 183.89], [534.83, 184.02]]}]\n```" - } - ] - }, - { - "id": 3028, - "data_source": "Mono3DRefer", - "scene_name": "004371", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[459.88, 209.5], [493.07, 209.11], [500.16, 213.44], [463.06, 213.93], [459.88, 173.39], [493.07, 173.38], [500.16, 173.45], [463.06, 173.45]]}, {\"category\": \"cyclist\", \"corners_3d\": [[594.32, 262.33], [634.63, 261.43], [656.18, 273.68], [610.47, 274.85], [594.32, 174.85], [634.63, 174.83], [656.18, 175.1], [610.47, 175.13]]}, {\"category\": \"van\", \"corners_3d\": [[380.7, 236.63], [321.74, 237.62], [344.59, 227.12], [393.82, 226.42], [380.7, 179.03], [321.74, 179.12], [344.59, 178.11], [393.82, 178.04]]}]\n```" - } - ] - }, - { - "id": 3029, - "data_source": "Mono3DRefer", - "scene_name": "004371", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[594.32, 262.33], [634.63, 261.43], [656.18, 273.68], [610.47, 274.85], [594.32, 174.85], [634.63, 174.83], [656.18, 175.1], [610.47, 175.13]]}, {\"category\": \"car\", \"corners_3d\": [[768.7, 254.14], [851.02, 258.01], [821.35, 286.03], [715.33, 279.3], [768.7, 183.99], [851.02, 184.52], [821.35, 188.36], [715.33, 187.43]]}, {\"category\": \"car\", \"corners_3d\": [[616.98, 239.24], [673.64, 237.97], [732.09, 253.82], [662.73, 255.8], [616.98, 184.16], [673.64, 183.94], [732.09, 186.64], [662.73, 186.98]]}, {\"category\": \"car\", \"corners_3d\": [[352.3, 249.51], [276.19, 250.97], [313.33, 236.02], [374.52, 235.06], [352.3, 185.97], [276.19, 186.22], [313.33, 183.66], [374.52, 183.5]]}, {\"category\": \"car\", \"corners_3d\": [[568.44, 223.73], [614.91, 222.98], [641.58, 230.67], [588.32, 231.68], [568.44, 178.75], [614.91, 178.66], [641.58, 179.55], [588.32, 179.67]]}, {\"category\": \"car\", \"corners_3d\": [[541.91, 216.73], [581.58, 216.05], [606.06, 221.97], [561.3, 222.84], [541.91, 180.87], [581.58, 180.74], [606.06, 181.82], [561.3, 181.98]]}, {\"category\": \"car\", \"corners_3d\": [[523.79, 210.7], [553.94, 210.29], [567.96, 214.16], [534.83, 214.65], [523.79, 182.97], [553.94, 182.86], [567.96, 183.89], [534.83, 184.02]]}]\n```" - } - ] - }, - { - "id": 3030, - "data_source": "Mono3DRefer", - "scene_name": "004371", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[459.88, 209.5], [493.07, 209.11], [500.16, 213.44], [463.06, 213.93], [459.88, 173.39], [493.07, 173.38], [500.16, 173.45], [463.06, 173.45]]}, {\"category\": \"car\", \"corners_3d\": [[768.7, 254.14], [851.02, 258.01], [821.35, 286.03], [715.33, 279.3], [768.7, 183.99], [851.02, 184.52], [821.35, 188.36], [715.33, 187.43]]}, {\"category\": \"car\", \"corners_3d\": [[616.98, 239.24], [673.64, 237.97], [732.09, 253.82], [662.73, 255.8], [616.98, 184.16], [673.64, 183.94], [732.09, 186.64], [662.73, 186.98]]}, {\"category\": \"car\", \"corners_3d\": [[352.3, 249.51], [276.19, 250.97], [313.33, 236.02], [374.52, 235.06], [352.3, 185.97], [276.19, 186.22], [313.33, 183.66], [374.52, 183.5]]}, {\"category\": \"van\", \"corners_3d\": [[380.7, 236.63], [321.74, 237.62], [344.59, 227.12], [393.82, 226.42], [380.7, 179.03], [321.74, 179.12], [344.59, 178.11], [393.82, 178.04]]}, {\"category\": \"car\", \"corners_3d\": [[568.44, 223.73], [614.91, 222.98], [641.58, 230.67], [588.32, 231.68], [568.44, 178.75], [614.91, 178.66], [641.58, 179.55], [588.32, 179.67]]}, {\"category\": \"car\", \"corners_3d\": [[541.91, 216.73], [581.58, 216.05], [606.06, 221.97], [561.3, 222.84], [541.91, 180.87], [581.58, 180.74], [606.06, 181.82], [561.3, 181.98]]}, {\"category\": \"car\", \"corners_3d\": [[523.79, 210.7], [553.94, 210.29], [567.96, 214.16], [534.83, 214.65], [523.79, 182.97], [553.94, 182.86], [567.96, 183.89], [534.83, 184.02]]}]\n```" - } - ] - }, - { - "id": 3031, - "data_source": "Mono3DRefer", - "scene_name": "004371", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[459.88, 209.5], [493.07, 209.11], [500.16, 213.44], [463.06, 213.93], [459.88, 173.39], [493.07, 173.38], [500.16, 173.45], [463.06, 173.45]]}, {\"category\": \"cyclist\", \"corners_3d\": [[594.32, 262.33], [634.63, 261.43], [656.18, 273.68], [610.47, 274.85], [594.32, 174.85], [634.63, 174.83], [656.18, 175.1], [610.47, 175.13]]}, {\"category\": \"car\", \"corners_3d\": [[768.7, 254.14], [851.02, 258.01], [821.35, 286.03], [715.33, 279.3], [768.7, 183.99], [851.02, 184.52], [821.35, 188.36], [715.33, 187.43]]}, {\"category\": \"car\", \"corners_3d\": [[616.98, 239.24], [673.64, 237.97], [732.09, 253.82], [662.73, 255.8], [616.98, 184.16], [673.64, 183.94], [732.09, 186.64], [662.73, 186.98]]}, {\"category\": \"car\", \"corners_3d\": [[352.3, 249.51], [276.19, 250.97], [313.33, 236.02], [374.52, 235.06], [352.3, 185.97], [276.19, 186.22], [313.33, 183.66], [374.52, 183.5]]}, {\"category\": \"van\", \"corners_3d\": [[380.7, 236.63], [321.74, 237.62], [344.59, 227.12], [393.82, 226.42], [380.7, 179.03], [321.74, 179.12], [344.59, 178.11], [393.82, 178.04]]}, {\"category\": \"car\", \"corners_3d\": [[568.44, 223.73], [614.91, 222.98], [641.58, 230.67], [588.32, 231.68], [568.44, 178.75], [614.91, 178.66], [641.58, 179.55], [588.32, 179.67]]}, {\"category\": \"car\", \"corners_3d\": [[541.91, 216.73], [581.58, 216.05], [606.06, 221.97], [561.3, 222.84], [541.91, 180.87], [581.58, 180.74], [606.06, 181.82], [561.3, 181.98]]}, {\"category\": \"car\", \"corners_3d\": [[523.79, 210.7], [553.94, 210.29], [567.96, 214.16], [534.83, 214.65], [523.79, 182.97], [553.94, 182.86], [567.96, 183.89], [534.83, 184.02]]}]\n```" - } - ] - }, - { - "id": 3032, - "data_source": "Mono3DRefer", - "scene_name": "004373", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[467.02, 270.57], [370.05, 270.05], [431.91, 247.6], [506.64, 247.91], [467.02, 183.82], [370.05, 183.76], [431.91, 181.24], [506.64, 181.28]]}, {\"category\": \"car\", \"corners_3d\": [[529.63, 195.99], [553.77, 195.95], [552.21, 197.73], [526.22, 197.77], [529.63, 171.61], [553.77, 171.61], [552.21, 171.51], [526.22, 171.51]]}]\n```" - } - ] - }, - { - "id": 3033, - "data_source": "Mono3DRefer", - "scene_name": "004382", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[156.34, 261.99], [137.44, 262.04], [179.32, 253.87], [196.47, 253.83], [156.34, 197.83], [137.44, 197.85], [179.32, 195.56], [196.47, 195.55]]}]\n```" - } - ] - }, - { - "id": 3034, - "data_source": "Mono3DRefer", - "scene_name": "004382", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[571.8, 225.69], [616.1, 225.66], [617.86, 232.74], [567.62, 232.78], [571.8, 182.72], [616.1, 182.71], [617.86, 184.04], [567.62, 184.04]]}, {\"category\": \"car\", \"corners_3d\": [[421.83, 240.51], [376.86, 240.56], [406.58, 231.61], [445.6, 231.58], [421.83, 198.96], [376.86, 198.98], [406.58, 195.53], [445.6, 195.51]]}, {\"category\": \"car\", \"corners_3d\": [[539.66, 220.94], [504.79, 220.94], [515.78, 215.87], [546.97, 215.87], [539.66, 190.28], [504.79, 190.28], [515.78, 188.44], [546.97, 188.44]]}]\n```" - } - ] - }, - { - "id": 3035, - "data_source": "Mono3DRefer", - "scene_name": "004382", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[571.8, 225.69], [616.1, 225.66], [617.86, 232.74], [567.62, 232.78], [571.8, 182.72], [616.1, 182.71], [617.86, 184.04], [567.62, 184.04]]}, {\"category\": \"car\", \"corners_3d\": [[421.83, 240.51], [376.86, 240.56], [406.58, 231.61], [445.6, 231.58], [421.83, 198.96], [376.86, 198.98], [406.58, 195.53], [445.6, 195.51]]}, {\"category\": \"car\", \"corners_3d\": [[539.66, 220.94], [504.79, 220.94], [515.78, 215.87], [546.97, 215.87], [539.66, 190.28], [504.79, 190.28], [515.78, 188.44], [546.97, 188.44]]}, {\"category\": \"cyclist\", \"corners_3d\": [[156.34, 261.99], [137.44, 262.04], [179.32, 253.87], [196.47, 253.83], [156.34, 197.83], [137.44, 197.85], [179.32, 195.56], [196.47, 195.55]]}]\n```" - } - ] - }, - { - "id": 3036, - "data_source": "Mono3DRefer", - "scene_name": "004394", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[552.73, 214.65], [514.45, 214.63], [525.45, 210.11], [559.6, 210.13], [552.73, 183.24], [514.45, 183.24], [525.45, 182.11], [559.6, 182.12]]}]\n```" - } - ] - }, - { - "id": 3037, - "data_source": "Mono3DRefer", - "scene_name": "004395", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[512.38, 274.34], [434.35, 274.57], [468.11, 253.13], [529.66, 252.99], [512.38, 207.92], [434.35, 208.0], [468.11, 200.59], [529.66, 200.54]]}, {\"category\": \"car\", \"corners_3d\": [[585.97, 201.79], [563.77, 201.81], [565.9, 199.8], [586.57, 199.79], [585.97, 178.42], [563.77, 178.43], [565.9, 178.04], [586.57, 178.04]]}]\n```" - } - ] - }, - { - "id": 3038, - "data_source": "Mono3DRefer", - "scene_name": "004399", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[508.74, 241.66], [453.93, 242.33], [461.59, 233.56], [509.48, 233.05], [508.74, 190.97], [453.93, 191.15], [461.59, 188.84], [509.48, 188.7]]}]\n```" - } - ] - }, - { - "id": 3039, - "data_source": "Mono3DRefer", - "scene_name": "004400", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[846.26, 326.19], [818.68, 314.22], [901.37, 312.07], [935.73, 323.63], [846.26, 128.67], [818.68, 132.93], [901.37, 133.69], [935.73, 129.58]]}]\n```" - } - ] - }, - { - "id": 3040, - "data_source": "Mono3DRefer", - "scene_name": "004410", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[834.41, 387.32], [797.96, 402.92], [685.04, 374.84], [724.15, 362.82], [834.41, 139.99], [797.96, 136.94], [685.04, 142.44], [724.15, 144.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[754.53, 396.93], [696.4, 421.05], [604.25, 394.54], [664.86, 375.23], [754.53, 127.58], [696.4, 121.68], [604.25, 128.16], [664.86, 132.88]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[478.36, 334.39], [498.54, 322.24], [591.96, 325.78], [579.46, 338.57], [478.36, 153.9], [498.54, 156.0], [591.96, 155.39], [579.46, 153.18]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[597.06, 226.79], [576.27, 227.31], [564.78, 225.14], [584.73, 224.66], [597.06, 144.79], [576.27, 144.39], [564.78, 146.07], [584.73, 146.44]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[555.19, 208.24], [531.21, 208.43], [528.25, 207.38], [551.35, 207.2], [555.19, 133.99], [531.21, 133.65], [528.25, 135.42], [551.35, 135.73]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[579.95, 207.37], [557.5, 207.57], [554.12, 206.85], [576.01, 206.66], [579.95, 142.52], [557.5, 142.22], [554.12, 143.24], [576.01, 143.52]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[627.64, 206.9], [610.74, 207.08], [603.58, 206.18], [619.96, 206.01], [627.64, 145.92], [610.74, 145.68], [603.58, 146.87], [619.96, 147.09]]}]\n```" - } - ] - }, - { - "id": 3041, - "data_source": "Mono3DRefer", - "scene_name": "004410", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[507.69, 185.35], [497.87, 185.29], [522.09, 185.11], [531.81, 185.16], [507.69, 142.86], [497.87, 143.31], [522.09, 144.71], [531.81, 144.3]]}]\n```" - } - ] - }, - { - "id": 3042, - "data_source": "Mono3DRefer", - "scene_name": "004410", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[834.41, 387.32], [797.96, 402.92], [685.04, 374.84], [724.15, 362.82], [834.41, 139.99], [797.96, 136.94], [685.04, 142.44], [724.15, 144.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[754.53, 396.93], [696.4, 421.05], [604.25, 394.54], [664.86, 375.23], [754.53, 127.58], [696.4, 121.68], [604.25, 128.16], [664.86, 132.88]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[478.36, 334.39], [498.54, 322.24], [591.96, 325.78], [579.46, 338.57], [478.36, 153.9], [498.54, 156.0], [591.96, 155.39], [579.46, 153.18]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[597.06, 226.79], [576.27, 227.31], [564.78, 225.14], [584.73, 224.66], [597.06, 144.79], [576.27, 144.39], [564.78, 146.07], [584.73, 146.44]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[555.19, 208.24], [531.21, 208.43], [528.25, 207.38], [551.35, 207.2], [555.19, 133.99], [531.21, 133.65], [528.25, 135.42], [551.35, 135.73]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[579.95, 207.37], [557.5, 207.57], [554.12, 206.85], [576.01, 206.66], [579.95, 142.52], [557.5, 142.22], [554.12, 143.24], [576.01, 143.52]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[627.64, 206.9], [610.74, 207.08], [603.58, 206.18], [619.96, 206.01], [627.64, 145.92], [610.74, 145.68], [603.58, 146.87], [619.96, 147.09]]}, {\"category\": \"cyclist\", \"corners_3d\": [[507.69, 185.35], [497.87, 185.29], [522.09, 185.11], [531.81, 185.16], [507.69, 142.86], [497.87, 143.31], [522.09, 144.71], [531.81, 144.3]]}]\n```" - } - ] - }, - { - "id": 3043, - "data_source": "Mono3DRefer", - "scene_name": "004432", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[560.71, 257.62], [612.72, 265.51], [432.68, 285.06], [389.04, 273.7], [560.71, 187.71], [612.72, 189.09], [432.68, 192.52], [389.04, 190.53]]}, {\"category\": \"car\", \"corners_3d\": [[481.9, 246.11], [533.46, 252.55], [338.46, 271.85], [294.85, 262.1], [481.9, 172.0], [533.46, 171.93], [338.46, 171.7], [294.85, 171.82]]}, {\"category\": \"car\", \"corners_3d\": [[172.34, 224.93], [125.24, 226.9], [107.97, 221.16], [150.79, 219.57], [172.34, 191.11], [125.24, 191.8], [107.97, 189.79], [150.79, 189.23]]}]\n```" - } - ] - }, - { - "id": 3044, - "data_source": "Mono3DRefer", - "scene_name": "004443", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[500.64, 246.77], [463.38, 247.34], [464.77, 240.33], [498.08, 239.87], [500.64, 163.37], [463.38, 163.21], [464.77, 165.16], [498.08, 165.29]]}]\n```" - } - ] - }, - { - "id": 3045, - "data_source": "Mono3DRefer", - "scene_name": "004443", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[-14.49, 468.76], [136.67, 432.15], [205.92, 471.59], [38.84, 521.79], [-14.49, 158.32], [136.67, 161.28], [205.92, 158.09], [38.84, 154.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[142.96, 411.57], [241.25, 394.33], [276.79, 425.21], [166.24, 448.08], [142.96, 174.24], [241.25, 174.78], [276.79, 173.81], [166.24, 173.08]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[-19.38, 397.72], [90.64, 383.1], [90.97, 413.15], [-36.8, 432.65], [-19.38, 172.08], [90.64, 172.72], [90.97, 171.41], [-36.8, 170.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[268.81, 334.47], [338.15, 321.78], [390.72, 331.45], [321.25, 346.05], [268.81, 162.23], [338.15, 163.83], [390.72, 162.61], [321.25, 160.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[234.23, 326.99], [302.11, 320.91], [313.67, 335.26], [239.02, 342.7], [234.23, 169.47], [302.11, 169.97], [313.67, 168.79], [239.02, 168.17]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[696.89, 224.64], [658.68, 224.9], [653.73, 223.35], [690.6, 223.1], [696.89, 158.4], [658.68, 158.26], [653.73, 159.1], [690.6, 159.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[520.67, 214.08], [542.47, 213.98], [542.9, 215.01], [520.4, 215.12], [520.67, 169.31], [542.47, 169.35], [542.9, 168.96], [520.4, 168.92]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[531.98, 213.37], [555.15, 213.26], [556.32, 214.37], [532.35, 214.49], [531.98, 166.81], [555.15, 166.86], [556.32, 166.35], [532.35, 166.3]]}]\n```" - } - ] - }, - { - "id": 3046, - "data_source": "Mono3DRefer", - "scene_name": "004443", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[390.61, 288.67], [272.81, 289.72], [353.83, 258.84], [437.67, 258.3], [390.61, 180.72], [272.81, 180.71], [353.83, 180.94], [437.67, 180.95]]}]\n```" - } - ] - }, - { - "id": 3047, - "data_source": "Mono3DRefer", - "scene_name": "004443", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[-14.49, 468.76], [136.67, 432.15], [205.92, 471.59], [38.84, 521.79], [-14.49, 158.32], [136.67, 161.28], [205.92, 158.09], [38.84, 154.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[142.96, 411.57], [241.25, 394.33], [276.79, 425.21], [166.24, 448.08], [142.96, 174.24], [241.25, 174.78], [276.79, 173.81], [166.24, 173.08]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[-19.38, 397.72], [90.64, 383.1], [90.97, 413.15], [-36.8, 432.65], [-19.38, 172.08], [90.64, 172.72], [90.97, 171.41], [-36.8, 170.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[268.81, 334.47], [338.15, 321.78], [390.72, 331.45], [321.25, 346.05], [268.81, 162.23], [338.15, 163.83], [390.72, 162.61], [321.25, 160.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[234.23, 326.99], [302.11, 320.91], [313.67, 335.26], [239.02, 342.7], [234.23, 169.47], [302.11, 169.97], [313.67, 168.79], [239.02, 168.17]]}, {\"category\": \"cyclist\", \"corners_3d\": [[500.64, 246.77], [463.38, 247.34], [464.77, 240.33], [498.08, 239.87], [500.64, 163.37], [463.38, 163.21], [464.77, 165.16], [498.08, 165.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[696.89, 224.64], [658.68, 224.9], [653.73, 223.35], [690.6, 223.1], [696.89, 158.4], [658.68, 158.26], [653.73, 159.1], [690.6, 159.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[520.67, 214.08], [542.47, 213.98], [542.9, 215.01], [520.4, 215.12], [520.67, 169.31], [542.47, 169.35], [542.9, 168.96], [520.4, 168.92]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[531.98, 213.37], [555.15, 213.26], [556.32, 214.37], [532.35, 214.49], [531.98, 166.81], [555.15, 166.86], [556.32, 166.35], [532.35, 166.3]]}]\n```" - } - ] - }, - { - "id": 3048, - "data_source": "Mono3DRefer", - "scene_name": "004443", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[390.61, 288.67], [272.81, 289.72], [353.83, 258.84], [437.67, 258.3], [390.61, 180.72], [272.81, 180.71], [353.83, 180.94], [437.67, 180.95]]}, {\"category\": \"cyclist\", \"corners_3d\": [[500.64, 246.77], [463.38, 247.34], [464.77, 240.33], [498.08, 239.87], [500.64, 163.37], [463.38, 163.21], [464.77, 165.16], [498.08, 165.29]]}]\n```" - } - ] - }, - { - "id": 3049, - "data_source": "Mono3DRefer", - "scene_name": "004443", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[-14.49, 468.76], [136.67, 432.15], [205.92, 471.59], [38.84, 521.79], [-14.49, 158.32], [136.67, 161.28], [205.92, 158.09], [38.84, 154.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[142.96, 411.57], [241.25, 394.33], [276.79, 425.21], [166.24, 448.08], [142.96, 174.24], [241.25, 174.78], [276.79, 173.81], [166.24, 173.08]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[-19.38, 397.72], [90.64, 383.1], [90.97, 413.15], [-36.8, 432.65], [-19.38, 172.08], [90.64, 172.72], [90.97, 171.41], [-36.8, 170.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[268.81, 334.47], [338.15, 321.78], [390.72, 331.45], [321.25, 346.05], [268.81, 162.23], [338.15, 163.83], [390.72, 162.61], [321.25, 160.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[234.23, 326.99], [302.11, 320.91], [313.67, 335.26], [239.02, 342.7], [234.23, 169.47], [302.11, 169.97], [313.67, 168.79], [239.02, 168.17]]}, {\"category\": \"car\", \"corners_3d\": [[390.61, 288.67], [272.81, 289.72], [353.83, 258.84], [437.67, 258.3], [390.61, 180.72], [272.81, 180.71], [353.83, 180.94], [437.67, 180.95]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[696.89, 224.64], [658.68, 224.9], [653.73, 223.35], [690.6, 223.1], [696.89, 158.4], [658.68, 158.26], [653.73, 159.1], [690.6, 159.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[520.67, 214.08], [542.47, 213.98], [542.9, 215.01], [520.4, 215.12], [520.67, 169.31], [542.47, 169.35], [542.9, 168.96], [520.4, 168.92]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[531.98, 213.37], [555.15, 213.26], [556.32, 214.37], [532.35, 214.49], [531.98, 166.81], [555.15, 166.86], [556.32, 166.35], [532.35, 166.3]]}]\n```" - } - ] - }, - { - "id": 3050, - "data_source": "Mono3DRefer", - "scene_name": "004443", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[-14.49, 468.76], [136.67, 432.15], [205.92, 471.59], [38.84, 521.79], [-14.49, 158.32], [136.67, 161.28], [205.92, 158.09], [38.84, 154.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[142.96, 411.57], [241.25, 394.33], [276.79, 425.21], [166.24, 448.08], [142.96, 174.24], [241.25, 174.78], [276.79, 173.81], [166.24, 173.08]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[-19.38, 397.72], [90.64, 383.1], [90.97, 413.15], [-36.8, 432.65], [-19.38, 172.08], [90.64, 172.72], [90.97, 171.41], [-36.8, 170.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[268.81, 334.47], [338.15, 321.78], [390.72, 331.45], [321.25, 346.05], [268.81, 162.23], [338.15, 163.83], [390.72, 162.61], [321.25, 160.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[234.23, 326.99], [302.11, 320.91], [313.67, 335.26], [239.02, 342.7], [234.23, 169.47], [302.11, 169.97], [313.67, 168.79], [239.02, 168.17]]}, {\"category\": \"car\", \"corners_3d\": [[390.61, 288.67], [272.81, 289.72], [353.83, 258.84], [437.67, 258.3], [390.61, 180.72], [272.81, 180.71], [353.83, 180.94], [437.67, 180.95]]}, {\"category\": \"cyclist\", \"corners_3d\": [[500.64, 246.77], [463.38, 247.34], [464.77, 240.33], [498.08, 239.87], [500.64, 163.37], [463.38, 163.21], [464.77, 165.16], [498.08, 165.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[696.89, 224.64], [658.68, 224.9], [653.73, 223.35], [690.6, 223.1], [696.89, 158.4], [658.68, 158.26], [653.73, 159.1], [690.6, 159.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[520.67, 214.08], [542.47, 213.98], [542.9, 215.01], [520.4, 215.12], [520.67, 169.31], [542.47, 169.35], [542.9, 168.96], [520.4, 168.92]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[531.98, 213.37], [555.15, 213.26], [556.32, 214.37], [532.35, 214.49], [531.98, 166.81], [555.15, 166.86], [556.32, 166.35], [532.35, 166.3]]}]\n```" - } - ] - }, - { - "id": 3051, - "data_source": "Mono3DRefer", - "scene_name": "004446", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[537.64, 222.89], [498.03, 222.87], [512.81, 216.61], [547.47, 216.63], [537.64, 185.29], [498.03, 185.29], [512.81, 183.73], [547.47, 183.74]]}, {\"category\": \"car\", \"corners_3d\": [[186.94, 434.35], [-37.86, 433.6], [252.67, 317.78], [377.92, 318.01], [186.94, 194.6], [-37.86, 194.54], [252.67, 184.9], [377.92, 184.92]]}]\n```" - } - ] - }, - { - "id": 3052, - "data_source": "Mono3DRefer", - "scene_name": "004446", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[373.41, 222.81], [361.18, 222.73], [375.65, 220.42], [387.33, 220.49], [373.41, 185.28], [361.18, 185.26], [375.65, 184.68], [387.33, 184.7]]}]\n```" - } - ] - }, - { - "id": 3053, - "data_source": "Mono3DRefer", - "scene_name": "004446", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[373.41, 222.81], [361.18, 222.73], [375.65, 220.42], [387.33, 220.49], [373.41, 185.28], [361.18, 185.26], [375.65, 184.68], [387.33, 184.7]]}, {\"category\": \"car\", \"corners_3d\": [[537.64, 222.89], [498.03, 222.87], [512.81, 216.61], [547.47, 216.63], [537.64, 185.29], [498.03, 185.29], [512.81, 183.73], [547.47, 183.74]]}, {\"category\": \"car\", \"corners_3d\": [[186.94, 434.35], [-37.86, 433.6], [252.67, 317.78], [377.92, 318.01], [186.94, 194.6], [-37.86, 194.54], [252.67, 184.9], [377.92, 184.92]]}]\n```" - } - ] - }, - { - "id": 3054, - "data_source": "Mono3DRefer", - "scene_name": "004453", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[593.29, 212.69], [556.22, 212.4], [574.55, 207.71], [607.31, 207.93], [593.29, 180.27], [556.22, 180.22], [574.55, 179.35], [607.31, 179.39]]}]\n```" - } - ] - }, - { - "id": 3055, - "data_source": "Mono3DRefer", - "scene_name": "004457", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[561.5, 205.72], [537.14, 205.7], [543.85, 203.14], [566.32, 203.16], [561.5, 178.59], [537.14, 178.59], [543.85, 178.14], [566.32, 178.15]]}, {\"category\": \"car\", \"corners_3d\": [[523.84, 240.15], [458.82, 240.21], [484.41, 228.15], [537.78, 228.11], [523.84, 171.8], [458.82, 171.8], [484.41, 171.99], [537.78, 171.99]]}]\n```" - } - ] - }, - { - "id": 3056, - "data_source": "Mono3DRefer", - "scene_name": "004464", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[548.12, 224.71], [501.53, 224.81], [513.11, 217.8], [553.41, 217.72], [548.12, 175.82], [501.53, 175.83], [513.11, 175.43], [553.41, 175.42]]}, {\"category\": \"car\", \"corners_3d\": [[569.36, 203.45], [548.78, 203.43], [553.75, 201.4], [572.96, 201.41], [569.36, 180.53], [548.78, 180.53], [553.75, 180.02], [572.96, 180.02]]}]\n```" - } - ] - }, - { - "id": 3057, - "data_source": "Mono3DRefer", - "scene_name": "004466", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-164.02, 407.12], [-365.97, 407.18], [9.77, 316.87], [133.85, 316.85], [-164.02, 221.68], [-365.97, 221.69], [9.77, 202.87], [133.85, 202.86]]}, {\"category\": \"car\", \"corners_3d\": [[611.91, 199.6], [633.83, 199.61], [635.04, 201.58], [611.51, 201.57], [611.91, 178.9], [633.83, 178.9], [635.04, 179.35], [611.51, 179.34]]}, {\"category\": \"car\", \"corners_3d\": [[494.34, 212.32], [466.41, 212.05], [488.93, 208.74], [514.62, 208.96], [494.34, 183.54], [466.41, 183.47], [488.93, 182.57], [514.62, 182.64]]}, {\"category\": \"car\", \"corners_3d\": [[517.42, 207.18], [492.1, 207.0], [510.8, 204.17], [534.1, 204.33], [517.42, 175.32], [492.1, 175.31], [510.8, 175.11], [534.1, 175.12]]}, {\"category\": \"car\", \"corners_3d\": [[656.21, 197.62], [679.68, 197.67], [681.79, 199.67], [656.43, 199.61], [656.21, 175.51], [679.68, 175.51], [681.79, 175.73], [656.43, 175.72]]}]\n```" - } - ] - }, - { - "id": 3058, - "data_source": "Mono3DRefer", - "scene_name": "004475", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[651.92, 193.2], [678.03, 193.2], [682.83, 193.89], [654.47, 193.88], [651.92, 169.5], [678.03, 169.49], [682.83, 168.14], [654.47, 168.15]]}, {\"category\": \"car\", \"corners_3d\": [[592.19, 232.23], [649.12, 232.27], [656.38, 242.27], [587.36, 242.21], [592.19, 181.53], [649.12, 181.52], [656.38, 180.74], [587.36, 180.74]]}, {\"category\": \"car\", \"corners_3d\": [[586.83, 214.95], [630.92, 214.92], [636.97, 219.71], [585.78, 219.75], [586.83, 177.48], [630.92, 177.49], [636.97, 176.24], [585.78, 176.23]]}, {\"category\": \"car\", \"corners_3d\": [[687.6, 184.88], [702.57, 184.88], [686.82, 184.87], [671.46, 184.87], [687.6, 168.76], [702.57, 168.54], [686.82, 167.86], [671.46, 168.09]]}]\n```" - } - ] - }, - { - "id": 3059, - "data_source": "Mono3DRefer", - "scene_name": "004497", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[388.6, 237.3], [365.85, 242.85], [205.42, 241.89], [240.45, 236.49], [388.6, 190.92], [365.85, 192.47], [205.42, 192.2], [240.45, 190.69]]}]\n```" - } - ] - }, - { - "id": 3060, - "data_source": "Mono3DRefer", - "scene_name": "004523", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[529.29, 214.65], [493.76, 214.7], [504.35, 210.3], [536.14, 210.27], [529.29, 179.64], [493.76, 179.65], [504.35, 178.94], [536.14, 178.93]]}]\n```" - } - ] - }, - { - "id": 3061, - "data_source": "Mono3DRefer", - "scene_name": "004535", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[571.95, 202.26], [548.82, 202.26], [553.24, 200.1], [574.67, 200.1], [571.95, 181.94], [548.82, 181.94], [553.24, 181.27], [574.67, 181.27]]}]\n```" - } - ] - }, - { - "id": 3062, - "data_source": "Mono3DRefer", - "scene_name": "004540", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[484.66, 302.2], [366.11, 302.43], [437.14, 263.38], [519.92, 263.26], [484.66, 193.22], [366.11, 193.26], [437.14, 187.11], [519.92, 187.09]]}, {\"category\": \"car\", \"corners_3d\": [[678.07, 233.73], [727.63, 233.61], [759.89, 247.69], [698.89, 247.88], [678.07, 181.63], [727.63, 181.62], [759.89, 183.65], [698.89, 183.67]]}, {\"category\": \"car\", \"corners_3d\": [[661.0, 222.78], [701.53, 222.64], [722.79, 230.94], [675.55, 231.13], [661.0, 182.21], [701.53, 182.18], [722.79, 183.74], [675.55, 183.77]]}, {\"category\": \"car\", \"corners_3d\": [[694.59, 196.7], [673.79, 196.73], [666.89, 195.1], [686.29, 195.07], [694.59, 174.92], [673.79, 174.92], [666.89, 174.78], [686.29, 174.78]]}]\n```" - } - ] - }, - { - "id": 3063, - "data_source": "Mono3DRefer", - "scene_name": "004540", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[720.71, 256.93], [806.28, 256.54], [908.56, 294.58], [784.54, 295.41], [720.71, 158.57], [806.28, 158.64], [908.56, 152.18], [784.54, 152.04]]}]\n```" - } - ] - }, - { - "id": 3064, - "data_source": "Mono3DRefer", - "scene_name": "004540", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[484.66, 302.2], [366.11, 302.43], [437.14, 263.38], [519.92, 263.26], [484.66, 193.22], [366.11, 193.26], [437.14, 187.11], [519.92, 187.09]]}, {\"category\": \"van\", \"corners_3d\": [[720.71, 256.93], [806.28, 256.54], [908.56, 294.58], [784.54, 295.41], [720.71, 158.57], [806.28, 158.64], [908.56, 152.18], [784.54, 152.04]]}, {\"category\": \"car\", \"corners_3d\": [[678.07, 233.73], [727.63, 233.61], [759.89, 247.69], [698.89, 247.88], [678.07, 181.63], [727.63, 181.62], [759.89, 183.65], [698.89, 183.67]]}, {\"category\": \"car\", \"corners_3d\": [[661.0, 222.78], [701.53, 222.64], [722.79, 230.94], [675.55, 231.13], [661.0, 182.21], [701.53, 182.18], [722.79, 183.74], [675.55, 183.77]]}, {\"category\": \"car\", \"corners_3d\": [[694.59, 196.7], [673.79, 196.73], [666.89, 195.1], [686.29, 195.07], [694.59, 174.92], [673.79, 174.92], [666.89, 174.78], [686.29, 174.78]]}]\n```" - } - ] - }, - { - "id": 3065, - "data_source": "Mono3DRefer", - "scene_name": "004550", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[66.1, 272.18], [-7.85, 272.67], [104.37, 253.38], [163.63, 253.06], [66.1, 208.32], [-7.85, 208.49], [104.37, 201.61], [163.63, 201.49]]}]\n```" - } - ] - }, - { - "id": 3066, - "data_source": "Mono3DRefer", - "scene_name": "004550", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[354.23, 243.2], [425.53, 243.13], [363.1, 267.87], [266.59, 267.99], [354.23, 152.91], [425.53, 152.92], [363.1, 145.91], [266.59, 145.88]]}]\n```" - } - ] - }, - { - "id": 3067, - "data_source": "Mono3DRefer", - "scene_name": "004550", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[354.23, 243.2], [425.53, 243.13], [363.1, 267.87], [266.59, 267.99], [354.23, 152.91], [425.53, 152.92], [363.1, 145.91], [266.59, 145.88]]}, {\"category\": \"car\", \"corners_3d\": [[66.1, 272.18], [-7.85, 272.67], [104.37, 253.38], [163.63, 253.06], [66.1, 208.32], [-7.85, 208.49], [104.37, 201.61], [163.63, 201.49]]}]\n```" - } - ] - }, - { - "id": 3068, - "data_source": "Mono3DRefer", - "scene_name": "004554", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[316.99, 243.57], [341.35, 237.98], [465.33, 238.23], [451.56, 243.86], [316.99, 175.41], [341.35, 175.2], [465.33, 175.21], [451.56, 175.42]]}, {\"category\": \"car\", \"corners_3d\": [[361.18, 230.59], [377.91, 226.82], [490.93, 226.92], [482.06, 230.7], [361.18, 180.58], [377.91, 180.08], [490.93, 180.09], [482.06, 180.6]]}, {\"category\": \"car\", \"corners_3d\": [[791.59, 198.64], [826.9, 198.69], [851.88, 202.13], [811.83, 202.06], [791.59, 162.85], [826.9, 162.83], [851.88, 161.49], [811.83, 161.52]]}, {\"category\": \"car\", \"corners_3d\": [[401.88, 220.1], [412.32, 217.71], [508.61, 217.7], [503.3, 220.08], [401.88, 183.67], [412.32, 183.13], [508.61, 183.12], [503.3, 183.67]]}, {\"category\": \"car\", \"corners_3d\": [[426.89, 213.51], [436.13, 211.54], [527.16, 211.6], [522.54, 213.57], [426.89, 177.04], [436.13, 176.84], [527.16, 176.84], [522.54, 177.04]]}]\n```" - } - ] - }, - { - "id": 3069, - "data_source": "Mono3DRefer", - "scene_name": "004559", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[834.52, 314.0], [692.5, 310.01], [700.43, 267.38], [797.59, 269.25], [834.52, 185.88], [692.5, 185.51], [700.43, 181.58], [797.59, 181.75]]}]\n```" - } - ] - }, - { - "id": 3070, - "data_source": "Mono3DRefer", - "scene_name": "004562", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[246.49, 282.67], [291.83, 277.9], [306.95, 280.59], [261.05, 285.64], [246.49, 155.76], [291.83, 156.97], [306.95, 156.29], [261.05, 155.0]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[593.58, 248.19], [558.85, 245.4], [588.27, 242.55], [622.61, 245.09], [593.58, 176.2], [558.85, 176.42], [588.27, 176.65], [622.61, 176.44]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[502.91, 248.27], [477.87, 250.34], [458.04, 248.08], [482.86, 246.14], [502.91, 170.69], [477.87, 170.36], [458.04, 170.72], [482.86, 171.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[565.52, 232.62], [532.11, 233.03], [529.61, 230.71], [561.54, 230.33], [565.52, 167.43], [532.11, 167.32], [529.61, 167.96], [561.54, 168.06]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[609.86, 219.16], [637.79, 219.22], [638.11, 220.72], [609.07, 220.66], [609.86, 166.63], [637.79, 166.61], [638.11, 166.01], [609.07, 166.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[615.82, 209.64], [599.89, 209.58], [601.62, 208.83], [617.12, 208.89], [615.82, 168.35], [599.89, 168.38], [601.62, 168.73], [617.12, 168.7]]}]\n```" - } - ] - }, - { - "id": 3071, - "data_source": "Mono3DRefer", - "scene_name": "004562", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[772.11, 311.46], [957.09, 312.48], [1854.57, 663.79], [1162.07, 650.2], [772.11, 84.93], [957.09, 84.17], [1854.57, -176.95], [1162.07, -166.84]]}]\n```" - } - ] - }, - { - "id": 3072, - "data_source": "Mono3DRefer", - "scene_name": "004562", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[246.49, 282.67], [291.83, 277.9], [306.95, 280.59], [261.05, 285.64], [246.49, 155.76], [291.83, 156.97], [306.95, 156.29], [261.05, 155.0]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[593.58, 248.19], [558.85, 245.4], [588.27, 242.55], [622.61, 245.09], [593.58, 176.2], [558.85, 176.42], [588.27, 176.65], [622.61, 176.44]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[502.91, 248.27], [477.87, 250.34], [458.04, 248.08], [482.86, 246.14], [502.91, 170.69], [477.87, 170.36], [458.04, 170.72], [482.86, 171.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[565.52, 232.62], [532.11, 233.03], [529.61, 230.71], [561.54, 230.33], [565.52, 167.43], [532.11, 167.32], [529.61, 167.96], [561.54, 168.06]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[609.86, 219.16], [637.79, 219.22], [638.11, 220.72], [609.07, 220.66], [609.86, 166.63], [637.79, 166.61], [638.11, 166.01], [609.07, 166.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[615.82, 209.64], [599.89, 209.58], [601.62, 208.83], [617.12, 208.89], [615.82, 168.35], [599.89, 168.38], [601.62, 168.73], [617.12, 168.7]]}, {\"category\": \"van\", \"corners_3d\": [[772.11, 311.46], [957.09, 312.48], [1854.57, 663.79], [1162.07, 650.2], [772.11, 84.93], [957.09, 84.17], [1854.57, -176.95], [1162.07, -166.84]]}]\n```" - } - ] - }, - { - "id": 3073, - "data_source": "Mono3DRefer", - "scene_name": "004566", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[623.38, 227.65], [593.59, 227.65], [595.07, 222.36], [621.99, 222.36], [623.38, 200.06], [593.59, 200.06], [595.07, 197.44], [621.99, 197.44]]}]\n```" - } - ] - }, - { - "id": 3074, - "data_source": "Mono3DRefer", - "scene_name": "004580", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[469.15, 301.32], [458.35, 308.98], [374.38, 308.21], [390.1, 300.64], [469.15, 159.26], [458.35, 157.91], [374.38, 158.05], [390.1, 159.38]]}]\n```" - } - ] - }, - { - "id": 3075, - "data_source": "Mono3DRefer", - "scene_name": "004587", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[574.65, 284.58], [682.36, 284.59], [722.53, 346.73], [554.92, 346.7], [574.65, 184.27], [682.36, 184.28], [722.53, 190.63], [554.92, 190.62]]}, {\"category\": \"car\", \"corners_3d\": [[119.97, 371.12], [318.21, 371.16], [-282.86, 779.49], [-888.62, 779.09], [119.97, 180.02], [318.21, 180.02], [-282.86, 194.79], [-888.62, 194.77]]}, {\"category\": \"car\", \"corners_3d\": [[455.48, 230.44], [515.87, 230.45], [491.9, 245.09], [416.16, 245.09], [455.48, 179.47], [515.87, 179.47], [491.9, 181.16], [416.16, 181.16]]}, {\"category\": \"car\", \"corners_3d\": [[694.63, 212.09], [734.85, 212.09], [755.21, 218.5], [708.42, 218.5], [694.63, 177.69], [734.85, 177.69], [755.21, 178.48], [708.42, 178.48]]}]\n```" - } - ] - }, - { - "id": 3076, - "data_source": "Mono3DRefer", - "scene_name": "004587", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[385.64, 254.43], [491.26, 254.44], [418.32, 304.5], [247.94, 304.48], [385.64, 146.81], [491.26, 146.8], [418.32, 130.82], [247.94, 130.83]]}, {\"category\": \"van\", \"corners_3d\": [[485.83, 215.3], [531.57, 215.25], [520.35, 222.67], [466.57, 222.74], [485.83, 168.3], [531.57, 168.31], [520.35, 167.52], [466.57, 167.51]]}, {\"category\": \"van\", \"corners_3d\": [[670.98, 201.75], [709.52, 201.75], [723.23, 205.73], [679.37, 205.73], [670.98, 165.67], [709.52, 165.67], [723.23, 164.68], [679.37, 164.68]]}]\n```" - } - ] - }, - { - "id": 3077, - "data_source": "Mono3DRefer", - "scene_name": "004587", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[574.65, 284.58], [682.36, 284.59], [722.53, 346.73], [554.92, 346.7], [574.65, 184.27], [682.36, 184.28], [722.53, 190.63], [554.92, 190.62]]}, {\"category\": \"car\", \"corners_3d\": [[119.97, 371.12], [318.21, 371.16], [-282.86, 779.49], [-888.62, 779.09], [119.97, 180.02], [318.21, 180.02], [-282.86, 194.79], [-888.62, 194.77]]}, {\"category\": \"van\", \"corners_3d\": [[385.64, 254.43], [491.26, 254.44], [418.32, 304.5], [247.94, 304.48], [385.64, 146.81], [491.26, 146.8], [418.32, 130.82], [247.94, 130.83]]}, {\"category\": \"car\", \"corners_3d\": [[455.48, 230.44], [515.87, 230.45], [491.9, 245.09], [416.16, 245.09], [455.48, 179.47], [515.87, 179.47], [491.9, 181.16], [416.16, 181.16]]}, {\"category\": \"van\", \"corners_3d\": [[485.83, 215.3], [531.57, 215.25], [520.35, 222.67], [466.57, 222.74], [485.83, 168.3], [531.57, 168.31], [520.35, 167.52], [466.57, 167.51]]}, {\"category\": \"car\", \"corners_3d\": [[694.63, 212.09], [734.85, 212.09], [755.21, 218.5], [708.42, 218.5], [694.63, 177.69], [734.85, 177.69], [755.21, 178.48], [708.42, 178.48]]}, {\"category\": \"van\", \"corners_3d\": [[670.98, 201.75], [709.52, 201.75], [723.23, 205.73], [679.37, 205.73], [670.98, 165.67], [709.52, 165.67], [723.23, 164.68], [679.37, 164.68]]}]\n```" - } - ] - }, - { - "id": 3078, - "data_source": "Mono3DRefer", - "scene_name": "004592", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[605.19, 199.12], [581.92, 199.09], [586.05, 197.15], [607.6, 197.18], [605.19, 179.17], [581.92, 179.16], [586.05, 178.7], [607.6, 178.7]]}, {\"category\": \"car\", \"corners_3d\": [[1258.05, 240.36], [1284.58, 247.4], [997.55, 236.68], [998.05, 231.45], [1258.05, 128.26], [1284.58, 123.61], [997.55, 130.69], [998.05, 134.15]]}, {\"category\": \"car\", \"corners_3d\": [[448.96, 257.37], [526.96, 257.94], [474.66, 293.1], [365.22, 291.98], [448.96, 177.78], [526.96, 177.82], [474.66, 179.87], [365.22, 179.8]]}, {\"category\": \"car\", \"corners_3d\": [[559.95, 230.21], [506.55, 230.0], [527.1, 221.52], [572.61, 221.68], [559.95, 178.93], [506.55, 178.91], [527.1, 178.01], [572.61, 178.03]]}]\n```" - } - ] - }, - { - "id": 3079, - "data_source": "Mono3DRefer", - "scene_name": "004613", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[420.33, 212.01], [410.61, 211.97], [419.75, 210.54], [429.12, 210.57], [420.33, 182.59], [410.61, 182.58], [419.75, 182.23], [429.12, 182.23]]}]\n```" - } - ] - }, - { - "id": 3080, - "data_source": "Mono3DRefer", - "scene_name": "004613", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[553.02, 208.12], [525.02, 208.13], [532.06, 204.9], [557.48, 204.89], [553.02, 181.62], [525.02, 181.62], [532.06, 180.82], [557.48, 180.82]]}, {\"category\": \"car\", \"corners_3d\": [[471.01, 257.1], [400.75, 256.94], [445.19, 240.15], [501.47, 240.26], [471.01, 182.35], [400.75, 182.33], [445.19, 180.44], [501.47, 180.45]]}]\n```" - } - ] - }, - { - "id": 3081, - "data_source": "Mono3DRefer", - "scene_name": "004613", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[420.33, 212.01], [410.61, 211.97], [419.75, 210.54], [429.12, 210.57], [420.33, 182.59], [410.61, 182.58], [419.75, 182.23], [429.12, 182.23]]}, {\"category\": \"car\", \"corners_3d\": [[553.02, 208.12], [525.02, 208.13], [532.06, 204.9], [557.48, 204.89], [553.02, 181.62], [525.02, 181.62], [532.06, 180.82], [557.48, 180.82]]}, {\"category\": \"car\", \"corners_3d\": [[471.01, 257.1], [400.75, 256.94], [445.19, 240.15], [501.47, 240.26], [471.01, 182.35], [400.75, 182.33], [445.19, 180.44], [501.47, 180.45]]}]\n```" - } - ] - }, - { - "id": 3082, - "data_source": "Mono3DRefer", - "scene_name": "004618", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[463.49, 253.36], [382.95, 253.55], [430.2, 235.53], [492.7, 235.42], [463.49, 179.26], [382.95, 179.27], [430.2, 177.84], [492.7, 177.83]]}]\n```" - } - ] - }, - { - "id": 3083, - "data_source": "Mono3DRefer", - "scene_name": "004628", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[553.95, 194.21], [537.01, 194.22], [540.25, 193.01], [556.24, 193.0], [553.95, 176.32], [537.01, 176.32], [540.25, 176.13], [556.24, 176.12]]}]\n```" - } - ] - }, - { - "id": 3084, - "data_source": "Mono3DRefer", - "scene_name": "004636", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[216.36, 470.58], [-82.54, 465.64], [298.51, 310.77], [441.07, 311.86], [216.36, 200.71], [-82.54, 200.24], [298.51, 185.76], [441.07, 185.86]]}, {\"category\": \"car\", \"corners_3d\": [[797.04, 335.42], [940.27, 334.47], [1234.48, 469.63], [973.34, 472.85], [797.04, 201.14], [940.27, 200.97], [1234.48, 224.49], [973.34, 225.05]]}, {\"category\": \"car\", \"corners_3d\": [[438.76, 285.96], [322.35, 285.61], [401.39, 256.01], [487.35, 256.2], [438.76, 177.02], [322.35, 177.01], [401.39, 175.92], [487.35, 175.92]]}, {\"category\": \"car\", \"corners_3d\": [[887.92, 312.33], [762.24, 312.35], [712.67, 267.23], [797.71, 267.22], [887.92, 201.95], [762.24, 201.96], [712.67, 192.54], [797.71, 192.54]]}, {\"category\": \"car\", \"corners_3d\": [[521.71, 230.71], [464.85, 230.76], [484.63, 222.39], [533.27, 222.36], [521.71, 175.85], [464.85, 175.85], [484.63, 175.42], [533.27, 175.41]]}, {\"category\": \"car\", \"corners_3d\": [[732.28, 223.7], [683.52, 223.67], [673.36, 216.0], [714.75, 216.02], [732.28, 181.42], [683.52, 181.42], [673.36, 180.12], [714.75, 180.13]]}, {\"category\": \"car\", \"corners_3d\": [[551.25, 216.1], [504.56, 216.15], [515.82, 210.74], [556.66, 210.69], [551.25, 178.08], [504.56, 178.09], [515.82, 177.43], [556.66, 177.43]]}, {\"category\": \"car\", \"corners_3d\": [[651.62, 198.57], [675.27, 198.59], [679.67, 200.83], [653.97, 200.81], [651.62, 175.72], [675.27, 175.73], [679.67, 175.98], [653.97, 175.97]]}]\n```" - } - ] - }, - { - "id": 3085, - "data_source": "Mono3DRefer", - "scene_name": "004636", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[644.0, 192.45], [667.43, 192.47], [671.04, 194.46], [645.24, 194.44], [644.0, 164.35], [667.43, 164.34], [671.04, 163.48], [645.24, 163.49]]}]\n```" - } - ] - }, - { - "id": 3086, - "data_source": "Mono3DRefer", - "scene_name": "004636", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[216.36, 470.58], [-82.54, 465.64], [298.51, 310.77], [441.07, 311.86], [216.36, 200.71], [-82.54, 200.24], [298.51, 185.76], [441.07, 185.86]]}, {\"category\": \"car\", \"corners_3d\": [[797.04, 335.42], [940.27, 334.47], [1234.48, 469.63], [973.34, 472.85], [797.04, 201.14], [940.27, 200.97], [1234.48, 224.49], [973.34, 225.05]]}, {\"category\": \"car\", \"corners_3d\": [[438.76, 285.96], [322.35, 285.61], [401.39, 256.01], [487.35, 256.2], [438.76, 177.02], [322.35, 177.01], [401.39, 175.92], [487.35, 175.92]]}, {\"category\": \"car\", \"corners_3d\": [[887.92, 312.33], [762.24, 312.35], [712.67, 267.23], [797.71, 267.22], [887.92, 201.95], [762.24, 201.96], [712.67, 192.54], [797.71, 192.54]]}, {\"category\": \"car\", \"corners_3d\": [[521.71, 230.71], [464.85, 230.76], [484.63, 222.39], [533.27, 222.36], [521.71, 175.85], [464.85, 175.85], [484.63, 175.42], [533.27, 175.41]]}, {\"category\": \"car\", \"corners_3d\": [[732.28, 223.7], [683.52, 223.67], [673.36, 216.0], [714.75, 216.02], [732.28, 181.42], [683.52, 181.42], [673.36, 180.12], [714.75, 180.13]]}, {\"category\": \"car\", \"corners_3d\": [[551.25, 216.1], [504.56, 216.15], [515.82, 210.74], [556.66, 210.69], [551.25, 178.08], [504.56, 178.09], [515.82, 177.43], [556.66, 177.43]]}, {\"category\": \"car\", \"corners_3d\": [[651.62, 198.57], [675.27, 198.59], [679.67, 200.83], [653.97, 200.81], [651.62, 175.72], [675.27, 175.73], [679.67, 175.98], [653.97, 175.97]]}, {\"category\": \"van\", \"corners_3d\": [[644.0, 192.45], [667.43, 192.47], [671.04, 194.46], [645.24, 194.44], [644.0, 164.35], [667.43, 164.34], [671.04, 163.48], [645.24, 163.49]]}]\n```" - } - ] - }, - { - "id": 3087, - "data_source": "Mono3DRefer", - "scene_name": "004651", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[623.5, 284.88], [651.32, 281.69], [670.57, 285.39], [642.37, 288.8], [623.5, 164.21], [651.32, 164.46], [670.57, 164.17], [642.37, 163.91]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[325.36, 245.57], [346.18, 244.05], [356.54, 245.64], [335.47, 247.23], [325.36, 175.75], [346.18, 175.69], [356.54, 175.75], [335.47, 175.81]]}]\n```" - } - ] - }, - { - "id": 3088, - "data_source": "Mono3DRefer", - "scene_name": "004651", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[646.87, 265.81], [742.69, 266.06], [787.46, 301.39], [655.29, 300.9], [646.87, 174.7], [742.69, 174.71], [787.46, 175.41], [655.29, 175.4]]}, {\"category\": \"car\", \"corners_3d\": [[248.4, 240.26], [279.66, 235.25], [393.43, 236.08], [370.99, 241.24], [248.4, 176.45], [279.66, 176.18], [393.43, 176.23], [370.99, 176.5]]}]\n```" - } - ] - }, - { - "id": 3089, - "data_source": "Mono3DRefer", - "scene_name": "004651", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[646.87, 265.81], [742.69, 266.06], [787.46, 301.39], [655.29, 300.9], [646.87, 174.7], [742.69, 174.71], [787.46, 175.41], [655.29, 175.4]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[623.5, 284.88], [651.32, 281.69], [670.57, 285.39], [642.37, 288.8], [623.5, 164.21], [651.32, 164.46], [670.57, 164.17], [642.37, 163.91]]}, {\"category\": \"car\", \"corners_3d\": [[248.4, 240.26], [279.66, 235.25], [393.43, 236.08], [370.99, 241.24], [248.4, 176.45], [279.66, 176.18], [393.43, 176.23], [370.99, 176.5]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[325.36, 245.57], [346.18, 244.05], [356.54, 245.64], [335.47, 247.23], [325.36, 175.75], [346.18, 175.69], [356.54, 175.75], [335.47, 175.81]]}]\n```" - } - ] - }, - { - "id": 3090, - "data_source": "Mono3DRefer", - "scene_name": "004654", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[698.65, 192.49], [722.8, 192.63], [719.54, 194.2], [693.51, 194.04], [698.65, 169.92], [722.8, 169.9], [719.54, 169.66], [693.51, 169.69]]}, {\"category\": \"car\", \"corners_3d\": [[-67.81, 377.3], [-233.53, 372.72], [118.06, 295.29], [223.59, 297.0], [-67.81, 151.78], [-233.53, 152.25], [118.06, 160.23], [223.59, 160.06]]}, {\"category\": \"car\", \"corners_3d\": [[745.86, 251.94], [830.52, 252.14], [906.22, 281.28], [790.29, 280.92], [745.86, 172.33], [830.52, 172.33], [906.22, 172.14], [790.29, 172.14]]}, {\"category\": \"car\", \"corners_3d\": [[624.56, 202.66], [587.9, 202.34], [605.86, 199.31], [638.89, 199.56], [624.56, 168.1], [587.9, 168.15], [605.86, 168.64], [638.89, 168.6]]}, {\"category\": \"car\", \"corners_3d\": [[739.43, 233.1], [797.38, 233.54], [828.26, 248.88], [755.51, 248.2], [739.43, 155.41], [797.38, 155.28], [828.26, 150.84], [755.51, 151.03]]}]\n```" - } - ] - }, - { - "id": 3091, - "data_source": "Mono3DRefer", - "scene_name": "004655", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[646.84, 211.04], [689.48, 210.97], [709.2, 218.41], [658.26, 218.5], [646.84, 172.84], [689.48, 172.84], [709.2, 172.84], [658.26, 172.84]]}]\n```" - } - ] - }, - { - "id": 3092, - "data_source": "Mono3DRefer", - "scene_name": "004656", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[354.21, 366.78], [158.88, 367.86], [324.52, 293.72], [445.27, 293.3], [354.21, 208.77], [158.88, 208.98], [324.52, 195.24], [445.27, 195.16]]}]\n```" - } - ] - }, - { - "id": 3093, - "data_source": "Mono3DRefer", - "scene_name": "004674", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[530.11, 302.59], [475.96, 302.69], [487.62, 290.87], [536.52, 290.79], [530.11, 157.91], [475.96, 157.89], [487.62, 160.08], [536.52, 160.09]]}]\n```" - } - ] - }, - { - "id": 3094, - "data_source": "Mono3DRefer", - "scene_name": "004679", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[731.06, 319.57], [721.84, 309.38], [819.12, 309.41], [836.03, 319.6], [731.06, 162.68], [721.84, 163.99], [819.12, 163.99], [836.03, 162.68]]}]\n```" - } - ] - }, - { - "id": 3095, - "data_source": "Mono3DRefer", - "scene_name": "004696", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[429.49, 299.94], [438.43, 293.53], [535.41, 293.35], [531.99, 299.74], [429.49, 136.13], [438.43, 138.52], [535.41, 138.58], [531.99, 136.21]]}]\n```" - } - ] - }, - { - "id": 3096, - "data_source": "Mono3DRefer", - "scene_name": "004717", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[645.39, 299.06], [675.17, 295.13], [697.22, 299.69], [667.02, 303.93], [645.39, 166.51], [675.17, 166.71], [697.22, 166.48], [667.02, 166.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[337.77, 256.26], [359.44, 254.51], [369.85, 256.46], [347.88, 258.3], [337.77, 183.64], [359.44, 183.42], [369.85, 183.67], [347.88, 183.91]]}]\n```" - } - ] - }, - { - "id": 3097, - "data_source": "Mono3DRefer", - "scene_name": "004717", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[674.07, 280.1], [778.85, 280.73], [838.72, 326.92], [688.95, 325.63], [674.07, 180.98], [778.85, 181.02], [838.72, 184.52], [688.95, 184.43]]}, {\"category\": \"car\", \"corners_3d\": [[247.98, 250.61], [282.77, 244.59], [399.73, 246.1], [374.18, 252.38], [247.98, 183.77], [282.77, 182.92], [399.73, 183.14], [374.18, 184.02]]}]\n```" - } - ] - }, - { - "id": 3098, - "data_source": "Mono3DRefer", - "scene_name": "004717", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[674.07, 280.1], [778.85, 280.73], [838.72, 326.92], [688.95, 325.63], [674.07, 180.98], [778.85, 181.02], [838.72, 184.52], [688.95, 184.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[645.39, 299.06], [675.17, 295.13], [697.22, 299.69], [667.02, 303.93], [645.39, 166.51], [675.17, 166.71], [697.22, 166.48], [667.02, 166.27]]}, {\"category\": \"car\", \"corners_3d\": [[247.98, 250.61], [282.77, 244.59], [399.73, 246.1], [374.18, 252.38], [247.98, 183.77], [282.77, 182.92], [399.73, 183.14], [374.18, 184.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[337.77, 256.26], [359.44, 254.51], [369.85, 256.46], [347.88, 258.3], [337.77, 183.64], [359.44, 183.42], [369.85, 183.67], [347.88, 183.91]]}]\n```" - } - ] - }, - { - "id": 3099, - "data_source": "Mono3DRefer", - "scene_name": "004727", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[447.51, 353.1], [257.84, 357.32], [353.06, 295.91], [479.09, 294.02], [447.51, 181.15], [257.84, 181.35], [353.06, 178.52], [479.09, 178.43]]}, {\"category\": \"car\", \"corners_3d\": [[507.81, 264.37], [428.8, 264.28], [470.42, 243.98], [531.9, 244.03], [507.81, 183.63], [428.8, 183.62], [470.42, 181.23], [531.9, 181.24]]}]\n```" - } - ] - }, - { - "id": 3100, - "data_source": "Mono3DRefer", - "scene_name": "004741", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[520.46, 265.69], [449.24, 266.96], [458.43, 252.07], [518.4, 251.17], [520.46, 199.94], [449.24, 200.31], [458.43, 195.97], [518.4, 195.7]]}, {\"category\": \"car\", \"corners_3d\": [[524.24, 244.12], [479.82, 244.21], [495.63, 234.38], [533.92, 234.31], [524.24, 199.0], [479.82, 199.03], [495.63, 195.42], [533.92, 195.4]]}]\n```" - } - ] - }, - { - "id": 3101, - "data_source": "Mono3DRefer", - "scene_name": "004751", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[199.47, 362.3], [10.7, 365.68], [227.65, 289.3], [340.13, 288.06], [199.47, 190.04], [10.7, 190.35], [227.65, 183.42], [340.13, 183.31]]}, {\"category\": \"car\", \"corners_3d\": [[1068.58, 320.14], [973.13, 291.82], [1266.15, 289.25], [1427.0, 316.24], [1068.58, 143.97], [973.13, 149.53], [1266.15, 150.03], [1427.0, 144.74]]}]\n```" - } - ] - }, - { - "id": 3102, - "data_source": "Mono3DRefer", - "scene_name": "004751", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[676.86, 207.93], [735.93, 209.6], [606.76, 224.58], [533.22, 221.34], [676.86, 131.7], [735.93, 129.74], [606.76, 112.16], [533.22, 115.97]]}]\n```" - } - ] - }, - { - "id": 3103, - "data_source": "Mono3DRefer", - "scene_name": "004751", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[932.03, 285.87], [874.67, 267.48], [1192.36, 265.72], [1308.83, 283.36], [932.03, 144.76], [874.67, 149.33], [1192.36, 149.77], [1308.83, 145.38]]}]\n```" - } - ] - }, - { - "id": 3104, - "data_source": "Mono3DRefer", - "scene_name": "004751", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[199.47, 362.3], [10.7, 365.68], [227.65, 289.3], [340.13, 288.06], [199.47, 190.04], [10.7, 190.35], [227.65, 183.42], [340.13, 183.31]]}, {\"category\": \"truck\", \"corners_3d\": [[676.86, 207.93], [735.93, 209.6], [606.76, 224.58], [533.22, 221.34], [676.86, 131.7], [735.93, 129.74], [606.76, 112.16], [533.22, 115.97]]}, {\"category\": \"car\", \"corners_3d\": [[1068.58, 320.14], [973.13, 291.82], [1266.15, 289.25], [1427.0, 316.24], [1068.58, 143.97], [973.13, 149.53], [1266.15, 150.03], [1427.0, 144.74]]}]\n```" - } - ] - }, - { - "id": 3105, - "data_source": "Mono3DRefer", - "scene_name": "004751", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[932.03, 285.87], [874.67, 267.48], [1192.36, 265.72], [1308.83, 283.36], [932.03, 144.76], [874.67, 149.33], [1192.36, 149.77], [1308.83, 145.38]]}, {\"category\": \"car\", \"corners_3d\": [[199.47, 362.3], [10.7, 365.68], [227.65, 289.3], [340.13, 288.06], [199.47, 190.04], [10.7, 190.35], [227.65, 183.42], [340.13, 183.31]]}, {\"category\": \"car\", \"corners_3d\": [[1068.58, 320.14], [973.13, 291.82], [1266.15, 289.25], [1427.0, 316.24], [1068.58, 143.97], [973.13, 149.53], [1266.15, 150.03], [1427.0, 144.74]]}]\n```" - } - ] - }, - { - "id": 3106, - "data_source": "Mono3DRefer", - "scene_name": "004751", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[932.03, 285.87], [874.67, 267.48], [1192.36, 265.72], [1308.83, 283.36], [932.03, 144.76], [874.67, 149.33], [1192.36, 149.77], [1308.83, 145.38]]}, {\"category\": \"truck\", \"corners_3d\": [[676.86, 207.93], [735.93, 209.6], [606.76, 224.58], [533.22, 221.34], [676.86, 131.7], [735.93, 129.74], [606.76, 112.16], [533.22, 115.97]]}]\n```" - } - ] - }, - { - "id": 3107, - "data_source": "Mono3DRefer", - "scene_name": "004751", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[932.03, 285.87], [874.67, 267.48], [1192.36, 265.72], [1308.83, 283.36], [932.03, 144.76], [874.67, 149.33], [1192.36, 149.77], [1308.83, 145.38]]}, {\"category\": \"car\", \"corners_3d\": [[199.47, 362.3], [10.7, 365.68], [227.65, 289.3], [340.13, 288.06], [199.47, 190.04], [10.7, 190.35], [227.65, 183.42], [340.13, 183.31]]}, {\"category\": \"truck\", \"corners_3d\": [[676.86, 207.93], [735.93, 209.6], [606.76, 224.58], [533.22, 221.34], [676.86, 131.7], [735.93, 129.74], [606.76, 112.16], [533.22, 115.97]]}, {\"category\": \"car\", \"corners_3d\": [[1068.58, 320.14], [973.13, 291.82], [1266.15, 289.25], [1427.0, 316.24], [1068.58, 143.97], [973.13, 149.53], [1266.15, 150.03], [1427.0, 144.74]]}]\n```" - } - ] - }, - { - "id": 3108, - "data_source": "Mono3DRefer", - "scene_name": "004803", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-90.73, 769.52], [-684.24, 779.52], [161.53, 378.28], [359.94, 377.12], [-90.73, 166.62], [-684.24, 166.52], [161.53, 170.71], [359.94, 170.72]]}, {\"category\": \"car\", \"corners_3d\": [[425.33, 319.57], [315.14, 319.14], [396.58, 280.49], [477.78, 280.72], [425.33, 218.77], [315.14, 218.64], [396.58, 206.54], [477.78, 206.61]]}, {\"category\": \"car\", \"corners_3d\": [[537.21, 243.78], [495.76, 243.7], [511.8, 234.8], [548.05, 234.86], [537.21, 197.58], [495.76, 197.55], [511.8, 194.45], [548.05, 194.47]]}, {\"category\": \"car\", \"corners_3d\": [[564.93, 224.98], [535.44, 225.02], [540.33, 220.71], [567.38, 220.67], [564.93, 196.12], [535.44, 196.14], [540.33, 194.22], [567.38, 194.2]]}, {\"category\": \"car\", \"corners_3d\": [[582.81, 215.88], [562.31, 215.54], [577.61, 212.89], [596.94, 213.19], [582.81, 194.23], [562.31, 194.06], [577.61, 192.74], [596.94, 192.89]]}]\n```" - } - ] - }, - { - "id": 3109, - "data_source": "Mono3DRefer", - "scene_name": "004805", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[544.47, 200.09], [513.03, 200.02], [524.9, 198.5], [553.78, 198.56], [544.47, 174.69], [513.03, 174.71], [524.9, 175.27], [553.78, 175.25]]}]\n```" - } - ] - }, - { - "id": 3110, - "data_source": "Mono3DRefer", - "scene_name": "004809", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[446.55, 219.51], [408.19, 218.93], [455.92, 212.85], [489.69, 213.29], [446.55, 183.88], [408.19, 183.75], [455.92, 182.31], [489.69, 182.41]]}, {\"category\": \"car\", \"corners_3d\": [[724.11, 193.86], [746.94, 194.32], [717.68, 195.8], [693.98, 195.28], [724.11, 170.07], [746.94, 170.01], [717.68, 169.81], [693.98, 169.88]]}, {\"category\": \"car\", \"corners_3d\": [[-11.63, 263.51], [-74.2, 262.36], [102.87, 241.98], [152.8, 242.66], [-11.63, 193.95], [-74.2, 193.68], [102.87, 188.94], [152.8, 189.09]]}, {\"category\": \"car\", \"corners_3d\": [[636.32, 227.96], [692.79, 228.3], [698.03, 240.03], [629.67, 239.52], [636.32, 174.88], [692.79, 174.9], [698.03, 175.33], [629.67, 175.31]]}, {\"category\": \"car\", \"corners_3d\": [[1001.72, 273.66], [1143.25, 287.24], [1044.11, 331.45], [872.08, 306.5], [1001.72, 173.55], [1143.25, 173.64], [1044.11, 173.95], [872.08, 173.77]]}, {\"category\": \"car\", \"corners_3d\": [[-185.07, 433.74], [-426.04, 427.2], [88.78, 306.59], [220.82, 308.37], [-185.07, 222.37], [-426.04, 221.12], [88.78, 198.23], [220.82, 198.57]]}, {\"category\": \"car\", \"corners_3d\": [[544.75, 214.2], [508.42, 213.84], [532.3, 209.47], [564.93, 209.76], [544.75, 172.13], [508.42, 172.13], [532.3, 172.21], [564.93, 172.21]]}, {\"category\": \"car\", \"corners_3d\": [[221.44, 256.05], [157.43, 254.75], [271.44, 238.31], [323.88, 239.14], [221.44, 186.63], [157.43, 186.42], [271.44, 183.69], [323.88, 183.83]]}]\n```" - } - ] - }, - { - "id": 3111, - "data_source": "Mono3DRefer", - "scene_name": "004820", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[419.07, 312.97], [306.59, 313.63], [378.07, 277.79], [461.76, 277.42], [419.07, 204.05], [306.59, 204.2], [378.07, 196.22], [461.76, 196.14]]}, {\"category\": \"car\", \"corners_3d\": [[477.53, 258.18], [410.3, 258.11], [446.84, 242.98], [502.16, 243.03], [477.53, 191.66], [410.3, 191.64], [446.84, 188.31], [502.16, 188.32]]}, {\"category\": \"car\", \"corners_3d\": [[527.11, 228.86], [481.64, 228.21], [515.08, 221.05], [554.96, 221.54], [527.11, 181.11], [481.64, 181.01], [515.08, 179.96], [554.96, 180.03]]}]\n```" - } - ] - }, - { - "id": 3112, - "data_source": "Mono3DRefer", - "scene_name": "004829", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[556.86, 219.78], [515.18, 219.51], [534.17, 214.17], [571.16, 214.39], [556.86, 182.33], [515.18, 182.28], [534.17, 181.2], [571.16, 181.24]]}]\n```" - } - ] - }, - { - "id": 3113, - "data_source": "Mono3DRefer", - "scene_name": "004831", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1028.24, 243.44], [1109.81, 248.65], [1060.28, 263.15], [968.38, 255.85], [1028.24, 180.54], [1109.81, 181.11], [1060.28, 182.69], [968.38, 181.89]]}, {\"category\": \"car\", \"corners_3d\": [[1203.03, 219.6], [1150.65, 217.42], [1185.73, 213.91], [1235.37, 215.75], [1203.03, 173.92], [1150.65, 173.87], [1185.73, 173.79], [1235.37, 173.83]]}, {\"category\": \"car\", \"corners_3d\": [[702.26, 300.31], [818.01, 318.28], [608.36, 378.48], [487.51, 344.3], [702.26, 171.21], [818.01, 170.98], [608.36, 170.2], [487.51, 170.64]]}, {\"category\": \"car\", \"corners_3d\": [[874.85, 308.36], [765.2, 293.22], [887.4, 267.48], [983.11, 276.59], [874.85, 174.53], [765.2, 174.34], [887.4, 174.02], [983.11, 174.14]]}]\n```" - } - ] - }, - { - "id": 3114, - "data_source": "Mono3DRefer", - "scene_name": "004834", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[462.08, 251.54], [524.4, 251.34], [508.46, 270.86], [430.54, 271.17], [462.08, 187.75], [524.4, 187.71], [508.46, 191.41], [430.54, 191.47]]}, {\"category\": \"car\", \"corners_3d\": [[528.09, 248.14], [466.7, 248.15], [487.72, 237.03], [540.03, 237.03], [528.09, 193.93], [466.7, 193.93], [487.72, 190.82], [540.03, 190.81]]}, {\"category\": \"car\", \"corners_3d\": [[657.1, 236.89], [706.69, 236.5], [734.3, 247.37], [676.38, 247.91], [657.1, 186.5], [706.69, 186.41], [734.3, 188.73], [676.38, 188.84]]}, {\"category\": \"car\", \"corners_3d\": [[637.68, 228.24], [679.69, 228.11], [695.47, 236.96], [646.75, 237.14], [637.68, 189.2], [679.69, 189.16], [695.47, 191.77], [646.75, 191.82]]}, {\"category\": \"car\", \"corners_3d\": [[556.3, 221.91], [523.65, 221.84], [533.95, 217.13], [563.46, 217.18], [556.3, 188.92], [523.65, 188.9], [533.95, 187.35], [563.46, 187.37]]}, {\"category\": \"car\", \"corners_3d\": [[571.88, 213.82], [541.55, 213.81], [548.76, 209.85], [576.17, 209.86], [571.88, 178.88], [541.55, 178.88], [548.76, 178.3], [576.17, 178.3]]}, {\"category\": \"car\", \"corners_3d\": [[585.51, 212.15], [558.74, 212.08], [566.17, 208.71], [590.64, 208.77], [585.51, 183.03], [558.74, 183.01], [566.17, 182.14], [590.64, 182.16]]}, {\"category\": \"car\", \"corners_3d\": [[454.23, 330.18], [319.21, 329.62], [411.88, 281.86], [505.9, 282.14], [454.23, 191.54], [319.21, 191.47], [411.88, 185.8], [505.9, 185.83]]}, {\"category\": \"car\", \"corners_3d\": [[737.03, 304.57], [841.71, 303.82], [968.55, 367.59], [813.54, 369.27], [737.03, 191.34], [841.71, 191.24], [968.55, 200.19], [813.54, 200.43]]}, {\"category\": \"car\", \"corners_3d\": [[674.46, 264.07], [747.95, 263.71], [796.3, 290.06], [701.64, 290.67], [674.46, 192.75], [747.95, 192.67], [796.3, 198.42], [701.64, 198.55]]}]\n```" - } - ] - }, - { - "id": 3115, - "data_source": "Mono3DRefer", - "scene_name": "004848", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[336.31, 360.21], [160.67, 358.86], [326.62, 293.18], [440.73, 293.74], [336.31, 202.29], [160.67, 202.08], [326.62, 191.76], [440.73, 191.85]]}, {\"category\": \"car\", \"corners_3d\": [[508.1, 209.25], [535.6, 209.17], [532.86, 212.37], [502.91, 212.46], [508.1, 181.51], [535.6, 181.49], [532.86, 182.25], [502.91, 182.27]]}]\n```" - } - ] - }, - { - "id": 3116, - "data_source": "Mono3DRefer", - "scene_name": "004865", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[517.12, 255.48], [446.71, 255.57], [476.32, 239.77], [533.28, 239.71], [517.12, 181.48], [446.71, 181.49], [476.32, 179.84], [533.28, 179.83]]}, {\"category\": \"car\", \"corners_3d\": [[558.72, 217.86], [533.45, 217.83], [540.7, 214.21], [563.93, 214.24], [558.72, 189.73], [533.45, 189.72], [540.7, 188.36], [563.93, 188.37]]}]\n```" - } - ] - }, - { - "id": 3117, - "data_source": "Mono3DRefer", - "scene_name": "004869", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[610.55, 323.02], [611.72, 312.51], [722.83, 313.17], [730.55, 323.78], [610.55, 146.36], [611.72, 148.88], [722.83, 148.72], [730.55, 146.18]]}]\n```" - } - ] - }, - { - "id": 3118, - "data_source": "Mono3DRefer", - "scene_name": "004875", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[166.52, 239.35], [162.11, 233.75], [364.1, 225.54], [383.74, 229.68], [166.52, 167.54], [162.11, 167.98], [364.1, 168.64], [383.74, 168.31]]}, {\"category\": \"car\", \"corners_3d\": [[1075.76, 236.01], [1170.48, 242.73], [1064.86, 253.74], [968.45, 244.86], [1075.76, 156.44], [1170.48, 154.69], [1064.86, 151.83], [968.45, 154.14]]}, {\"category\": \"car\", \"corners_3d\": [[1235.89, 202.64], [1189.26, 200.83], [1280.49, 200.02], [1331.41, 201.72], [1235.89, 153.8], [1189.26, 154.96], [1280.49, 155.48], [1331.41, 154.39]]}]\n```" - } - ] - }, - { - "id": 3119, - "data_source": "Mono3DRefer", - "scene_name": "004875", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[882.38, 209.93], [851.97, 207.3], [971.89, 205.97], [1009.85, 208.4], [882.38, 141.54], [851.97, 143.76], [971.89, 144.88], [1009.85, 142.83]]}]\n```" - } - ] - }, - { - "id": 3120, - "data_source": "Mono3DRefer", - "scene_name": "004875", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[166.52, 239.35], [162.11, 233.75], [364.1, 225.54], [383.74, 229.68], [166.52, 167.54], [162.11, 167.98], [364.1, 168.64], [383.74, 168.31]]}, {\"category\": \"car\", \"corners_3d\": [[1075.76, 236.01], [1170.48, 242.73], [1064.86, 253.74], [968.45, 244.86], [1075.76, 156.44], [1170.48, 154.69], [1064.86, 151.83], [968.45, 154.14]]}, {\"category\": \"van\", \"corners_3d\": [[882.38, 209.93], [851.97, 207.3], [971.89, 205.97], [1009.85, 208.4], [882.38, 141.54], [851.97, 143.76], [971.89, 144.88], [1009.85, 142.83]]}, {\"category\": \"car\", \"corners_3d\": [[1235.89, 202.64], [1189.26, 200.83], [1280.49, 200.02], [1331.41, 201.72], [1235.89, 153.8], [1189.26, 154.96], [1280.49, 155.48], [1331.41, 154.39]]}]\n```" - } - ] - }, - { - "id": 3121, - "data_source": "Mono3DRefer", - "scene_name": "004877", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1914.15, 442.62], [2799.51, 622.71], [1644.37, 646.49], [1208.31, 451.0], [1914.15, 183.24], [2799.51, 190.17], [1644.37, 191.08], [1208.31, 183.56]]}, {\"category\": \"car\", \"corners_3d\": [[881.05, 331.36], [821.3, 301.39], [1157.27, 297.63], [1290.42, 325.68], [881.05, 145.73], [821.3, 150.86], [1157.27, 151.5], [1290.42, 146.7]]}, {\"category\": \"car\", \"corners_3d\": [[-99.03, 317.24], [0.73, 296.01], [267.68, 294.18], [214.87, 314.73], [-99.03, 203.14], [0.73, 198.69], [267.68, 198.31], [214.87, 202.62]]}, {\"category\": \"car\", \"corners_3d\": [[198.91, 255.7], [229.02, 248.13], [376.5, 246.5], [361.58, 253.73], [198.91, 191.75], [229.02, 190.03], [376.5, 189.65], [361.58, 191.3]]}, {\"category\": \"car\", \"corners_3d\": [[363.52, 221.12], [373.21, 218.62], [462.03, 218.17], [457.25, 220.62], [363.52, 184.91], [373.21, 184.29], [462.03, 184.18], [457.25, 184.79]]}]\n```" - } - ] - }, - { - "id": 3122, - "data_source": "Mono3DRefer", - "scene_name": "004877", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[587.72, 216.2], [603.41, 216.16], [605.16, 219.58], [588.23, 219.62], [587.72, 173.59], [603.41, 173.59], [605.16, 173.65], [588.23, 173.65]]}]\n```" - } - ] - }, - { - "id": 3123, - "data_source": "Mono3DRefer", - "scene_name": "004877", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1914.15, 442.62], [2799.51, 622.71], [1644.37, 646.49], [1208.31, 451.0], [1914.15, 183.24], [2799.51, 190.17], [1644.37, 191.08], [1208.31, 183.56]]}, {\"category\": \"car\", \"corners_3d\": [[881.05, 331.36], [821.3, 301.39], [1157.27, 297.63], [1290.42, 325.68], [881.05, 145.73], [821.3, 150.86], [1157.27, 151.5], [1290.42, 146.7]]}, {\"category\": \"cyclist\", \"corners_3d\": [[587.72, 216.2], [603.41, 216.16], [605.16, 219.58], [588.23, 219.62], [587.72, 173.59], [603.41, 173.59], [605.16, 173.65], [588.23, 173.65]]}, {\"category\": \"car\", \"corners_3d\": [[-99.03, 317.24], [0.73, 296.01], [267.68, 294.18], [214.87, 314.73], [-99.03, 203.14], [0.73, 198.69], [267.68, 198.31], [214.87, 202.62]]}, {\"category\": \"car\", \"corners_3d\": [[198.91, 255.7], [229.02, 248.13], [376.5, 246.5], [361.58, 253.73], [198.91, 191.75], [229.02, 190.03], [376.5, 189.65], [361.58, 191.3]]}, {\"category\": \"car\", \"corners_3d\": [[363.52, 221.12], [373.21, 218.62], [462.03, 218.17], [457.25, 220.62], [363.52, 184.91], [373.21, 184.29], [462.03, 184.18], [457.25, 184.79]]}]\n```" - } - ] - }, - { - "id": 3124, - "data_source": "Mono3DRefer", - "scene_name": "004878", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[439.33, 321.55], [304.73, 322.12], [395.99, 275.12], [488.07, 274.85], [439.33, 207.39], [304.73, 207.53], [395.99, 196.61], [488.07, 196.55]]}, {\"category\": \"car\", \"corners_3d\": [[578.79, 195.04], [553.9, 195.06], [556.47, 193.35], [579.45, 193.33], [578.79, 168.86], [553.9, 168.86], [556.47, 169.17], [579.45, 169.17]]}]\n```" - } - ] - }, - { - "id": 3125, - "data_source": "Mono3DRefer", - "scene_name": "004899", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[123.18, 230.62], [79.85, 230.32], [159.17, 222.56], [196.96, 222.78], [123.18, 183.52], [79.85, 183.46], [159.17, 182.03], [196.96, 182.07]]}, {\"category\": \"car\", \"corners_3d\": [[-47.38, 247.65], [-110.53, 247.31], [38.5, 232.63], [89.7, 232.84], [-47.38, 173.67], [-110.53, 173.67], [38.5, 173.51], [89.7, 173.51]]}]\n```" - } - ] - }, - { - "id": 3126, - "data_source": "Mono3DRefer", - "scene_name": "004904", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[819.56, 188.87], [820.96, 188.52], [838.96, 188.65], [837.96, 189.01], [819.56, 156.2], [820.96, 156.57], [838.96, 156.43], [837.96, 156.05]]}]\n```" - } - ] - }, - { - "id": 3127, - "data_source": "Mono3DRefer", - "scene_name": "004904", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[823.02, 239.99], [904.89, 241.24], [973.04, 268.51], [857.62, 266.08], [823.02, 159.9], [904.89, 159.66], [973.04, 154.4], [857.62, 154.87]]}]\n```" - } - ] - }, - { - "id": 3128, - "data_source": "Mono3DRefer", - "scene_name": "004904", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[803.57, 206.86], [857.41, 207.8], [846.5, 215.51], [781.51, 214.12], [803.57, 140.12], [857.41, 139.22], [846.5, 131.79], [781.51, 133.13]]}]\n```" - } - ] - }, - { - "id": 3129, - "data_source": "Mono3DRefer", - "scene_name": "004904", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[823.02, 239.99], [904.89, 241.24], [973.04, 268.51], [857.62, 266.08], [823.02, 159.9], [904.89, 159.66], [973.04, 154.4], [857.62, 154.87]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[819.56, 188.87], [820.96, 188.52], [838.96, 188.65], [837.96, 189.01], [819.56, 156.2], [820.96, 156.57], [838.96, 156.43], [837.96, 156.05]]}]\n```" - } - ] - }, - { - "id": 3130, - "data_source": "Mono3DRefer", - "scene_name": "004904", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[803.57, 206.86], [857.41, 207.8], [846.5, 215.51], [781.51, 214.12], [803.57, 140.12], [857.41, 139.22], [846.5, 131.79], [781.51, 133.13]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[819.56, 188.87], [820.96, 188.52], [838.96, 188.65], [837.96, 189.01], [819.56, 156.2], [820.96, 156.57], [838.96, 156.43], [837.96, 156.05]]}]\n```" - } - ] - }, - { - "id": 3131, - "data_source": "Mono3DRefer", - "scene_name": "004904", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[823.02, 239.99], [904.89, 241.24], [973.04, 268.51], [857.62, 266.08], [823.02, 159.9], [904.89, 159.66], [973.04, 154.4], [857.62, 154.87]]}, {\"category\": \"van\", \"corners_3d\": [[803.57, 206.86], [857.41, 207.8], [846.5, 215.51], [781.51, 214.12], [803.57, 140.12], [857.41, 139.22], [846.5, 131.79], [781.51, 133.13]]}]\n```" - } - ] - }, - { - "id": 3132, - "data_source": "Mono3DRefer", - "scene_name": "004904", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[823.02, 239.99], [904.89, 241.24], [973.04, 268.51], [857.62, 266.08], [823.02, 159.9], [904.89, 159.66], [973.04, 154.4], [857.62, 154.87]]}, {\"category\": \"van\", \"corners_3d\": [[803.57, 206.86], [857.41, 207.8], [846.5, 215.51], [781.51, 214.12], [803.57, 140.12], [857.41, 139.22], [846.5, 131.79], [781.51, 133.13]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[819.56, 188.87], [820.96, 188.52], [838.96, 188.65], [837.96, 189.01], [819.56, 156.2], [820.96, 156.57], [838.96, 156.43], [837.96, 156.05]]}]\n```" - } - ] - }, - { - "id": 3133, - "data_source": "Mono3DRefer", - "scene_name": "004908", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[381.93, 261.08], [295.33, 260.76], [367.11, 241.94], [435.32, 242.14], [381.93, 187.87], [295.33, 187.81], [367.11, 184.61], [435.32, 184.64]]}]\n```" - } - ] - }, - { - "id": 3134, - "data_source": "Mono3DRefer", - "scene_name": "004908", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[742.83, 202.86], [753.46, 202.83], [763.62, 204.4], [752.45, 204.43], [742.83, 165.72], [753.46, 165.73], [763.62, 165.36], [752.45, 165.35]]}]\n```" - } - ] - }, - { - "id": 3135, - "data_source": "Mono3DRefer", - "scene_name": "004908", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[381.93, 261.08], [295.33, 260.76], [367.11, 241.94], [435.32, 242.14], [381.93, 187.87], [295.33, 187.81], [367.11, 184.61], [435.32, 184.64]]}, {\"category\": \"cyclist\", \"corners_3d\": [[742.83, 202.86], [753.46, 202.83], [763.62, 204.4], [752.45, 204.43], [742.83, 165.72], [753.46, 165.73], [763.62, 165.36], [752.45, 165.35]]}]\n```" - } - ] - }, - { - "id": 3136, - "data_source": "Mono3DRefer", - "scene_name": "004912", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[231.59, 306.38], [253.75, 299.19], [343.54, 299.52], [326.77, 306.75], [231.59, 158.07], [253.75, 159.35], [343.54, 159.29], [326.77, 158.0]]}]\n```" - } - ] - }, - { - "id": 3137, - "data_source": "Mono3DRefer", - "scene_name": "004920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[661.87, 217.39], [717.75, 217.36], [754.34, 231.54], [680.66, 231.59], [661.87, 150.86], [717.75, 150.87], [754.34, 143.87], [680.66, 143.84]]}, {\"category\": \"van\", \"corners_3d\": [[637.46, 199.89], [670.82, 199.86], [684.53, 204.79], [645.09, 204.83], [637.46, 156.73], [670.82, 156.74], [684.53, 153.8], [645.09, 153.78]]}, {\"category\": \"van\", \"corners_3d\": [[562.69, 194.73], [561.0, 195.42], [491.37, 195.4], [495.17, 194.71], [562.69, 161.44], [561.0, 161.08], [491.37, 161.1], [495.17, 161.45]]}]\n```" - } - ] - }, - { - "id": 3138, - "data_source": "Mono3DRefer", - "scene_name": "004920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[882.44, 279.09], [844.65, 265.17], [1084.41, 264.5], [1157.73, 278.2], [882.44, 167.1], [844.65, 167.86], [1084.41, 167.89], [1157.73, 167.15]]}, {\"category\": \"car\", \"corners_3d\": [[250.8, 302.65], [172.66, 330.92], [-209.03, 330.92], [-62.62, 302.65], [250.8, 132.11], [172.66, 123.24], [-209.03, 123.24], [-62.62, 132.11]]}, {\"category\": \"car\", \"corners_3d\": [[16.54, 285.97], [89.72, 271.81], [290.54, 271.49], [246.25, 285.55], [16.54, 186.97], [89.72, 185.2], [290.54, 185.16], [246.25, 186.92]]}, {\"category\": \"car\", \"corners_3d\": [[246.84, 240.77], [277.6, 234.99], [441.99, 234.97], [426.53, 240.74], [246.84, 179.92], [277.6, 179.32], [441.99, 179.32], [426.53, 179.92]]}, {\"category\": \"car\", \"corners_3d\": [[341.88, 227.87], [355.14, 223.81], [490.62, 222.69], [488.22, 226.57], [341.88, 168.91], [355.14, 169.2], [490.62, 169.28], [488.22, 169.0]]}, {\"category\": \"car\", \"corners_3d\": [[439.78, 205.05], [447.07, 203.79], [507.64, 203.84], [502.83, 205.11], [439.78, 174.76], [447.07, 174.69], [507.64, 174.69], [502.83, 174.77]]}, {\"category\": \"car\", \"corners_3d\": [[533.25, 203.82], [530.01, 205.13], [457.01, 205.13], [463.23, 203.82], [533.25, 174.91], [530.01, 175.0], [457.01, 175.0], [463.23, 174.91]]}]\n```" - } - ] - }, - { - "id": 3139, - "data_source": "Mono3DRefer", - "scene_name": "004920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[882.44, 279.09], [844.65, 265.17], [1084.41, 264.5], [1157.73, 278.2], [882.44, 167.1], [844.65, 167.86], [1084.41, 167.89], [1157.73, 167.15]]}, {\"category\": \"car\", \"corners_3d\": [[250.8, 302.65], [172.66, 330.92], [-209.03, 330.92], [-62.62, 302.65], [250.8, 132.11], [172.66, 123.24], [-209.03, 123.24], [-62.62, 132.11]]}, {\"category\": \"car\", \"corners_3d\": [[16.54, 285.97], [89.72, 271.81], [290.54, 271.49], [246.25, 285.55], [16.54, 186.97], [89.72, 185.2], [290.54, 185.16], [246.25, 186.92]]}, {\"category\": \"van\", \"corners_3d\": [[661.87, 217.39], [717.75, 217.36], [754.34, 231.54], [680.66, 231.59], [661.87, 150.86], [717.75, 150.87], [754.34, 143.87], [680.66, 143.84]]}, {\"category\": \"car\", \"corners_3d\": [[246.84, 240.77], [277.6, 234.99], [441.99, 234.97], [426.53, 240.74], [246.84, 179.92], [277.6, 179.32], [441.99, 179.32], [426.53, 179.92]]}, {\"category\": \"car\", \"corners_3d\": [[341.88, 227.87], [355.14, 223.81], [490.62, 222.69], [488.22, 226.57], [341.88, 168.91], [355.14, 169.2], [490.62, 169.28], [488.22, 169.0]]}, {\"category\": \"van\", \"corners_3d\": [[637.46, 199.89], [670.82, 199.86], [684.53, 204.79], [645.09, 204.83], [637.46, 156.73], [670.82, 156.74], [684.53, 153.8], [645.09, 153.78]]}, {\"category\": \"car\", \"corners_3d\": [[439.78, 205.05], [447.07, 203.79], [507.64, 203.84], [502.83, 205.11], [439.78, 174.76], [447.07, 174.69], [507.64, 174.69], [502.83, 174.77]]}, {\"category\": \"car\", \"corners_3d\": [[533.25, 203.82], [530.01, 205.13], [457.01, 205.13], [463.23, 203.82], [533.25, 174.91], [530.01, 175.0], [457.01, 175.0], [463.23, 174.91]]}, {\"category\": \"van\", \"corners_3d\": [[562.69, 194.73], [561.0, 195.42], [491.37, 195.4], [495.17, 194.71], [562.69, 161.44], [561.0, 161.08], [491.37, 161.1], [495.17, 161.45]]}]\n```" - } - ] - }, - { - "id": 3140, - "data_source": "Mono3DRefer", - "scene_name": "004936", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[541.5, 220.59], [498.27, 220.59], [512.22, 214.58], [550.0, 214.57], [541.5, 184.4], [498.27, 184.4], [512.22, 182.95], [550.0, 182.95]]}]\n```" - } - ] - }, - { - "id": 3141, - "data_source": "Mono3DRefer", - "scene_name": "004940", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[122.63, 360.22], [164.62, 342.98], [289.2, 341.28], [260.76, 358.14], [122.63, 149.56], [164.62, 152.53], [289.2, 152.82], [260.76, 149.92]]}]\n```" - } - ] - }, - { - "id": 3142, - "data_source": "Mono3DRefer", - "scene_name": "004949", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[206.86, 262.78], [220.16, 264.7], [176.36, 268.8], [163.48, 266.69], [206.86, 169.01], [220.16, 168.74], [176.36, 168.17], [163.48, 168.46]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[270.95, 253.2], [280.92, 255.04], [233.17, 258.56], [223.98, 256.55], [270.95, 170.35], [280.92, 170.09], [233.17, 169.6], [223.98, 169.88]]}]\n```" - } - ] - }, - { - "id": 3143, - "data_source": "Mono3DRefer", - "scene_name": "004949", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[735.44, 296.18], [771.93, 301.89], [684.09, 318.18], [647.62, 310.88], [735.44, 176.06], [771.93, 175.85], [684.09, 175.22], [647.62, 175.5]]}]\n```" - } - ] - }, - { - "id": 3144, - "data_source": "Mono3DRefer", - "scene_name": "004949", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[735.44, 296.18], [771.93, 301.89], [684.09, 318.18], [647.62, 310.88], [735.44, 176.06], [771.93, 175.85], [684.09, 175.22], [647.62, 175.5]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[206.86, 262.78], [220.16, 264.7], [176.36, 268.8], [163.48, 266.69], [206.86, 169.01], [220.16, 168.74], [176.36, 168.17], [163.48, 168.46]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[270.95, 253.2], [280.92, 255.04], [233.17, 258.56], [223.98, 256.55], [270.95, 170.35], [280.92, 170.09], [233.17, 169.6], [223.98, 169.88]]}]\n```" - } - ] - }, - { - "id": 3145, - "data_source": "Mono3DRefer", - "scene_name": "004951", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[464.35, 249.74], [394.31, 249.9], [431.88, 235.44], [488.74, 235.34], [464.35, 180.93], [394.31, 180.95], [431.88, 179.43], [488.74, 179.42]]}]\n```" - } - ] - }, - { - "id": 3146, - "data_source": "Mono3DRefer", - "scene_name": "004961", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[16.03, 698.66], [-525.57, 717.92], [184.0, 365.82], [371.13, 363.35], [16.03, 154.92], [-525.57, 154.26], [184.0, 166.27], [371.13, 166.36]]}, {\"category\": \"car\", \"corners_3d\": [[550.89, 242.88], [488.3, 241.76], [533.34, 229.44], [585.2, 230.2], [550.89, 182.1], [488.3, 181.95], [533.34, 180.32], [585.2, 180.42]]}, {\"category\": \"car\", \"corners_3d\": [[596.98, 224.25], [556.86, 223.4], [589.64, 217.2], [625.26, 217.86], [596.98, 177.6], [556.86, 177.52], [589.64, 176.95], [625.26, 177.01]]}, {\"category\": \"car\", \"corners_3d\": [[652.09, 212.54], [617.27, 211.91], [640.09, 208.05], [671.74, 208.56], [652.09, 176.96], [617.27, 176.9], [640.09, 176.5], [671.74, 176.55]]}, {\"category\": \"car\", \"corners_3d\": [[820.0, 185.64], [841.58, 185.85], [824.24, 186.78], [801.44, 186.54], [820.0, 164.08], [841.58, 163.94], [824.24, 163.3], [801.44, 163.46]]}]\n```" - } - ] - }, - { - "id": 3147, - "data_source": "Mono3DRefer", - "scene_name": "004969", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[780.73, 216.04], [825.71, 216.61], [833.98, 223.29], [782.12, 222.53], [780.73, 175.2], [825.71, 175.23], [833.98, 175.59], [782.12, 175.55]]}]\n```" - } - ] - }, - { - "id": 3148, - "data_source": "Mono3DRefer", - "scene_name": "004982", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[583.45, 217.77], [626.78, 217.75], [630.5, 224.76], [580.41, 224.8], [583.45, 176.83], [626.78, 176.83], [630.5, 177.45], [580.41, 177.45]]}, {\"category\": \"car\", \"corners_3d\": [[599.98, 207.26], [634.19, 207.26], [637.51, 212.01], [598.57, 212.01], [599.98, 171.43], [634.19, 171.43], [637.51, 171.23], [598.57, 171.23]]}, {\"category\": \"car\", \"corners_3d\": [[871.56, 365.09], [1068.73, 366.69], [2036.5, 796.0], [1388.78, 779.76], [871.56, 157.29], [1068.73, 157.16], [2036.5, 122.39], [1388.78, 123.7]]}, {\"category\": \"car\", \"corners_3d\": [[736.4, 256.73], [813.49, 256.56], [882.02, 282.9], [780.79, 283.2], [736.4, 175.49], [813.49, 175.48], [882.02, 176.31], [780.79, 176.32]]}, {\"category\": \"car\", \"corners_3d\": [[668.18, 211.51], [705.19, 211.51], [717.35, 216.46], [675.6, 216.46], [668.18, 173.56], [705.19, 173.56], [717.35, 173.65], [675.6, 173.65]]}, {\"category\": \"car\", \"corners_3d\": [[651.69, 205.59], [681.73, 205.54], [694.28, 209.62], [660.5, 209.69], [651.69, 175.61], [681.73, 175.6], [694.28, 175.95], [660.5, 175.95]]}]\n```" - } - ] - }, - { - "id": 3149, - "data_source": "Mono3DRefer", - "scene_name": "004994", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[444.1, 274.82], [358.66, 275.3], [409.51, 251.78], [475.22, 251.49], [444.1, 195.62], [358.66, 195.73], [409.51, 190.48], [475.22, 190.41]]}, {\"category\": \"car\", \"corners_3d\": [[556.97, 206.58], [522.14, 206.63], [528.94, 203.1], [560.13, 203.06], [556.97, 177.61], [522.14, 177.62], [528.94, 177.12], [560.13, 177.11]]}]\n```" - } - ] - }, - { - "id": 3150, - "data_source": "Mono3DRefer", - "scene_name": "004994", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[496.21, 240.81], [431.57, 240.88], [462.05, 228.69], [515.1, 228.65], [496.21, 170.51], [431.57, 170.51], [462.05, 170.93], [515.1, 170.93]]}]\n```" - } - ] - }, - { - "id": 3151, - "data_source": "Mono3DRefer", - "scene_name": "004994", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[444.1, 274.82], [358.66, 275.3], [409.51, 251.78], [475.22, 251.49], [444.1, 195.62], [358.66, 195.73], [409.51, 190.48], [475.22, 190.41]]}, {\"category\": \"van\", \"corners_3d\": [[496.21, 240.81], [431.57, 240.88], [462.05, 228.69], [515.1, 228.65], [496.21, 170.51], [431.57, 170.51], [462.05, 170.93], [515.1, 170.93]]}, {\"category\": \"car\", \"corners_3d\": [[556.97, 206.58], [522.14, 206.63], [528.94, 203.1], [560.13, 203.06], [556.97, 177.61], [522.14, 177.62], [528.94, 177.12], [560.13, 177.11]]}]\n```" - } - ] - }, - { - "id": 3152, - "data_source": "Mono3DRefer", - "scene_name": "005001", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[375.11, 318.53], [251.61, 319.54], [350.25, 275.49], [436.37, 275.0], [375.11, 204.75], [251.61, 204.97], [350.25, 195.33], [436.37, 195.22]]}, {\"category\": \"car\", \"corners_3d\": [[552.11, 213.13], [512.67, 213.16], [523.06, 208.46], [557.9, 208.44], [552.11, 180.27], [512.67, 180.28], [523.06, 179.41], [557.9, 179.41]]}]\n```" - } - ] - }, - { - "id": 3153, - "data_source": "Mono3DRefer", - "scene_name": "005001", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[467.18, 259.6], [385.07, 259.7], [432.08, 240.86], [496.34, 240.8], [467.18, 170.37], [385.07, 170.37], [432.08, 170.91], [496.34, 170.91]]}]\n```" - } - ] - }, - { - "id": 3154, - "data_source": "Mono3DRefer", - "scene_name": "005001", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[375.11, 318.53], [251.61, 319.54], [350.25, 275.49], [436.37, 275.0], [375.11, 204.75], [251.61, 204.97], [350.25, 195.33], [436.37, 195.22]]}, {\"category\": \"van\", \"corners_3d\": [[467.18, 259.6], [385.07, 259.7], [432.08, 240.86], [496.34, 240.8], [467.18, 170.37], [385.07, 170.37], [432.08, 170.91], [496.34, 170.91]]}, {\"category\": \"car\", \"corners_3d\": [[552.11, 213.13], [512.67, 213.16], [523.06, 208.46], [557.9, 208.44], [552.11, 180.27], [512.67, 180.28], [523.06, 179.41], [557.9, 179.41]]}]\n```" - } - ] - }, - { - "id": 3155, - "data_source": "Mono3DRefer", - "scene_name": "005014", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[310.32, 308.09], [186.83, 308.56], [301.11, 270.53], [389.78, 270.28], [310.32, 189.0], [186.83, 189.06], [301.11, 184.52], [389.78, 184.49]]}, {\"category\": \"car\", \"corners_3d\": [[482.09, 224.19], [434.7, 224.2], [462.95, 215.87], [502.66, 215.87], [482.09, 182.77], [434.7, 182.77], [462.95, 181.17], [502.66, 181.17]]}, {\"category\": \"car\", \"corners_3d\": [[640.79, 197.31], [667.03, 197.36], [668.22, 199.51], [639.67, 199.45], [640.79, 175.83], [667.03, 175.84], [668.22, 176.1], [639.67, 176.1]]}, {\"category\": \"car\", \"corners_3d\": [[741.47, 289.8], [860.92, 289.62], [1005.7, 355.14], [819.41, 355.58], [741.47, 174.26], [860.92, 174.26], [1005.7, 175.05], [819.41, 175.06]]}]\n```" - } - ] - }, - { - "id": 3156, - "data_source": "Mono3DRefer", - "scene_name": "005014", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[508.29, 232.1], [488.03, 232.08], [498.87, 227.08], [517.41, 227.09], [508.29, 172.52], [488.03, 172.52], [498.87, 172.55], [517.41, 172.55]]}, {\"category\": \"cyclist\", \"corners_3d\": [[536.4, 229.24], [515.59, 229.34], [519.28, 225.18], [538.55, 225.1], [536.4, 174.1], [515.59, 174.1], [519.28, 174.01], [538.55, 174.0]]}, {\"category\": \"cyclist\", \"corners_3d\": [[547.93, 209.39], [535.08, 209.4], [537.97, 207.62], [550.2, 207.61], [547.93, 174.79], [535.08, 174.79], [537.97, 174.7], [550.2, 174.7]]}]\n```" - } - ] - }, - { - "id": 3157, - "data_source": "Mono3DRefer", - "scene_name": "005014", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[310.32, 308.09], [186.83, 308.56], [301.11, 270.53], [389.78, 270.28], [310.32, 189.0], [186.83, 189.06], [301.11, 184.52], [389.78, 184.49]]}, {\"category\": \"cyclist\", \"corners_3d\": [[508.29, 232.1], [488.03, 232.08], [498.87, 227.08], [517.41, 227.09], [508.29, 172.52], [488.03, 172.52], [498.87, 172.55], [517.41, 172.55]]}, {\"category\": \"cyclist\", \"corners_3d\": [[536.4, 229.24], [515.59, 229.34], [519.28, 225.18], [538.55, 225.1], [536.4, 174.1], [515.59, 174.1], [519.28, 174.01], [538.55, 174.0]]}, {\"category\": \"car\", \"corners_3d\": [[482.09, 224.19], [434.7, 224.2], [462.95, 215.87], [502.66, 215.87], [482.09, 182.77], [434.7, 182.77], [462.95, 181.17], [502.66, 181.17]]}, {\"category\": \"cyclist\", \"corners_3d\": [[547.93, 209.39], [535.08, 209.4], [537.97, 207.62], [550.2, 207.61], [547.93, 174.79], [535.08, 174.79], [537.97, 174.7], [550.2, 174.7]]}, {\"category\": \"car\", \"corners_3d\": [[640.79, 197.31], [667.03, 197.36], [668.22, 199.51], [639.67, 199.45], [640.79, 175.83], [667.03, 175.84], [668.22, 176.1], [639.67, 176.1]]}, {\"category\": \"car\", \"corners_3d\": [[741.47, 289.8], [860.92, 289.62], [1005.7, 355.14], [819.41, 355.58], [741.47, 174.26], [860.92, 174.26], [1005.7, 175.05], [819.41, 175.06]]}]\n```" - } - ] - }, - { - "id": 3158, - "data_source": "Mono3DRefer", - "scene_name": "005020", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1146.71, 218.74], [1193.15, 220.25], [1191.71, 225.54], [1140.31, 223.68], [1146.71, 185.43], [1193.15, 185.84], [1191.71, 187.29], [1140.31, 186.78]]}]\n```" - } - ] - }, - { - "id": 3159, - "data_source": "Mono3DRefer", - "scene_name": "005033", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-1089.3, 1643.84], [-2877.4, 1809.36], [50.51, 420.62], [297.6, 416.47], [-1089.3, 87.5], [-2877.4, 77.9], [50.51, 158.48], [297.6, 158.72]]}, {\"category\": \"car\", \"corners_3d\": [[541.75, 252.82], [472.68, 251.49], [523.12, 236.24], [579.31, 237.1], [541.75, 185.72], [472.68, 185.51], [523.12, 183.06], [579.31, 183.19]]}, {\"category\": \"car\", \"corners_3d\": [[593.16, 231.31], [550.14, 230.27], [585.9, 222.78], [623.79, 223.57], [593.16, 181.15], [550.14, 181.0], [585.9, 179.94], [623.79, 180.05]]}, {\"category\": \"car\", \"corners_3d\": [[652.54, 219.5], [615.57, 218.75], [639.0, 213.96], [672.4, 214.56], [652.54, 181.82], [615.57, 181.67], [639.0, 180.75], [672.4, 180.87]]}, {\"category\": \"car\", \"corners_3d\": [[822.92, 187.37], [845.16, 187.62], [827.55, 188.71], [804.01, 188.43], [822.92, 165.21], [845.16, 165.08], [827.55, 164.51], [804.01, 164.66]]}]\n```" - } - ] - }, - { - "id": 3160, - "data_source": "Mono3DRefer", - "scene_name": "005034", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[741.89, 274.03], [831.61, 274.04], [928.19, 318.17], [799.34, 318.15], [741.89, 187.21], [831.61, 187.21], [928.19, 193.47], [799.34, 193.47]]}, {\"category\": \"car\", \"corners_3d\": [[695.72, 243.89], [763.1, 243.7], [810.53, 262.94], [724.96, 263.26], [695.72, 180.51], [763.1, 180.49], [810.53, 182.56], [724.96, 182.6]]}, {\"category\": \"car\", \"corners_3d\": [[545.22, 223.8], [503.58, 224.01], [510.61, 217.44], [546.9, 217.29], [545.22, 181.93], [503.58, 181.97], [510.61, 180.8], [546.9, 180.77]]}, {\"category\": \"car\", \"corners_3d\": [[651.65, 208.73], [683.45, 208.7], [693.43, 212.77], [658.01, 212.81], [651.65, 180.06], [683.45, 180.06], [693.43, 180.88], [658.01, 180.88]]}]\n```" - } - ] - }, - { - "id": 3161, - "data_source": "Mono3DRefer", - "scene_name": "005034", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[663.58, 224.52], [718.8, 224.48], [750.99, 238.82], [680.44, 238.88], [663.58, 158.04], [718.8, 158.05], [750.99, 153.94], [680.44, 153.92]]}]\n```" - } - ] - }, - { - "id": 3162, - "data_source": "Mono3DRefer", - "scene_name": "005034", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[741.89, 274.03], [831.61, 274.04], [928.19, 318.17], [799.34, 318.15], [741.89, 187.21], [831.61, 187.21], [928.19, 193.47], [799.34, 193.47]]}, {\"category\": \"car\", \"corners_3d\": [[695.72, 243.89], [763.1, 243.7], [810.53, 262.94], [724.96, 263.26], [695.72, 180.51], [763.1, 180.49], [810.53, 182.56], [724.96, 182.6]]}, {\"category\": \"van\", \"corners_3d\": [[663.58, 224.52], [718.8, 224.48], [750.99, 238.82], [680.44, 238.88], [663.58, 158.04], [718.8, 158.05], [750.99, 153.94], [680.44, 153.92]]}, {\"category\": \"car\", \"corners_3d\": [[545.22, 223.8], [503.58, 224.01], [510.61, 217.44], [546.9, 217.29], [545.22, 181.93], [503.58, 181.97], [510.61, 180.8], [546.9, 180.77]]}, {\"category\": \"car\", \"corners_3d\": [[651.65, 208.73], [683.45, 208.7], [693.43, 212.77], [658.01, 212.81], [651.65, 180.06], [683.45, 180.06], [693.43, 180.88], [658.01, 180.88]]}]\n```" - } - ] - }, - { - "id": 3163, - "data_source": "Mono3DRefer", - "scene_name": "005045", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1197.87, 357.14], [990.78, 361.12], [802.63, 279.2], [922.95, 277.92], [1197.87, 132.68], [990.78, 131.81], [802.63, 149.67], [922.95, 149.95]]}, {\"category\": \"car\", \"corners_3d\": [[283.92, 367.19], [125.92, 368.48], [297.13, 295.9], [396.03, 295.38], [283.92, 216.47], [125.92, 216.76], [297.13, 200.47], [396.03, 200.36]]}, {\"category\": \"car\", \"corners_3d\": [[463.73, 261.42], [391.1, 261.78], [426.7, 245.03], [485.58, 244.79], [463.73, 202.21], [391.1, 202.33], [426.7, 196.78], [485.58, 196.7]]}, {\"category\": \"car\", \"corners_3d\": [[499.63, 237.42], [444.06, 237.57], [466.08, 227.62], [513.09, 227.51], [499.63, 182.88], [444.06, 182.9], [466.08, 181.36], [513.09, 181.34]]}, {\"category\": \"car\", \"corners_3d\": [[600.66, 205.81], [627.71, 205.75], [632.22, 208.51], [602.91, 208.59], [600.66, 179.94], [627.71, 179.92], [632.22, 180.52], [602.91, 180.53]]}]\n```" - } - ] - }, - { - "id": 3164, - "data_source": "Mono3DRefer", - "scene_name": "005045", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[566.96, 206.46], [536.31, 206.42], [547.86, 202.31], [574.76, 202.34], [566.96, 159.98], [536.31, 159.99], [547.86, 161.57], [574.76, 161.55]]}]\n```" - } - ] - }, - { - "id": 3165, - "data_source": "Mono3DRefer", - "scene_name": "005045", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[593.12, 209.51], [600.33, 209.22], [605.09, 209.52], [597.86, 209.82], [593.12, 178.71], [600.33, 178.67], [605.09, 178.72], [597.86, 178.76]]}]\n```" - } - ] - }, - { - "id": 3166, - "data_source": "Mono3DRefer", - "scene_name": "005045", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1197.87, 357.14], [990.78, 361.12], [802.63, 279.2], [922.95, 277.92], [1197.87, 132.68], [990.78, 131.81], [802.63, 149.67], [922.95, 149.95]]}, {\"category\": \"car\", \"corners_3d\": [[283.92, 367.19], [125.92, 368.48], [297.13, 295.9], [396.03, 295.38], [283.92, 216.47], [125.92, 216.76], [297.13, 200.47], [396.03, 200.36]]}, {\"category\": \"car\", \"corners_3d\": [[463.73, 261.42], [391.1, 261.78], [426.7, 245.03], [485.58, 244.79], [463.73, 202.21], [391.1, 202.33], [426.7, 196.78], [485.58, 196.7]]}, {\"category\": \"car\", \"corners_3d\": [[499.63, 237.42], [444.06, 237.57], [466.08, 227.62], [513.09, 227.51], [499.63, 182.88], [444.06, 182.9], [466.08, 181.36], [513.09, 181.34]]}, {\"category\": \"truck\", \"corners_3d\": [[566.96, 206.46], [536.31, 206.42], [547.86, 202.31], [574.76, 202.34], [566.96, 159.98], [536.31, 159.99], [547.86, 161.57], [574.76, 161.55]]}, {\"category\": \"car\", \"corners_3d\": [[600.66, 205.81], [627.71, 205.75], [632.22, 208.51], [602.91, 208.59], [600.66, 179.94], [627.71, 179.92], [632.22, 180.52], [602.91, 180.53]]}]\n```" - } - ] - }, - { - "id": 3167, - "data_source": "Mono3DRefer", - "scene_name": "005045", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1197.87, 357.14], [990.78, 361.12], [802.63, 279.2], [922.95, 277.92], [1197.87, 132.68], [990.78, 131.81], [802.63, 149.67], [922.95, 149.95]]}, {\"category\": \"car\", \"corners_3d\": [[283.92, 367.19], [125.92, 368.48], [297.13, 295.9], [396.03, 295.38], [283.92, 216.47], [125.92, 216.76], [297.13, 200.47], [396.03, 200.36]]}, {\"category\": \"car\", \"corners_3d\": [[463.73, 261.42], [391.1, 261.78], [426.7, 245.03], [485.58, 244.79], [463.73, 202.21], [391.1, 202.33], [426.7, 196.78], [485.58, 196.7]]}, {\"category\": \"car\", \"corners_3d\": [[499.63, 237.42], [444.06, 237.57], [466.08, 227.62], [513.09, 227.51], [499.63, 182.88], [444.06, 182.9], [466.08, 181.36], [513.09, 181.34]]}, {\"category\": \"car\", \"corners_3d\": [[600.66, 205.81], [627.71, 205.75], [632.22, 208.51], [602.91, 208.59], [600.66, 179.94], [627.71, 179.92], [632.22, 180.52], [602.91, 180.53]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[593.12, 209.51], [600.33, 209.22], [605.09, 209.52], [597.86, 209.82], [593.12, 178.71], [600.33, 178.67], [605.09, 178.72], [597.86, 178.76]]}]\n```" - } - ] - }, - { - "id": 3168, - "data_source": "Mono3DRefer", - "scene_name": "005045", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[566.96, 206.46], [536.31, 206.42], [547.86, 202.31], [574.76, 202.34], [566.96, 159.98], [536.31, 159.99], [547.86, 161.57], [574.76, 161.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[593.12, 209.51], [600.33, 209.22], [605.09, 209.52], [597.86, 209.82], [593.12, 178.71], [600.33, 178.67], [605.09, 178.72], [597.86, 178.76]]}]\n```" - } - ] - }, - { - "id": 3169, - "data_source": "Mono3DRefer", - "scene_name": "005045", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1197.87, 357.14], [990.78, 361.12], [802.63, 279.2], [922.95, 277.92], [1197.87, 132.68], [990.78, 131.81], [802.63, 149.67], [922.95, 149.95]]}, {\"category\": \"car\", \"corners_3d\": [[283.92, 367.19], [125.92, 368.48], [297.13, 295.9], [396.03, 295.38], [283.92, 216.47], [125.92, 216.76], [297.13, 200.47], [396.03, 200.36]]}, {\"category\": \"car\", \"corners_3d\": [[463.73, 261.42], [391.1, 261.78], [426.7, 245.03], [485.58, 244.79], [463.73, 202.21], [391.1, 202.33], [426.7, 196.78], [485.58, 196.7]]}, {\"category\": \"car\", \"corners_3d\": [[499.63, 237.42], [444.06, 237.57], [466.08, 227.62], [513.09, 227.51], [499.63, 182.88], [444.06, 182.9], [466.08, 181.36], [513.09, 181.34]]}, {\"category\": \"truck\", \"corners_3d\": [[566.96, 206.46], [536.31, 206.42], [547.86, 202.31], [574.76, 202.34], [566.96, 159.98], [536.31, 159.99], [547.86, 161.57], [574.76, 161.55]]}, {\"category\": \"car\", \"corners_3d\": [[600.66, 205.81], [627.71, 205.75], [632.22, 208.51], [602.91, 208.59], [600.66, 179.94], [627.71, 179.92], [632.22, 180.52], [602.91, 180.53]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[593.12, 209.51], [600.33, 209.22], [605.09, 209.52], [597.86, 209.82], [593.12, 178.71], [600.33, 178.67], [605.09, 178.72], [597.86, 178.76]]}]\n```" - } - ] - }, - { - "id": 3170, - "data_source": "Mono3DRefer", - "scene_name": "005047", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[571.52, 199.09], [547.32, 199.09], [551.71, 197.23], [574.18, 197.23], [571.52, 179.27], [547.32, 179.27], [551.71, 178.82], [574.18, 178.82]]}]\n```" - } - ] - }, - { - "id": 3171, - "data_source": "Mono3DRefer", - "scene_name": "005051", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[312.58, 366.74], [114.6, 374.35], [249.77, 303.67], [376.67, 300.42], [312.58, 199.02], [114.6, 200.05], [249.77, 190.51], [376.67, 190.07]]}, {\"category\": \"car\", \"corners_3d\": [[598.19, 204.51], [593.96, 203.16], [676.3, 202.63], [684.06, 203.94], [598.19, 172.1], [593.96, 172.13], [676.3, 172.14], [684.06, 172.11]]}, {\"category\": \"car\", \"corners_3d\": [[469.5, 206.75], [441.54, 206.87], [450.02, 204.05], [475.65, 203.95], [469.5, 182.07], [441.54, 182.1], [450.02, 181.33], [475.65, 181.3]]}, {\"category\": \"car\", \"corners_3d\": [[725.97, 181.77], [742.35, 181.74], [754.24, 182.21], [737.04, 182.24], [725.97, 163.31], [742.35, 163.34], [754.24, 162.84], [737.04, 162.81]]}]\n```" - } - ] - }, - { - "id": 3172, - "data_source": "Mono3DRefer", - "scene_name": "005052", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[686.76, 243.22], [746.43, 242.99], [794.89, 263.57], [717.85, 263.95], [686.76, 180.31], [746.43, 180.29], [794.89, 182.47], [717.85, 182.51]]}, {\"category\": \"car\", \"corners_3d\": [[662.85, 226.79], [709.7, 226.58], [738.0, 237.22], [681.97, 237.52], [662.85, 179.77], [709.7, 179.75], [738.0, 181.11], [681.97, 181.15]]}, {\"category\": \"car\", \"corners_3d\": [[693.3, 194.24], [671.16, 194.28], [663.51, 192.72], [684.06, 192.69], [693.3, 171.03], [671.16, 171.03], [663.51, 171.16], [684.06, 171.17]]}, {\"category\": \"car\", \"corners_3d\": [[411.61, 365.87], [224.26, 366.9], [374.04, 288.41], [485.37, 288.04], [411.61, 194.88], [224.26, 195.0], [374.04, 186.04], [485.37, 186.0]]}]\n```" - } - ] - }, - { - "id": 3173, - "data_source": "Mono3DRefer", - "scene_name": "005052", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[753.81, 288.1], [868.21, 287.19], [1080.73, 369.79], [885.48, 372.5], [753.81, 155.61], [868.21, 155.74], [1080.73, 143.38], [885.48, 142.97]]}]\n```" - } - ] - }, - { - "id": 3174, - "data_source": "Mono3DRefer", - "scene_name": "005052", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[686.76, 243.22], [746.43, 242.99], [794.89, 263.57], [717.85, 263.95], [686.76, 180.31], [746.43, 180.29], [794.89, 182.47], [717.85, 182.51]]}, {\"category\": \"car\", \"corners_3d\": [[662.85, 226.79], [709.7, 226.58], [738.0, 237.22], [681.97, 237.52], [662.85, 179.77], [709.7, 179.75], [738.0, 181.11], [681.97, 181.15]]}, {\"category\": \"car\", \"corners_3d\": [[693.3, 194.24], [671.16, 194.28], [663.51, 192.72], [684.06, 192.69], [693.3, 171.03], [671.16, 171.03], [663.51, 171.16], [684.06, 171.17]]}, {\"category\": \"car\", \"corners_3d\": [[411.61, 365.87], [224.26, 366.9], [374.04, 288.41], [485.37, 288.04], [411.61, 194.88], [224.26, 195.0], [374.04, 186.04], [485.37, 186.0]]}, {\"category\": \"van\", \"corners_3d\": [[753.81, 288.1], [868.21, 287.19], [1080.73, 369.79], [885.48, 372.5], [753.81, 155.61], [868.21, 155.74], [1080.73, 143.38], [885.48, 142.97]]}]\n```" - } - ] - }, - { - "id": 3175, - "data_source": "Mono3DRefer", - "scene_name": "005083", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[336.74, 251.48], [270.7, 253.16], [297.46, 239.21], [351.76, 238.06], [336.74, 188.5], [270.7, 188.83], [297.46, 186.06], [351.76, 185.83]]}, {\"category\": \"van\", \"corners_3d\": [[431.86, 209.32], [465.01, 208.92], [469.6, 213.16], [432.57, 213.66], [431.86, 173.56], [465.01, 173.55], [469.6, 173.64], [432.57, 173.65]]}]\n```" - } - ] - }, - { - "id": 3176, - "data_source": "Mono3DRefer", - "scene_name": "005083", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[545.48, 262.96], [585.99, 261.8], [605.75, 273.94], [559.92, 275.44], [545.48, 175.35], [585.99, 175.32], [605.75, 175.66], [559.92, 175.7]]}]\n```" - } - ] - }, - { - "id": 3177, - "data_source": "Mono3DRefer", - "scene_name": "005083", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[543.08, 229.1], [593.08, 228.09], [621.94, 237.21], [564.12, 238.58], [543.08, 180.71], [593.08, 180.57], [621.94, 181.84], [564.12, 182.03]]}, {\"category\": \"car\", \"corners_3d\": [[516.31, 220.95], [558.43, 220.06], [585.41, 226.86], [537.67, 228.03], [516.31, 182.87], [558.43, 182.68], [585.41, 184.1], [537.67, 184.34]]}, {\"category\": \"car\", \"corners_3d\": [[354.15, 233.05], [304.98, 234.27], [313.16, 226.52], [356.09, 225.59], [354.15, 192.16], [304.98, 192.55], [313.16, 190.06], [356.09, 189.76]]}, {\"category\": \"car\", \"corners_3d\": [[500.46, 214.97], [532.51, 214.47], [545.84, 219.02], [510.43, 219.63], [500.46, 185.68], [532.51, 185.53], [545.84, 186.91], [510.43, 187.1]]}, {\"category\": \"car\", \"corners_3d\": [[768.94, 260.86], [862.01, 265.26], [841.09, 301.55], [715.21, 293.17], [768.94, 182.21], [862.01, 182.68], [841.09, 186.54], [715.21, 185.65]]}, {\"category\": \"car\", \"corners_3d\": [[604.25, 246.54], [666.51, 244.85], [734.19, 264.48], [656.48, 267.22], [604.25, 185.75], [666.51, 185.45], [734.19, 188.89], [656.48, 189.37]]}, {\"category\": \"car\", \"corners_3d\": [[297.96, 266.21], [209.98, 268.44], [261.77, 248.25], [330.56, 246.86], [297.96, 194.39], [209.98, 194.9], [261.77, 190.24], [330.56, 189.92]]}]\n```" - } - ] - }, - { - "id": 3178, - "data_source": "Mono3DRefer", - "scene_name": "005083", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[336.74, 251.48], [270.7, 253.16], [297.46, 239.21], [351.76, 238.06], [336.74, 188.5], [270.7, 188.83], [297.46, 186.06], [351.76, 185.83]]}, {\"category\": \"van\", \"corners_3d\": [[431.86, 209.32], [465.01, 208.92], [469.6, 213.16], [432.57, 213.66], [431.86, 173.56], [465.01, 173.55], [469.6, 173.64], [432.57, 173.65]]}, {\"category\": \"cyclist\", \"corners_3d\": [[545.48, 262.96], [585.99, 261.8], [605.75, 273.94], [559.92, 275.44], [545.48, 175.35], [585.99, 175.32], [605.75, 175.66], [559.92, 175.7]]}]\n```" - } - ] - }, - { - "id": 3179, - "data_source": "Mono3DRefer", - "scene_name": "005083", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[336.74, 251.48], [270.7, 253.16], [297.46, 239.21], [351.76, 238.06], [336.74, 188.5], [270.7, 188.83], [297.46, 186.06], [351.76, 185.83]]}, {\"category\": \"car\", \"corners_3d\": [[543.08, 229.1], [593.08, 228.09], [621.94, 237.21], [564.12, 238.58], [543.08, 180.71], [593.08, 180.57], [621.94, 181.84], [564.12, 182.03]]}, {\"category\": \"car\", \"corners_3d\": [[516.31, 220.95], [558.43, 220.06], [585.41, 226.86], [537.67, 228.03], [516.31, 182.87], [558.43, 182.68], [585.41, 184.1], [537.67, 184.34]]}, {\"category\": \"car\", \"corners_3d\": [[354.15, 233.05], [304.98, 234.27], [313.16, 226.52], [356.09, 225.59], [354.15, 192.16], [304.98, 192.55], [313.16, 190.06], [356.09, 189.76]]}, {\"category\": \"car\", \"corners_3d\": [[500.46, 214.97], [532.51, 214.47], [545.84, 219.02], [510.43, 219.63], [500.46, 185.68], [532.51, 185.53], [545.84, 186.91], [510.43, 187.1]]}, {\"category\": \"van\", \"corners_3d\": [[431.86, 209.32], [465.01, 208.92], [469.6, 213.16], [432.57, 213.66], [431.86, 173.56], [465.01, 173.55], [469.6, 173.64], [432.57, 173.65]]}, {\"category\": \"car\", \"corners_3d\": [[768.94, 260.86], [862.01, 265.26], [841.09, 301.55], [715.21, 293.17], [768.94, 182.21], [862.01, 182.68], [841.09, 186.54], [715.21, 185.65]]}, {\"category\": \"car\", \"corners_3d\": [[604.25, 246.54], [666.51, 244.85], [734.19, 264.48], [656.48, 267.22], [604.25, 185.75], [666.51, 185.45], [734.19, 188.89], [656.48, 189.37]]}, {\"category\": \"car\", \"corners_3d\": [[297.96, 266.21], [209.98, 268.44], [261.77, 248.25], [330.56, 246.86], [297.96, 194.39], [209.98, 194.9], [261.77, 190.24], [330.56, 189.92]]}]\n```" - } - ] - }, - { - "id": 3180, - "data_source": "Mono3DRefer", - "scene_name": "005083", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[543.08, 229.1], [593.08, 228.09], [621.94, 237.21], [564.12, 238.58], [543.08, 180.71], [593.08, 180.57], [621.94, 181.84], [564.12, 182.03]]}, {\"category\": \"car\", \"corners_3d\": [[516.31, 220.95], [558.43, 220.06], [585.41, 226.86], [537.67, 228.03], [516.31, 182.87], [558.43, 182.68], [585.41, 184.1], [537.67, 184.34]]}, {\"category\": \"car\", \"corners_3d\": [[354.15, 233.05], [304.98, 234.27], [313.16, 226.52], [356.09, 225.59], [354.15, 192.16], [304.98, 192.55], [313.16, 190.06], [356.09, 189.76]]}, {\"category\": \"car\", \"corners_3d\": [[500.46, 214.97], [532.51, 214.47], [545.84, 219.02], [510.43, 219.63], [500.46, 185.68], [532.51, 185.53], [545.84, 186.91], [510.43, 187.1]]}, {\"category\": \"cyclist\", \"corners_3d\": [[545.48, 262.96], [585.99, 261.8], [605.75, 273.94], [559.92, 275.44], [545.48, 175.35], [585.99, 175.32], [605.75, 175.66], [559.92, 175.7]]}, {\"category\": \"car\", \"corners_3d\": [[768.94, 260.86], [862.01, 265.26], [841.09, 301.55], [715.21, 293.17], [768.94, 182.21], [862.01, 182.68], [841.09, 186.54], [715.21, 185.65]]}, {\"category\": \"car\", \"corners_3d\": [[604.25, 246.54], [666.51, 244.85], [734.19, 264.48], [656.48, 267.22], [604.25, 185.75], [666.51, 185.45], [734.19, 188.89], [656.48, 189.37]]}, {\"category\": \"car\", \"corners_3d\": [[297.96, 266.21], [209.98, 268.44], [261.77, 248.25], [330.56, 246.86], [297.96, 194.39], [209.98, 194.9], [261.77, 190.24], [330.56, 189.92]]}]\n```" - } - ] - }, - { - "id": 3181, - "data_source": "Mono3DRefer", - "scene_name": "005083", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[336.74, 251.48], [270.7, 253.16], [297.46, 239.21], [351.76, 238.06], [336.74, 188.5], [270.7, 188.83], [297.46, 186.06], [351.76, 185.83]]}, {\"category\": \"car\", \"corners_3d\": [[543.08, 229.1], [593.08, 228.09], [621.94, 237.21], [564.12, 238.58], [543.08, 180.71], [593.08, 180.57], [621.94, 181.84], [564.12, 182.03]]}, {\"category\": \"car\", \"corners_3d\": [[516.31, 220.95], [558.43, 220.06], [585.41, 226.86], [537.67, 228.03], [516.31, 182.87], [558.43, 182.68], [585.41, 184.1], [537.67, 184.34]]}, {\"category\": \"car\", \"corners_3d\": [[354.15, 233.05], [304.98, 234.27], [313.16, 226.52], [356.09, 225.59], [354.15, 192.16], [304.98, 192.55], [313.16, 190.06], [356.09, 189.76]]}, {\"category\": \"car\", \"corners_3d\": [[500.46, 214.97], [532.51, 214.47], [545.84, 219.02], [510.43, 219.63], [500.46, 185.68], [532.51, 185.53], [545.84, 186.91], [510.43, 187.1]]}, {\"category\": \"van\", \"corners_3d\": [[431.86, 209.32], [465.01, 208.92], [469.6, 213.16], [432.57, 213.66], [431.86, 173.56], [465.01, 173.55], [469.6, 173.64], [432.57, 173.65]]}, {\"category\": \"cyclist\", \"corners_3d\": [[545.48, 262.96], [585.99, 261.8], [605.75, 273.94], [559.92, 275.44], [545.48, 175.35], [585.99, 175.32], [605.75, 175.66], [559.92, 175.7]]}, {\"category\": \"car\", \"corners_3d\": [[768.94, 260.86], [862.01, 265.26], [841.09, 301.55], [715.21, 293.17], [768.94, 182.21], [862.01, 182.68], [841.09, 186.54], [715.21, 185.65]]}, {\"category\": \"car\", \"corners_3d\": [[604.25, 246.54], [666.51, 244.85], [734.19, 264.48], [656.48, 267.22], [604.25, 185.75], [666.51, 185.45], [734.19, 188.89], [656.48, 189.37]]}, {\"category\": \"car\", \"corners_3d\": [[297.96, 266.21], [209.98, 268.44], [261.77, 248.25], [330.56, 246.86], [297.96, 194.39], [209.98, 194.9], [261.77, 190.24], [330.56, 189.92]]}]\n```" - } - ] - }, - { - "id": 3182, - "data_source": "Mono3DRefer", - "scene_name": "005100", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[552.05, 224.15], [506.98, 224.16], [520.64, 217.28], [559.67, 217.28], [552.05, 185.6], [506.98, 185.6], [520.64, 183.9], [559.67, 183.9]]}, {\"category\": \"car\", \"corners_3d\": [[590.21, 186.2], [571.42, 186.21], [572.8, 185.41], [590.47, 185.41], [590.21, 166.43], [571.42, 166.42], [572.8, 166.81], [590.47, 166.81]]}]\n```" - } - ] - }, - { - "id": 3183, - "data_source": "Mono3DRefer", - "scene_name": "005116", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[658.08, 212.23], [693.75, 212.29], [701.95, 217.51], [661.56, 217.43], [658.08, 171.7], [693.75, 171.7], [701.95, 171.55], [661.56, 171.55]]}, {\"category\": \"van\", \"corners_3d\": [[467.48, 205.89], [446.27, 205.89], [457.99, 203.51], [477.68, 203.51], [467.48, 182.64], [446.27, 182.64], [457.99, 181.93], [477.68, 181.93]]}]\n```" - } - ] - }, - { - "id": 3184, - "data_source": "Mono3DRefer", - "scene_name": "005116", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[770.09, 222.75], [766.48, 221.2], [819.86, 221.42], [825.21, 222.99], [770.09, 177.43], [766.48, 177.29], [819.86, 177.31], [825.21, 177.45]]}]\n```" - } - ] - }, - { - "id": 3185, - "data_source": "Mono3DRefer", - "scene_name": "005116", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1114.49, 266.17], [1082.1, 266.26], [1049.73, 260.05], [1080.0, 259.98], [1114.49, 164.19], [1082.1, 164.18], [1049.73, 164.76], [1080.0, 164.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[537.5, 223.45], [537.99, 224.57], [521.47, 224.72], [521.34, 223.59], [537.5, 177.29], [537.99, 177.39], [521.47, 177.4], [521.34, 177.3]]}]\n```" - } - ] - }, - { - "id": 3186, - "data_source": "Mono3DRefer", - "scene_name": "005116", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.57, 234.6], [665.34, 234.61], [676.74, 247.36], [602.22, 247.36], [603.57, 184.9], [665.34, 184.9], [676.74, 187.39], [602.22, 187.39]]}, {\"category\": \"car\", \"corners_3d\": [[1151.84, 226.56], [1109.36, 222.76], [1231.93, 222.03], [1282.91, 225.71], [1151.84, 178.77], [1109.36, 178.36], [1231.93, 178.27], [1282.91, 178.68]]}, {\"category\": \"car\", \"corners_3d\": [[387.61, 221.57], [351.6, 221.55], [378.32, 216.63], [410.7, 216.65], [387.61, 188.18], [351.6, 188.17], [378.32, 186.62], [410.7, 186.63]]}, {\"category\": \"car\", \"corners_3d\": [[480.77, 213.02], [503.33, 211.71], [552.75, 214.05], [530.55, 215.53], [480.77, 182.89], [503.33, 182.56], [552.75, 183.15], [530.55, 183.52]]}, {\"category\": \"car\", \"corners_3d\": [[588.5, 220.2], [627.27, 220.25], [627.63, 226.49], [583.76, 226.42], [588.5, 184.92], [627.27, 184.93], [627.63, 186.52], [583.76, 186.5]]}]\n```" - } - ] - }, - { - "id": 3187, - "data_source": "Mono3DRefer", - "scene_name": "005116", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[770.09, 222.75], [766.48, 221.2], [819.86, 221.42], [825.21, 222.99], [770.09, 177.43], [766.48, 177.29], [819.86, 177.31], [825.21, 177.45]]}, {\"category\": \"van\", \"corners_3d\": [[658.08, 212.23], [693.75, 212.29], [701.95, 217.51], [661.56, 217.43], [658.08, 171.7], [693.75, 171.7], [701.95, 171.55], [661.56, 171.55]]}, {\"category\": \"van\", \"corners_3d\": [[467.48, 205.89], [446.27, 205.89], [457.99, 203.51], [477.68, 203.51], [467.48, 182.64], [446.27, 182.64], [457.99, 181.93], [477.68, 181.93]]}]\n```" - } - ] - }, - { - "id": 3188, - "data_source": "Mono3DRefer", - "scene_name": "005116", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1114.49, 266.17], [1082.1, 266.26], [1049.73, 260.05], [1080.0, 259.98], [1114.49, 164.19], [1082.1, 164.18], [1049.73, 164.76], [1080.0, 164.77]]}, {\"category\": \"van\", \"corners_3d\": [[658.08, 212.23], [693.75, 212.29], [701.95, 217.51], [661.56, 217.43], [658.08, 171.7], [693.75, 171.7], [701.95, 171.55], [661.56, 171.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[537.5, 223.45], [537.99, 224.57], [521.47, 224.72], [521.34, 223.59], [537.5, 177.29], [537.99, 177.39], [521.47, 177.4], [521.34, 177.3]]}, {\"category\": \"van\", \"corners_3d\": [[467.48, 205.89], [446.27, 205.89], [457.99, 203.51], [477.68, 203.51], [467.48, 182.64], [446.27, 182.64], [457.99, 181.93], [477.68, 181.93]]}]\n```" - } - ] - }, - { - "id": 3189, - "data_source": "Mono3DRefer", - "scene_name": "005116", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.57, 234.6], [665.34, 234.61], [676.74, 247.36], [602.22, 247.36], [603.57, 184.9], [665.34, 184.9], [676.74, 187.39], [602.22, 187.39]]}, {\"category\": \"car\", \"corners_3d\": [[1151.84, 226.56], [1109.36, 222.76], [1231.93, 222.03], [1282.91, 225.71], [1151.84, 178.77], [1109.36, 178.36], [1231.93, 178.27], [1282.91, 178.68]]}, {\"category\": \"car\", \"corners_3d\": [[387.61, 221.57], [351.6, 221.55], [378.32, 216.63], [410.7, 216.65], [387.61, 188.18], [351.6, 188.17], [378.32, 186.62], [410.7, 186.63]]}, {\"category\": \"van\", \"corners_3d\": [[658.08, 212.23], [693.75, 212.29], [701.95, 217.51], [661.56, 217.43], [658.08, 171.7], [693.75, 171.7], [701.95, 171.55], [661.56, 171.55]]}, {\"category\": \"van\", \"corners_3d\": [[467.48, 205.89], [446.27, 205.89], [457.99, 203.51], [477.68, 203.51], [467.48, 182.64], [446.27, 182.64], [457.99, 181.93], [477.68, 181.93]]}, {\"category\": \"car\", \"corners_3d\": [[480.77, 213.02], [503.33, 211.71], [552.75, 214.05], [530.55, 215.53], [480.77, 182.89], [503.33, 182.56], [552.75, 183.15], [530.55, 183.52]]}, {\"category\": \"car\", \"corners_3d\": [[588.5, 220.2], [627.27, 220.25], [627.63, 226.49], [583.76, 226.42], [588.5, 184.92], [627.27, 184.93], [627.63, 186.52], [583.76, 186.5]]}]\n```" - } - ] - }, - { - "id": 3190, - "data_source": "Mono3DRefer", - "scene_name": "005116", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1114.49, 266.17], [1082.1, 266.26], [1049.73, 260.05], [1080.0, 259.98], [1114.49, 164.19], [1082.1, 164.18], [1049.73, 164.76], [1080.0, 164.77]]}, {\"category\": \"cyclist\", \"corners_3d\": [[770.09, 222.75], [766.48, 221.2], [819.86, 221.42], [825.21, 222.99], [770.09, 177.43], [766.48, 177.29], [819.86, 177.31], [825.21, 177.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[537.5, 223.45], [537.99, 224.57], [521.47, 224.72], [521.34, 223.59], [537.5, 177.29], [537.99, 177.39], [521.47, 177.4], [521.34, 177.3]]}]\n```" - } - ] - }, - { - "id": 3191, - "data_source": "Mono3DRefer", - "scene_name": "005116", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.57, 234.6], [665.34, 234.61], [676.74, 247.36], [602.22, 247.36], [603.57, 184.9], [665.34, 184.9], [676.74, 187.39], [602.22, 187.39]]}, {\"category\": \"car\", \"corners_3d\": [[1151.84, 226.56], [1109.36, 222.76], [1231.93, 222.03], [1282.91, 225.71], [1151.84, 178.77], [1109.36, 178.36], [1231.93, 178.27], [1282.91, 178.68]]}, {\"category\": \"car\", \"corners_3d\": [[387.61, 221.57], [351.6, 221.55], [378.32, 216.63], [410.7, 216.65], [387.61, 188.18], [351.6, 188.17], [378.32, 186.62], [410.7, 186.63]]}, {\"category\": \"cyclist\", \"corners_3d\": [[770.09, 222.75], [766.48, 221.2], [819.86, 221.42], [825.21, 222.99], [770.09, 177.43], [766.48, 177.29], [819.86, 177.31], [825.21, 177.45]]}, {\"category\": \"car\", \"corners_3d\": [[480.77, 213.02], [503.33, 211.71], [552.75, 214.05], [530.55, 215.53], [480.77, 182.89], [503.33, 182.56], [552.75, 183.15], [530.55, 183.52]]}, {\"category\": \"car\", \"corners_3d\": [[588.5, 220.2], [627.27, 220.25], [627.63, 226.49], [583.76, 226.42], [588.5, 184.92], [627.27, 184.93], [627.63, 186.52], [583.76, 186.5]]}]\n```" - } - ] - }, - { - "id": 3192, - "data_source": "Mono3DRefer", - "scene_name": "005116", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.57, 234.6], [665.34, 234.61], [676.74, 247.36], [602.22, 247.36], [603.57, 184.9], [665.34, 184.9], [676.74, 187.39], [602.22, 187.39]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1114.49, 266.17], [1082.1, 266.26], [1049.73, 260.05], [1080.0, 259.98], [1114.49, 164.19], [1082.1, 164.18], [1049.73, 164.76], [1080.0, 164.77]]}, {\"category\": \"car\", \"corners_3d\": [[1151.84, 226.56], [1109.36, 222.76], [1231.93, 222.03], [1282.91, 225.71], [1151.84, 178.77], [1109.36, 178.36], [1231.93, 178.27], [1282.91, 178.68]]}, {\"category\": \"car\", \"corners_3d\": [[387.61, 221.57], [351.6, 221.55], [378.32, 216.63], [410.7, 216.65], [387.61, 188.18], [351.6, 188.17], [378.32, 186.62], [410.7, 186.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[537.5, 223.45], [537.99, 224.57], [521.47, 224.72], [521.34, 223.59], [537.5, 177.29], [537.99, 177.39], [521.47, 177.4], [521.34, 177.3]]}, {\"category\": \"car\", \"corners_3d\": [[480.77, 213.02], [503.33, 211.71], [552.75, 214.05], [530.55, 215.53], [480.77, 182.89], [503.33, 182.56], [552.75, 183.15], [530.55, 183.52]]}, {\"category\": \"car\", \"corners_3d\": [[588.5, 220.2], [627.27, 220.25], [627.63, 226.49], [583.76, 226.42], [588.5, 184.92], [627.27, 184.93], [627.63, 186.52], [583.76, 186.5]]}]\n```" - } - ] - }, - { - "id": 3193, - "data_source": "Mono3DRefer", - "scene_name": "005116", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1114.49, 266.17], [1082.1, 266.26], [1049.73, 260.05], [1080.0, 259.98], [1114.49, 164.19], [1082.1, 164.18], [1049.73, 164.76], [1080.0, 164.77]]}, {\"category\": \"cyclist\", \"corners_3d\": [[770.09, 222.75], [766.48, 221.2], [819.86, 221.42], [825.21, 222.99], [770.09, 177.43], [766.48, 177.29], [819.86, 177.31], [825.21, 177.45]]}, {\"category\": \"van\", \"corners_3d\": [[658.08, 212.23], [693.75, 212.29], [701.95, 217.51], [661.56, 217.43], [658.08, 171.7], [693.75, 171.7], [701.95, 171.55], [661.56, 171.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[537.5, 223.45], [537.99, 224.57], [521.47, 224.72], [521.34, 223.59], [537.5, 177.29], [537.99, 177.39], [521.47, 177.4], [521.34, 177.3]]}, {\"category\": \"van\", \"corners_3d\": [[467.48, 205.89], [446.27, 205.89], [457.99, 203.51], [477.68, 203.51], [467.48, 182.64], [446.27, 182.64], [457.99, 181.93], [477.68, 181.93]]}]\n```" - } - ] - }, - { - "id": 3194, - "data_source": "Mono3DRefer", - "scene_name": "005116", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.57, 234.6], [665.34, 234.61], [676.74, 247.36], [602.22, 247.36], [603.57, 184.9], [665.34, 184.9], [676.74, 187.39], [602.22, 187.39]]}, {\"category\": \"car\", \"corners_3d\": [[1151.84, 226.56], [1109.36, 222.76], [1231.93, 222.03], [1282.91, 225.71], [1151.84, 178.77], [1109.36, 178.36], [1231.93, 178.27], [1282.91, 178.68]]}, {\"category\": \"car\", \"corners_3d\": [[387.61, 221.57], [351.6, 221.55], [378.32, 216.63], [410.7, 216.65], [387.61, 188.18], [351.6, 188.17], [378.32, 186.62], [410.7, 186.63]]}, {\"category\": \"cyclist\", \"corners_3d\": [[770.09, 222.75], [766.48, 221.2], [819.86, 221.42], [825.21, 222.99], [770.09, 177.43], [766.48, 177.29], [819.86, 177.31], [825.21, 177.45]]}, {\"category\": \"van\", \"corners_3d\": [[658.08, 212.23], [693.75, 212.29], [701.95, 217.51], [661.56, 217.43], [658.08, 171.7], [693.75, 171.7], [701.95, 171.55], [661.56, 171.55]]}, {\"category\": \"van\", \"corners_3d\": [[467.48, 205.89], [446.27, 205.89], [457.99, 203.51], [477.68, 203.51], [467.48, 182.64], [446.27, 182.64], [457.99, 181.93], [477.68, 181.93]]}, {\"category\": \"car\", \"corners_3d\": [[480.77, 213.02], [503.33, 211.71], [552.75, 214.05], [530.55, 215.53], [480.77, 182.89], [503.33, 182.56], [552.75, 183.15], [530.55, 183.52]]}, {\"category\": \"car\", \"corners_3d\": [[588.5, 220.2], [627.27, 220.25], [627.63, 226.49], [583.76, 226.42], [588.5, 184.92], [627.27, 184.93], [627.63, 186.52], [583.76, 186.5]]}]\n```" - } - ] - }, - { - "id": 3195, - "data_source": "Mono3DRefer", - "scene_name": "005116", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.57, 234.6], [665.34, 234.61], [676.74, 247.36], [602.22, 247.36], [603.57, 184.9], [665.34, 184.9], [676.74, 187.39], [602.22, 187.39]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1114.49, 266.17], [1082.1, 266.26], [1049.73, 260.05], [1080.0, 259.98], [1114.49, 164.19], [1082.1, 164.18], [1049.73, 164.76], [1080.0, 164.77]]}, {\"category\": \"car\", \"corners_3d\": [[1151.84, 226.56], [1109.36, 222.76], [1231.93, 222.03], [1282.91, 225.71], [1151.84, 178.77], [1109.36, 178.36], [1231.93, 178.27], [1282.91, 178.68]]}, {\"category\": \"car\", \"corners_3d\": [[387.61, 221.57], [351.6, 221.55], [378.32, 216.63], [410.7, 216.65], [387.61, 188.18], [351.6, 188.17], [378.32, 186.62], [410.7, 186.63]]}, {\"category\": \"van\", \"corners_3d\": [[658.08, 212.23], [693.75, 212.29], [701.95, 217.51], [661.56, 217.43], [658.08, 171.7], [693.75, 171.7], [701.95, 171.55], [661.56, 171.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[537.5, 223.45], [537.99, 224.57], [521.47, 224.72], [521.34, 223.59], [537.5, 177.29], [537.99, 177.39], [521.47, 177.4], [521.34, 177.3]]}, {\"category\": \"van\", \"corners_3d\": [[467.48, 205.89], [446.27, 205.89], [457.99, 203.51], [477.68, 203.51], [467.48, 182.64], [446.27, 182.64], [457.99, 181.93], [477.68, 181.93]]}, {\"category\": \"car\", \"corners_3d\": [[480.77, 213.02], [503.33, 211.71], [552.75, 214.05], [530.55, 215.53], [480.77, 182.89], [503.33, 182.56], [552.75, 183.15], [530.55, 183.52]]}, {\"category\": \"car\", \"corners_3d\": [[588.5, 220.2], [627.27, 220.25], [627.63, 226.49], [583.76, 226.42], [588.5, 184.92], [627.27, 184.93], [627.63, 186.52], [583.76, 186.5]]}]\n```" - } - ] - }, - { - "id": 3196, - "data_source": "Mono3DRefer", - "scene_name": "005116", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.57, 234.6], [665.34, 234.61], [676.74, 247.36], [602.22, 247.36], [603.57, 184.9], [665.34, 184.9], [676.74, 187.39], [602.22, 187.39]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1114.49, 266.17], [1082.1, 266.26], [1049.73, 260.05], [1080.0, 259.98], [1114.49, 164.19], [1082.1, 164.18], [1049.73, 164.76], [1080.0, 164.77]]}, {\"category\": \"car\", \"corners_3d\": [[1151.84, 226.56], [1109.36, 222.76], [1231.93, 222.03], [1282.91, 225.71], [1151.84, 178.77], [1109.36, 178.36], [1231.93, 178.27], [1282.91, 178.68]]}, {\"category\": \"car\", \"corners_3d\": [[387.61, 221.57], [351.6, 221.55], [378.32, 216.63], [410.7, 216.65], [387.61, 188.18], [351.6, 188.17], [378.32, 186.62], [410.7, 186.63]]}, {\"category\": \"cyclist\", \"corners_3d\": [[770.09, 222.75], [766.48, 221.2], [819.86, 221.42], [825.21, 222.99], [770.09, 177.43], [766.48, 177.29], [819.86, 177.31], [825.21, 177.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[537.5, 223.45], [537.99, 224.57], [521.47, 224.72], [521.34, 223.59], [537.5, 177.29], [537.99, 177.39], [521.47, 177.4], [521.34, 177.3]]}, {\"category\": \"car\", \"corners_3d\": [[480.77, 213.02], [503.33, 211.71], [552.75, 214.05], [530.55, 215.53], [480.77, 182.89], [503.33, 182.56], [552.75, 183.15], [530.55, 183.52]]}, {\"category\": \"car\", \"corners_3d\": [[588.5, 220.2], [627.27, 220.25], [627.63, 226.49], [583.76, 226.42], [588.5, 184.92], [627.27, 184.93], [627.63, 186.52], [583.76, 186.5]]}]\n```" - } - ] - }, - { - "id": 3197, - "data_source": "Mono3DRefer", - "scene_name": "005116", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.57, 234.6], [665.34, 234.61], [676.74, 247.36], [602.22, 247.36], [603.57, 184.9], [665.34, 184.9], [676.74, 187.39], [602.22, 187.39]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1114.49, 266.17], [1082.1, 266.26], [1049.73, 260.05], [1080.0, 259.98], [1114.49, 164.19], [1082.1, 164.18], [1049.73, 164.76], [1080.0, 164.77]]}, {\"category\": \"car\", \"corners_3d\": [[1151.84, 226.56], [1109.36, 222.76], [1231.93, 222.03], [1282.91, 225.71], [1151.84, 178.77], [1109.36, 178.36], [1231.93, 178.27], [1282.91, 178.68]]}, {\"category\": \"car\", \"corners_3d\": [[387.61, 221.57], [351.6, 221.55], [378.32, 216.63], [410.7, 216.65], [387.61, 188.18], [351.6, 188.17], [378.32, 186.62], [410.7, 186.63]]}, {\"category\": \"cyclist\", \"corners_3d\": [[770.09, 222.75], [766.48, 221.2], [819.86, 221.42], [825.21, 222.99], [770.09, 177.43], [766.48, 177.29], [819.86, 177.31], [825.21, 177.45]]}, {\"category\": \"van\", \"corners_3d\": [[658.08, 212.23], [693.75, 212.29], [701.95, 217.51], [661.56, 217.43], [658.08, 171.7], [693.75, 171.7], [701.95, 171.55], [661.56, 171.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[537.5, 223.45], [537.99, 224.57], [521.47, 224.72], [521.34, 223.59], [537.5, 177.29], [537.99, 177.39], [521.47, 177.4], [521.34, 177.3]]}, {\"category\": \"van\", \"corners_3d\": [[467.48, 205.89], [446.27, 205.89], [457.99, 203.51], [477.68, 203.51], [467.48, 182.64], [446.27, 182.64], [457.99, 181.93], [477.68, 181.93]]}, {\"category\": \"car\", \"corners_3d\": [[480.77, 213.02], [503.33, 211.71], [552.75, 214.05], [530.55, 215.53], [480.77, 182.89], [503.33, 182.56], [552.75, 183.15], [530.55, 183.52]]}, {\"category\": \"car\", \"corners_3d\": [[588.5, 220.2], [627.27, 220.25], [627.63, 226.49], [583.76, 226.42], [588.5, 184.92], [627.27, 184.93], [627.63, 186.52], [583.76, 186.5]]}]\n```" - } - ] - }, - { - "id": 3198, - "data_source": "Mono3DRefer", - "scene_name": "005127", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[982.51, 268.6], [1110.35, 277.32], [1108.61, 318.91], [935.83, 302.41], [982.51, 180.76], [1110.35, 181.48], [1108.61, 184.91], [935.83, 183.55]]}, {\"category\": \"car\", \"corners_3d\": [[918.81, 211.66], [966.52, 213.18], [932.58, 219.52], [879.16, 217.5], [918.81, 171.48], [966.52, 171.42], [932.58, 171.2], [879.16, 171.27]]}, {\"category\": \"car\", \"corners_3d\": [[1072.99, 220.18], [1145.98, 222.9], [1126.91, 231.55], [1043.31, 227.84], [1072.99, 163.44], [1145.98, 162.9], [1126.91, 161.18], [1043.31, 161.92]]}, {\"category\": \"car\", \"corners_3d\": [[961.43, 205.44], [1006.05, 206.76], [971.33, 210.63], [923.35, 208.99], [961.43, 170.19], [1006.05, 170.08], [971.33, 169.76], [923.35, 169.9]]}]\n```" - } - ] - }, - { - "id": 3199, - "data_source": "Mono3DRefer", - "scene_name": "005128", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[70.46, 517.07], [-285.07, 515.49], [205.6, 328.96], [368.19, 329.28], [70.46, 192.51], [-285.07, 192.42], [205.6, 181.77], [368.19, 181.79]]}, {\"category\": \"car\", \"corners_3d\": [[805.75, 332.86], [969.86, 332.89], [1379.5, 515.13], [1028.43, 515.0], [805.75, 197.83], [969.86, 197.83], [1379.5, 226.27], [1028.43, 226.25]]}, {\"category\": \"car\", \"corners_3d\": [[796.05, 268.26], [712.36, 268.38], [686.41, 245.97], [750.51, 245.9], [796.05, 193.4], [712.36, 193.42], [686.41, 188.6], [750.51, 188.58]]}, {\"category\": \"car\", \"corners_3d\": [[491.92, 260.36], [409.79, 260.86], [442.39, 243.27], [508.03, 242.95], [491.92, 190.14], [409.79, 190.24], [442.39, 186.77], [508.03, 186.7]]}, {\"category\": \"car\", \"corners_3d\": [[667.78, 226.69], [716.05, 226.65], [736.81, 236.53], [679.69, 236.57], [667.78, 184.47], [716.05, 184.46], [736.81, 186.59], [679.69, 186.6]]}, {\"category\": \"car\", \"corners_3d\": [[712.52, 224.78], [665.66, 224.85], [656.18, 217.92], [696.81, 217.86], [712.52, 183.11], [665.66, 183.13], [656.18, 181.76], [696.81, 181.75]]}]\n```" - } - ] - }, - { - "id": 3200, - "data_source": "Mono3DRefer", - "scene_name": "005130", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[489.8, 264.85], [406.87, 264.54], [451.34, 246.32], [517.86, 246.52], [489.8, 193.99], [406.87, 193.92], [451.34, 189.74], [517.86, 189.78]]}]\n```" - } - ] - }, - { - "id": 3201, - "data_source": "Mono3DRefer", - "scene_name": "005132", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[708.01, 227.19], [765.93, 227.8], [777.5, 239.55], [707.2, 238.65], [708.01, 173.53], [765.93, 173.54], [777.5, 173.68], [707.2, 173.67]]}, {\"category\": \"car\", \"corners_3d\": [[386.51, 290.6], [292.99, 287.49], [403.45, 259.64], [476.04, 261.41], [386.51, 174.8], [292.99, 174.75], [403.45, 174.29], [476.04, 174.32]]}, {\"category\": \"car\", \"corners_3d\": [[-3.07, 320.19], [-103.13, 315.52], [152.03, 272.74], [227.2, 275.0], [-3.07, 199.2], [-103.13, 198.36], [152.03, 190.71], [227.2, 191.12]]}, {\"category\": \"car\", \"corners_3d\": [[444.19, 234.87], [395.78, 233.54], [465.79, 223.62], [507.42, 224.55], [444.19, 187.77], [395.78, 187.45], [465.79, 185.06], [507.42, 185.29]]}, {\"category\": \"car\", \"corners_3d\": [[902.03, 191.46], [922.09, 192.19], [863.55, 192.73], [845.24, 191.95], [902.03, 164.42], [922.09, 164.09], [863.55, 163.84], [845.24, 164.2]]}]\n```" - } - ] - }, - { - "id": 3202, - "data_source": "Mono3DRefer", - "scene_name": "005133", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[761.43, 272.78], [780.7, 269.62], [894.31, 282.46], [876.61, 286.54], [761.43, 147.29], [780.7, 148.1], [894.31, 144.82], [876.61, 143.77]]}, {\"category\": \"cyclist\", \"corners_3d\": [[740.59, 254.3], [750.53, 251.63], [838.57, 256.8], [831.14, 259.84], [740.59, 153.21], [750.53, 153.85], [838.57, 152.61], [831.14, 151.87]]}]\n```" - } - ] - }, - { - "id": 3203, - "data_source": "Mono3DRefer", - "scene_name": "005133", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[599.25, 200.64], [624.19, 200.62], [626.51, 202.88], [599.54, 202.9], [599.25, 175.99], [624.19, 175.98], [626.51, 176.24], [599.54, 176.24]]}, {\"category\": \"car\", \"corners_3d\": [[219.8, 368.65], [393.71, 368.69], [-158.98, 868.8], [-776.33, 868.32], [219.8, 183.18], [393.71, 183.18], [-158.98, 209.56], [-776.33, 209.53]]}]\n```" - } - ] - }, - { - "id": 3204, - "data_source": "Mono3DRefer", - "scene_name": "005133", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[769.44, 268.41], [763.58, 265.08], [789.3, 265.04], [796.1, 268.37], [769.44, 161.47], [763.58, 161.86], [789.3, 161.87], [796.1, 161.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[736.34, 257.89], [721.02, 255.68], [735.55, 254.49], [751.02, 256.63], [736.34, 172.26], [721.02, 172.28], [735.55, 172.29], [751.02, 172.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[258.38, 281.91], [245.97, 285.85], [225.32, 285.88], [238.45, 281.94], [258.38, 176.14], [245.97, 176.25], [225.32, 176.26], [238.45, 176.14]]}]\n```" - } - ] - }, - { - "id": 3205, - "data_source": "Mono3DRefer", - "scene_name": "005133", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[599.25, 200.64], [624.19, 200.62], [626.51, 202.88], [599.54, 202.9], [599.25, 175.99], [624.19, 175.98], [626.51, 176.24], [599.54, 176.24]]}, {\"category\": \"car\", \"corners_3d\": [[219.8, 368.65], [393.71, 368.69], [-158.98, 868.8], [-776.33, 868.32], [219.8, 183.18], [393.71, 183.18], [-158.98, 209.56], [-776.33, 209.53]]}, {\"category\": \"cyclist\", \"corners_3d\": [[761.43, 272.78], [780.7, 269.62], [894.31, 282.46], [876.61, 286.54], [761.43, 147.29], [780.7, 148.1], [894.31, 144.82], [876.61, 143.77]]}, {\"category\": \"cyclist\", \"corners_3d\": [[740.59, 254.3], [750.53, 251.63], [838.57, 256.8], [831.14, 259.84], [740.59, 153.21], [750.53, 153.85], [838.57, 152.61], [831.14, 151.87]]}]\n```" - } - ] - }, - { - "id": 3206, - "data_source": "Mono3DRefer", - "scene_name": "005133", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[761.43, 272.78], [780.7, 269.62], [894.31, 282.46], [876.61, 286.54], [761.43, 147.29], [780.7, 148.1], [894.31, 144.82], [876.61, 143.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[769.44, 268.41], [763.58, 265.08], [789.3, 265.04], [796.1, 268.37], [769.44, 161.47], [763.58, 161.86], [789.3, 161.87], [796.1, 161.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[736.34, 257.89], [721.02, 255.68], [735.55, 254.49], [751.02, 256.63], [736.34, 172.26], [721.02, 172.28], [735.55, 172.29], [751.02, 172.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[258.38, 281.91], [245.97, 285.85], [225.32, 285.88], [238.45, 281.94], [258.38, 176.14], [245.97, 176.25], [225.32, 176.26], [238.45, 176.14]]}, {\"category\": \"cyclist\", \"corners_3d\": [[740.59, 254.3], [750.53, 251.63], [838.57, 256.8], [831.14, 259.84], [740.59, 153.21], [750.53, 153.85], [838.57, 152.61], [831.14, 151.87]]}]\n```" - } - ] - }, - { - "id": 3207, - "data_source": "Mono3DRefer", - "scene_name": "005133", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[599.25, 200.64], [624.19, 200.62], [626.51, 202.88], [599.54, 202.9], [599.25, 175.99], [624.19, 175.98], [626.51, 176.24], [599.54, 176.24]]}, {\"category\": \"car\", \"corners_3d\": [[219.8, 368.65], [393.71, 368.69], [-158.98, 868.8], [-776.33, 868.32], [219.8, 183.18], [393.71, 183.18], [-158.98, 209.56], [-776.33, 209.53]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[769.44, 268.41], [763.58, 265.08], [789.3, 265.04], [796.1, 268.37], [769.44, 161.47], [763.58, 161.86], [789.3, 161.87], [796.1, 161.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[736.34, 257.89], [721.02, 255.68], [735.55, 254.49], [751.02, 256.63], [736.34, 172.26], [721.02, 172.28], [735.55, 172.29], [751.02, 172.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[258.38, 281.91], [245.97, 285.85], [225.32, 285.88], [238.45, 281.94], [258.38, 176.14], [245.97, 176.25], [225.32, 176.26], [238.45, 176.14]]}]\n```" - } - ] - }, - { - "id": 3208, - "data_source": "Mono3DRefer", - "scene_name": "005133", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[599.25, 200.64], [624.19, 200.62], [626.51, 202.88], [599.54, 202.9], [599.25, 175.99], [624.19, 175.98], [626.51, 176.24], [599.54, 176.24]]}, {\"category\": \"car\", \"corners_3d\": [[219.8, 368.65], [393.71, 368.69], [-158.98, 868.8], [-776.33, 868.32], [219.8, 183.18], [393.71, 183.18], [-158.98, 209.56], [-776.33, 209.53]]}, {\"category\": \"cyclist\", \"corners_3d\": [[761.43, 272.78], [780.7, 269.62], [894.31, 282.46], [876.61, 286.54], [761.43, 147.29], [780.7, 148.1], [894.31, 144.82], [876.61, 143.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[769.44, 268.41], [763.58, 265.08], [789.3, 265.04], [796.1, 268.37], [769.44, 161.47], [763.58, 161.86], [789.3, 161.87], [796.1, 161.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[736.34, 257.89], [721.02, 255.68], [735.55, 254.49], [751.02, 256.63], [736.34, 172.26], [721.02, 172.28], [735.55, 172.29], [751.02, 172.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[258.38, 281.91], [245.97, 285.85], [225.32, 285.88], [238.45, 281.94], [258.38, 176.14], [245.97, 176.25], [225.32, 176.26], [238.45, 176.14]]}, {\"category\": \"cyclist\", \"corners_3d\": [[740.59, 254.3], [750.53, 251.63], [838.57, 256.8], [831.14, 259.84], [740.59, 153.21], [750.53, 153.85], [838.57, 152.61], [831.14, 151.87]]}]\n```" - } - ] - }, - { - "id": 3209, - "data_source": "Mono3DRefer", - "scene_name": "005142", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[549.57, 304.33], [546.33, 311.69], [441.26, 311.69], [450.39, 304.33], [549.57, 148.13], [546.33, 146.2], [441.26, 146.2], [450.39, 148.13]]}]\n```" - } - ] - }, - { - "id": 3210, - "data_source": "Mono3DRefer", - "scene_name": "005169", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[489.05, 217.55], [444.73, 217.58], [463.76, 212.16], [502.7, 212.14], [489.05, 180.66], [444.73, 180.67], [463.76, 179.72], [502.7, 179.72]]}]\n```" - } - ] - }, - { - "id": 3211, - "data_source": "Mono3DRefer", - "scene_name": "005170", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[1244.59, 370.2], [1145.33, 366.54], [1101.59, 347.7], [1189.79, 350.65], [1244.59, 190.96], [1145.33, 190.78], [1101.59, 189.83], [1189.79, 189.98]]}]\n```" - } - ] - }, - { - "id": 3212, - "data_source": "Mono3DRefer", - "scene_name": "005170", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[434.49, 316.75], [248.36, 314.28], [398.47, 264.68], [515.99, 265.65], [434.49, 116.99], [248.36, 118.17], [398.47, 141.83], [515.99, 141.37]]}]\n```" - } - ] - }, - { - "id": 3213, - "data_source": "Mono3DRefer", - "scene_name": "005170", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[542.77, 245.98], [452.61, 245.66], [498.32, 228.99], [565.14, 229.17], [542.77, 139.76], [452.61, 139.97], [498.32, 150.76], [565.14, 150.65]]}]\n```" - } - ] - }, - { - "id": 3214, - "data_source": "Mono3DRefer", - "scene_name": "005170", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[899.21, 427.61], [1028.38, 418.94], [1196.6, 490.51], [1034.72, 505.35], [899.21, 133.32], [1028.38, 135.02], [1196.6, 121.0], [1034.72, 118.1]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[506.14, 324.08], [423.6, 320.86], [461.65, 303.99], [534.77, 306.47], [506.14, 146.68], [423.6, 147.46], [461.65, 151.59], [534.77, 150.98]]}]\n```" - } - ] - }, - { - "id": 3215, - "data_source": "Mono3DRefer", - "scene_name": "005170", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[434.49, 316.75], [248.36, 314.28], [398.47, 264.68], [515.99, 265.65], [434.49, 116.99], [248.36, 118.17], [398.47, 141.83], [515.99, 141.37]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1244.59, 370.2], [1145.33, 366.54], [1101.59, 347.7], [1189.79, 350.65], [1244.59, 190.96], [1145.33, 190.78], [1101.59, 189.83], [1189.79, 189.98]]}]\n```" - } - ] - }, - { - "id": 3216, - "data_source": "Mono3DRefer", - "scene_name": "005170", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[542.77, 245.98], [452.61, 245.66], [498.32, 228.99], [565.14, 229.17], [542.77, 139.76], [452.61, 139.97], [498.32, 150.76], [565.14, 150.65]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1244.59, 370.2], [1145.33, 366.54], [1101.59, 347.7], [1189.79, 350.65], [1244.59, 190.96], [1145.33, 190.78], [1101.59, 189.83], [1189.79, 189.98]]}]\n```" - } - ] - }, - { - "id": 3217, - "data_source": "Mono3DRefer", - "scene_name": "005170", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[899.21, 427.61], [1028.38, 418.94], [1196.6, 490.51], [1034.72, 505.35], [899.21, 133.32], [1028.38, 135.02], [1196.6, 121.0], [1034.72, 118.1]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[506.14, 324.08], [423.6, 320.86], [461.65, 303.99], [534.77, 306.47], [506.14, 146.68], [423.6, 147.46], [461.65, 151.59], [534.77, 150.98]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1244.59, 370.2], [1145.33, 366.54], [1101.59, 347.7], [1189.79, 350.65], [1244.59, 190.96], [1145.33, 190.78], [1101.59, 189.83], [1189.79, 189.98]]}]\n```" - } - ] - }, - { - "id": 3218, - "data_source": "Mono3DRefer", - "scene_name": "005170", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[434.49, 316.75], [248.36, 314.28], [398.47, 264.68], [515.99, 265.65], [434.49, 116.99], [248.36, 118.17], [398.47, 141.83], [515.99, 141.37]]}, {\"category\": \"van\", \"corners_3d\": [[542.77, 245.98], [452.61, 245.66], [498.32, 228.99], [565.14, 229.17], [542.77, 139.76], [452.61, 139.97], [498.32, 150.76], [565.14, 150.65]]}]\n```" - } - ] - }, - { - "id": 3219, - "data_source": "Mono3DRefer", - "scene_name": "005170", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[899.21, 427.61], [1028.38, 418.94], [1196.6, 490.51], [1034.72, 505.35], [899.21, 133.32], [1028.38, 135.02], [1196.6, 121.0], [1034.72, 118.1]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[506.14, 324.08], [423.6, 320.86], [461.65, 303.99], [534.77, 306.47], [506.14, 146.68], [423.6, 147.46], [461.65, 151.59], [534.77, 150.98]]}, {\"category\": \"car\", \"corners_3d\": [[434.49, 316.75], [248.36, 314.28], [398.47, 264.68], [515.99, 265.65], [434.49, 116.99], [248.36, 118.17], [398.47, 141.83], [515.99, 141.37]]}]\n```" - } - ] - }, - { - "id": 3220, - "data_source": "Mono3DRefer", - "scene_name": "005170", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[899.21, 427.61], [1028.38, 418.94], [1196.6, 490.51], [1034.72, 505.35], [899.21, 133.32], [1028.38, 135.02], [1196.6, 121.0], [1034.72, 118.1]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[506.14, 324.08], [423.6, 320.86], [461.65, 303.99], [534.77, 306.47], [506.14, 146.68], [423.6, 147.46], [461.65, 151.59], [534.77, 150.98]]}, {\"category\": \"van\", \"corners_3d\": [[542.77, 245.98], [452.61, 245.66], [498.32, 228.99], [565.14, 229.17], [542.77, 139.76], [452.61, 139.97], [498.32, 150.76], [565.14, 150.65]]}]\n```" - } - ] - }, - { - "id": 3221, - "data_source": "Mono3DRefer", - "scene_name": "005170", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[434.49, 316.75], [248.36, 314.28], [398.47, 264.68], [515.99, 265.65], [434.49, 116.99], [248.36, 118.17], [398.47, 141.83], [515.99, 141.37]]}, {\"category\": \"van\", \"corners_3d\": [[542.77, 245.98], [452.61, 245.66], [498.32, 228.99], [565.14, 229.17], [542.77, 139.76], [452.61, 139.97], [498.32, 150.76], [565.14, 150.65]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1244.59, 370.2], [1145.33, 366.54], [1101.59, 347.7], [1189.79, 350.65], [1244.59, 190.96], [1145.33, 190.78], [1101.59, 189.83], [1189.79, 189.98]]}]\n```" - } - ] - }, - { - "id": 3222, - "data_source": "Mono3DRefer", - "scene_name": "005170", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[899.21, 427.61], [1028.38, 418.94], [1196.6, 490.51], [1034.72, 505.35], [899.21, 133.32], [1028.38, 135.02], [1196.6, 121.0], [1034.72, 118.1]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[506.14, 324.08], [423.6, 320.86], [461.65, 303.99], [534.77, 306.47], [506.14, 146.68], [423.6, 147.46], [461.65, 151.59], [534.77, 150.98]]}, {\"category\": \"car\", \"corners_3d\": [[434.49, 316.75], [248.36, 314.28], [398.47, 264.68], [515.99, 265.65], [434.49, 116.99], [248.36, 118.17], [398.47, 141.83], [515.99, 141.37]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1244.59, 370.2], [1145.33, 366.54], [1101.59, 347.7], [1189.79, 350.65], [1244.59, 190.96], [1145.33, 190.78], [1101.59, 189.83], [1189.79, 189.98]]}]\n```" - } - ] - }, - { - "id": 3223, - "data_source": "Mono3DRefer", - "scene_name": "005170", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[899.21, 427.61], [1028.38, 418.94], [1196.6, 490.51], [1034.72, 505.35], [899.21, 133.32], [1028.38, 135.02], [1196.6, 121.0], [1034.72, 118.1]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[506.14, 324.08], [423.6, 320.86], [461.65, 303.99], [534.77, 306.47], [506.14, 146.68], [423.6, 147.46], [461.65, 151.59], [534.77, 150.98]]}, {\"category\": \"van\", \"corners_3d\": [[542.77, 245.98], [452.61, 245.66], [498.32, 228.99], [565.14, 229.17], [542.77, 139.76], [452.61, 139.97], [498.32, 150.76], [565.14, 150.65]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1244.59, 370.2], [1145.33, 366.54], [1101.59, 347.7], [1189.79, 350.65], [1244.59, 190.96], [1145.33, 190.78], [1101.59, 189.83], [1189.79, 189.98]]}]\n```" - } - ] - }, - { - "id": 3224, - "data_source": "Mono3DRefer", - "scene_name": "005170", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[899.21, 427.61], [1028.38, 418.94], [1196.6, 490.51], [1034.72, 505.35], [899.21, 133.32], [1028.38, 135.02], [1196.6, 121.0], [1034.72, 118.1]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[506.14, 324.08], [423.6, 320.86], [461.65, 303.99], [534.77, 306.47], [506.14, 146.68], [423.6, 147.46], [461.65, 151.59], [534.77, 150.98]]}, {\"category\": \"car\", \"corners_3d\": [[434.49, 316.75], [248.36, 314.28], [398.47, 264.68], [515.99, 265.65], [434.49, 116.99], [248.36, 118.17], [398.47, 141.83], [515.99, 141.37]]}, {\"category\": \"van\", \"corners_3d\": [[542.77, 245.98], [452.61, 245.66], [498.32, 228.99], [565.14, 229.17], [542.77, 139.76], [452.61, 139.97], [498.32, 150.76], [565.14, 150.65]]}]\n```" - } - ] - }, - { - "id": 3225, - "data_source": "Mono3DRefer", - "scene_name": "005170", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, car, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[899.21, 427.61], [1028.38, 418.94], [1196.6, 490.51], [1034.72, 505.35], [899.21, 133.32], [1028.38, 135.02], [1196.6, 121.0], [1034.72, 118.1]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[506.14, 324.08], [423.6, 320.86], [461.65, 303.99], [534.77, 306.47], [506.14, 146.68], [423.6, 147.46], [461.65, 151.59], [534.77, 150.98]]}, {\"category\": \"car\", \"corners_3d\": [[434.49, 316.75], [248.36, 314.28], [398.47, 264.68], [515.99, 265.65], [434.49, 116.99], [248.36, 118.17], [398.47, 141.83], [515.99, 141.37]]}, {\"category\": \"van\", \"corners_3d\": [[542.77, 245.98], [452.61, 245.66], [498.32, 228.99], [565.14, 229.17], [542.77, 139.76], [452.61, 139.97], [498.32, 150.76], [565.14, 150.65]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1244.59, 370.2], [1145.33, 366.54], [1101.59, 347.7], [1189.79, 350.65], [1244.59, 190.96], [1145.33, 190.78], [1101.59, 189.83], [1189.79, 189.98]]}]\n```" - } - ] - }, - { - "id": 3226, - "data_source": "Mono3DRefer", - "scene_name": "005177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[741.3, 223.61], [799.55, 224.25], [815.07, 235.18], [744.38, 234.23], [741.3, 170.1], [799.55, 170.06], [815.07, 169.47], [744.38, 169.52]]}, {\"category\": \"car\", \"corners_3d\": [[-167.27, 516.28], [-389.81, 490.73], [150.57, 339.73], [285.06, 346.51], [-167.27, 172.78], [-389.81, 172.79], [150.57, 172.82], [285.06, 172.82]]}, {\"category\": \"car\", \"corners_3d\": [[-215.08, 371.75], [-349.97, 363.85], [56.43, 292.17], [149.74, 295.2], [-215.08, 201.23], [-349.97, 200.11], [56.43, 189.88], [149.74, 190.31]]}, {\"category\": \"car\", \"corners_3d\": [[437.12, 238.08], [382.15, 236.43], [464.22, 224.83], [510.69, 225.93], [437.12, 183.9], [382.15, 183.62], [464.22, 181.66], [510.69, 181.85]]}, {\"category\": \"car\", \"corners_3d\": [[1001.17, 188.89], [1021.05, 189.61], [949.06, 189.73], [932.16, 189.0], [1001.17, 159.48], [1021.05, 158.88], [949.06, 158.78], [932.16, 159.39]]}]\n```" - } - ] - }, - { - "id": 3227, - "data_source": "Mono3DRefer", - "scene_name": "005184", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[540.46, 245.26], [472.09, 245.27], [499.4, 230.82], [554.13, 230.82], [540.46, 182.2], [472.09, 182.2], [499.4, 180.33], [554.13, 180.33]]}, {\"category\": \"car\", \"corners_3d\": [[668.2, 219.92], [705.36, 219.85], [724.56, 227.57], [681.32, 227.67], [668.2, 180.9], [705.36, 180.88], [724.56, 182.2], [681.32, 182.22]]}]\n```" - } - ] - }, - { - "id": 3228, - "data_source": "Mono3DRefer", - "scene_name": "005184", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[795.82, 332.39], [952.96, 331.72], [1403.43, 532.03], [1050.58, 535.48], [795.82, 166.13], [952.96, 166.16], [1403.43, 157.72], [1050.58, 157.57]]}, {\"category\": \"van\", \"corners_3d\": [[686.87, 230.91], [743.63, 230.78], [784.06, 245.87], [712.58, 246.09], [686.87, 166.03], [743.63, 166.05], [784.06, 164.27], [712.58, 164.25]]}]\n```" - } - ] - }, - { - "id": 3229, - "data_source": "Mono3DRefer", - "scene_name": "005184", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[795.82, 332.39], [952.96, 331.72], [1403.43, 532.03], [1050.58, 535.48], [795.82, 166.13], [952.96, 166.16], [1403.43, 157.72], [1050.58, 157.57]]}, {\"category\": \"car\", \"corners_3d\": [[540.46, 245.26], [472.09, 245.27], [499.4, 230.82], [554.13, 230.82], [540.46, 182.2], [472.09, 182.2], [499.4, 180.33], [554.13, 180.33]]}, {\"category\": \"van\", \"corners_3d\": [[686.87, 230.91], [743.63, 230.78], [784.06, 245.87], [712.58, 246.09], [686.87, 166.03], [743.63, 166.05], [784.06, 164.27], [712.58, 164.25]]}, {\"category\": \"car\", \"corners_3d\": [[668.2, 219.92], [705.36, 219.85], [724.56, 227.57], [681.32, 227.67], [668.2, 180.9], [705.36, 180.88], [724.56, 182.2], [681.32, 182.22]]}]\n```" - } - ] - }, - { - "id": 3230, - "data_source": "Mono3DRefer", - "scene_name": "005185", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[727.05, 222.0], [768.05, 222.31], [783.29, 231.95], [734.25, 231.51], [727.05, 182.26], [768.05, 182.32], [783.29, 184.16], [734.25, 184.08]]}, {\"category\": \"car\", \"corners_3d\": [[298.64, 263.15], [335.56, 255.88], [456.33, 258.96], [428.92, 266.81], [298.64, 192.41], [335.56, 190.84], [456.33, 191.51], [428.92, 193.21]]}, {\"category\": \"car\", \"corners_3d\": [[752.07, 305.85], [883.01, 308.48], [995.45, 387.42], [787.2, 380.92], [752.07, 185.22], [883.01, 185.46], [995.45, 192.8], [787.2, 192.2]]}]\n```" - } - ] - }, - { - "id": 3231, - "data_source": "Mono3DRefer", - "scene_name": "005185", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[715.83, 206.54], [745.71, 206.71], [751.09, 210.46], [717.89, 210.25], [715.83, 172.52], [745.71, 172.52], [751.09, 172.48], [717.89, 172.48]]}]\n```" - } - ] - }, - { - "id": 3232, - "data_source": "Mono3DRefer", - "scene_name": "005185", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[733.49, 341.64], [768.49, 335.46], [799.43, 343.31], [763.91, 350.11], [733.49, 175.78], [768.49, 175.67], [799.43, 175.81], [763.91, 175.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[386.71, 272.03], [409.76, 269.99], [420.41, 272.69], [396.93, 274.85], [386.71, 193.79], [409.76, 193.36], [420.41, 193.93], [396.93, 194.39]]}]\n```" - } - ] - }, - { - "id": 3233, - "data_source": "Mono3DRefer", - "scene_name": "005185", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[727.05, 222.0], [768.05, 222.31], [783.29, 231.95], [734.25, 231.51], [727.05, 182.26], [768.05, 182.32], [783.29, 184.16], [734.25, 184.08]]}, {\"category\": \"car\", \"corners_3d\": [[298.64, 263.15], [335.56, 255.88], [456.33, 258.96], [428.92, 266.81], [298.64, 192.41], [335.56, 190.84], [456.33, 191.51], [428.92, 193.21]]}, {\"category\": \"van\", \"corners_3d\": [[715.83, 206.54], [745.71, 206.71], [751.09, 210.46], [717.89, 210.25], [715.83, 172.52], [745.71, 172.52], [751.09, 172.48], [717.89, 172.48]]}, {\"category\": \"car\", \"corners_3d\": [[752.07, 305.85], [883.01, 308.48], [995.45, 387.42], [787.2, 380.92], [752.07, 185.22], [883.01, 185.46], [995.45, 192.8], [787.2, 192.2]]}]\n```" - } - ] - }, - { - "id": 3234, - "data_source": "Mono3DRefer", - "scene_name": "005185", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[733.49, 341.64], [768.49, 335.46], [799.43, 343.31], [763.91, 350.11], [733.49, 175.78], [768.49, 175.67], [799.43, 175.81], [763.91, 175.93]]}, {\"category\": \"car\", \"corners_3d\": [[727.05, 222.0], [768.05, 222.31], [783.29, 231.95], [734.25, 231.51], [727.05, 182.26], [768.05, 182.32], [783.29, 184.16], [734.25, 184.08]]}, {\"category\": \"car\", \"corners_3d\": [[298.64, 263.15], [335.56, 255.88], [456.33, 258.96], [428.92, 266.81], [298.64, 192.41], [335.56, 190.84], [456.33, 191.51], [428.92, 193.21]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[386.71, 272.03], [409.76, 269.99], [420.41, 272.69], [396.93, 274.85], [386.71, 193.79], [409.76, 193.36], [420.41, 193.93], [396.93, 194.39]]}, {\"category\": \"car\", \"corners_3d\": [[752.07, 305.85], [883.01, 308.48], [995.45, 387.42], [787.2, 380.92], [752.07, 185.22], [883.01, 185.46], [995.45, 192.8], [787.2, 192.2]]}]\n```" - } - ] - }, - { - "id": 3235, - "data_source": "Mono3DRefer", - "scene_name": "005185", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[733.49, 341.64], [768.49, 335.46], [799.43, 343.31], [763.91, 350.11], [733.49, 175.78], [768.49, 175.67], [799.43, 175.81], [763.91, 175.93]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[386.71, 272.03], [409.76, 269.99], [420.41, 272.69], [396.93, 274.85], [386.71, 193.79], [409.76, 193.36], [420.41, 193.93], [396.93, 194.39]]}, {\"category\": \"van\", \"corners_3d\": [[715.83, 206.54], [745.71, 206.71], [751.09, 210.46], [717.89, 210.25], [715.83, 172.52], [745.71, 172.52], [751.09, 172.48], [717.89, 172.48]]}]\n```" - } - ] - }, - { - "id": 3236, - "data_source": "Mono3DRefer", - "scene_name": "005185", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[733.49, 341.64], [768.49, 335.46], [799.43, 343.31], [763.91, 350.11], [733.49, 175.78], [768.49, 175.67], [799.43, 175.81], [763.91, 175.93]]}, {\"category\": \"car\", \"corners_3d\": [[727.05, 222.0], [768.05, 222.31], [783.29, 231.95], [734.25, 231.51], [727.05, 182.26], [768.05, 182.32], [783.29, 184.16], [734.25, 184.08]]}, {\"category\": \"car\", \"corners_3d\": [[298.64, 263.15], [335.56, 255.88], [456.33, 258.96], [428.92, 266.81], [298.64, 192.41], [335.56, 190.84], [456.33, 191.51], [428.92, 193.21]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[386.71, 272.03], [409.76, 269.99], [420.41, 272.69], [396.93, 274.85], [386.71, 193.79], [409.76, 193.36], [420.41, 193.93], [396.93, 194.39]]}, {\"category\": \"van\", \"corners_3d\": [[715.83, 206.54], [745.71, 206.71], [751.09, 210.46], [717.89, 210.25], [715.83, 172.52], [745.71, 172.52], [751.09, 172.48], [717.89, 172.48]]}, {\"category\": \"car\", \"corners_3d\": [[752.07, 305.85], [883.01, 308.48], [995.45, 387.42], [787.2, 380.92], [752.07, 185.22], [883.01, 185.46], [995.45, 192.8], [787.2, 192.2]]}]\n```" - } - ] - }, - { - "id": 3237, - "data_source": "Mono3DRefer", - "scene_name": "005190", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[637.05, 204.63], [664.59, 204.6], [671.78, 207.59], [641.64, 207.64], [637.05, 179.45], [664.59, 179.44], [671.78, 180.06], [641.64, 180.07]]}, {\"category\": \"car\", \"corners_3d\": [[654.12, 217.71], [694.79, 217.68], [707.99, 224.12], [661.48, 224.15], [654.12, 180.16], [694.79, 180.16], [707.99, 181.2], [661.48, 181.21]]}, {\"category\": \"car\", \"corners_3d\": [[676.34, 240.05], [740.61, 239.75], [789.71, 259.47], [706.67, 259.96], [676.34, 173.22], [740.61, 173.22], [789.71, 173.33], [706.67, 173.33]]}, {\"category\": \"car\", \"corners_3d\": [[56.05, 285.52], [131.73, 271.26], [335.45, 273.05], [288.33, 287.86], [56.05, 184.76], [131.73, 183.25], [335.45, 183.44], [288.33, 185.01]]}, {\"category\": \"car\", \"corners_3d\": [[405.54, 256.55], [381.77, 267.46], [129.94, 268.44], [183.23, 257.32], [405.54, 182.57], [381.77, 183.84], [129.94, 183.95], [183.23, 182.66]]}, {\"category\": \"car\", \"corners_3d\": [[222.33, 256.04], [261.31, 248.27], [421.59, 248.98], [398.89, 256.91], [222.33, 187.17], [261.31, 185.83], [421.59, 185.95], [398.89, 187.32]]}, {\"category\": \"car\", \"corners_3d\": [[486.53, 228.02], [477.58, 232.03], [331.37, 232.03], [350.23, 228.02], [486.53, 169.86], [477.58, 169.64], [331.37, 169.64], [350.23, 169.86]]}, {\"category\": \"car\", \"corners_3d\": [[323.21, 237.1], [343.94, 232.46], [452.93, 232.48], [440.67, 237.11], [323.21, 178.17], [343.94, 177.78], [452.93, 177.78], [440.67, 178.17]]}, {\"category\": \"car\", \"corners_3d\": [[664.25, 224.95], [714.02, 224.74], [737.13, 232.88], [679.62, 233.17], [664.25, 181.33], [714.02, 181.29], [737.13, 182.62], [679.62, 182.66]]}, {\"category\": \"car\", \"corners_3d\": [[362.5, 224.59], [376.96, 221.46], [473.04, 221.38], [464.78, 224.51], [362.5, 170.53], [376.96, 170.67], [473.04, 170.67], [464.78, 170.53]]}, {\"category\": \"car\", \"corners_3d\": [[389.82, 217.28], [402.57, 214.94], [501.66, 215.13], [494.37, 217.48], [389.82, 175.52], [402.57, 175.38], [501.66, 175.39], [494.37, 175.53]]}]\n```" - } - ] - }, - { - "id": 3238, - "data_source": "Mono3DRefer", - "scene_name": "005199", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[557.9, 205.03], [524.06, 205.08], [530.49, 201.81], [560.88, 201.77], [557.9, 176.89], [524.06, 176.9], [530.49, 176.49], [560.88, 176.48]]}, {\"category\": \"car\", \"corners_3d\": [[454.5, 267.48], [375.5, 267.8], [421.36, 247.24], [483.18, 247.05], [454.5, 193.98], [375.5, 194.06], [421.36, 189.47], [483.18, 189.42]]}]\n```" - } - ] - }, - { - "id": 3239, - "data_source": "Mono3DRefer", - "scene_name": "005199", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[502.26, 237.41], [440.86, 237.47], [468.49, 226.37], [519.33, 226.33], [502.26, 170.63], [440.86, 170.63], [468.49, 171.01], [519.33, 171.01]]}]\n```" - } - ] - }, - { - "id": 3240, - "data_source": "Mono3DRefer", - "scene_name": "005199", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[502.26, 237.41], [440.86, 237.47], [468.49, 226.37], [519.33, 226.33], [502.26, 170.63], [440.86, 170.63], [468.49, 171.01], [519.33, 171.01]]}, {\"category\": \"car\", \"corners_3d\": [[557.9, 205.03], [524.06, 205.08], [530.49, 201.81], [560.88, 201.77], [557.9, 176.89], [524.06, 176.9], [530.49, 176.49], [560.88, 176.48]]}, {\"category\": \"car\", \"corners_3d\": [[454.5, 267.48], [375.5, 267.8], [421.36, 247.24], [483.18, 247.05], [454.5, 193.98], [375.5, 194.06], [421.36, 189.47], [483.18, 189.42]]}]\n```" - } - ] - }, - { - "id": 3241, - "data_source": "Mono3DRefer", - "scene_name": "005213", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[689.51, 236.81], [702.04, 235.6], [760.02, 240.2], [747.74, 241.59], [689.51, 163.94], [702.04, 164.11], [760.02, 163.47], [747.74, 163.28]]}, {\"category\": \"cyclist\", \"corners_3d\": [[688.38, 226.93], [697.19, 225.87], [747.67, 228.36], [739.49, 229.53], [688.38, 163.51], [697.19, 163.7], [747.67, 163.27], [739.49, 163.07]]}]\n```" - } - ] - }, - { - "id": 3242, - "data_source": "Mono3DRefer", - "scene_name": "005213", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[700.29, 235.17], [697.88, 233.85], [713.54, 233.81], [716.29, 235.13], [700.29, 170.83], [697.88, 170.87], [713.54, 170.87], [716.29, 170.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[682.14, 230.21], [672.98, 229.25], [682.85, 228.71], [692.08, 229.66], [682.14, 175.05], [672.98, 175.01], [682.85, 174.99], [692.08, 175.03]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[387.86, 242.5], [383.35, 244.02], [370.85, 244.04], [375.64, 242.52], [387.86, 177.7], [383.35, 177.81], [370.85, 177.81], [375.64, 177.7]]}]\n```" - } - ] - }, - { - "id": 3243, - "data_source": "Mono3DRefer", - "scene_name": "005213", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[591.53, 201.9], [616.4, 201.88], [618.09, 204.24], [591.19, 204.26], [591.53, 177.32], [616.4, 177.32], [618.09, 177.68], [591.19, 177.68]]}, {\"category\": \"car\", \"corners_3d\": [[420.46, 267.64], [503.55, 267.76], [443.83, 317.56], [317.33, 317.28], [420.46, 178.9], [503.55, 178.9], [443.83, 182.08], [317.33, 182.06]]}, {\"category\": \"car\", \"corners_3d\": [[56.09, 281.87], [-28.4, 282.13], [94.02, 260.65], [161.7, 260.49], [56.09, 204.2], [-28.4, 204.27], [94.02, 198.1], [161.7, 198.05]]}]\n```" - } - ] - }, - { - "id": 3244, - "data_source": "Mono3DRefer", - "scene_name": "005213", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[689.51, 236.81], [702.04, 235.6], [760.02, 240.2], [747.74, 241.59], [689.51, 163.94], [702.04, 164.11], [760.02, 163.47], [747.74, 163.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[700.29, 235.17], [697.88, 233.85], [713.54, 233.81], [716.29, 235.13], [700.29, 170.83], [697.88, 170.87], [713.54, 170.87], [716.29, 170.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[682.14, 230.21], [672.98, 229.25], [682.85, 228.71], [692.08, 229.66], [682.14, 175.05], [672.98, 175.01], [682.85, 174.99], [692.08, 175.03]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[387.86, 242.5], [383.35, 244.02], [370.85, 244.04], [375.64, 242.52], [387.86, 177.7], [383.35, 177.81], [370.85, 177.81], [375.64, 177.7]]}, {\"category\": \"cyclist\", \"corners_3d\": [[688.38, 226.93], [697.19, 225.87], [747.67, 228.36], [739.49, 229.53], [688.38, 163.51], [697.19, 163.7], [747.67, 163.27], [739.49, 163.07]]}]\n```" - } - ] - }, - { - "id": 3245, - "data_source": "Mono3DRefer", - "scene_name": "005213", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[591.53, 201.9], [616.4, 201.88], [618.09, 204.24], [591.19, 204.26], [591.53, 177.32], [616.4, 177.32], [618.09, 177.68], [591.19, 177.68]]}, {\"category\": \"car\", \"corners_3d\": [[420.46, 267.64], [503.55, 267.76], [443.83, 317.56], [317.33, 317.28], [420.46, 178.9], [503.55, 178.9], [443.83, 182.08], [317.33, 182.06]]}, {\"category\": \"car\", \"corners_3d\": [[56.09, 281.87], [-28.4, 282.13], [94.02, 260.65], [161.7, 260.49], [56.09, 204.2], [-28.4, 204.27], [94.02, 198.1], [161.7, 198.05]]}, {\"category\": \"cyclist\", \"corners_3d\": [[689.51, 236.81], [702.04, 235.6], [760.02, 240.2], [747.74, 241.59], [689.51, 163.94], [702.04, 164.11], [760.02, 163.47], [747.74, 163.28]]}, {\"category\": \"cyclist\", \"corners_3d\": [[688.38, 226.93], [697.19, 225.87], [747.67, 228.36], [739.49, 229.53], [688.38, 163.51], [697.19, 163.7], [747.67, 163.27], [739.49, 163.07]]}]\n```" - } - ] - }, - { - "id": 3246, - "data_source": "Mono3DRefer", - "scene_name": "005213", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[591.53, 201.9], [616.4, 201.88], [618.09, 204.24], [591.19, 204.26], [591.53, 177.32], [616.4, 177.32], [618.09, 177.68], [591.19, 177.68]]}, {\"category\": \"car\", \"corners_3d\": [[420.46, 267.64], [503.55, 267.76], [443.83, 317.56], [317.33, 317.28], [420.46, 178.9], [503.55, 178.9], [443.83, 182.08], [317.33, 182.06]]}, {\"category\": \"car\", \"corners_3d\": [[56.09, 281.87], [-28.4, 282.13], [94.02, 260.65], [161.7, 260.49], [56.09, 204.2], [-28.4, 204.27], [94.02, 198.1], [161.7, 198.05]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[700.29, 235.17], [697.88, 233.85], [713.54, 233.81], [716.29, 235.13], [700.29, 170.83], [697.88, 170.87], [713.54, 170.87], [716.29, 170.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[682.14, 230.21], [672.98, 229.25], [682.85, 228.71], [692.08, 229.66], [682.14, 175.05], [672.98, 175.01], [682.85, 174.99], [692.08, 175.03]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[387.86, 242.5], [383.35, 244.02], [370.85, 244.04], [375.64, 242.52], [387.86, 177.7], [383.35, 177.81], [370.85, 177.81], [375.64, 177.7]]}]\n```" - } - ] - }, - { - "id": 3247, - "data_source": "Mono3DRefer", - "scene_name": "005213", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[591.53, 201.9], [616.4, 201.88], [618.09, 204.24], [591.19, 204.26], [591.53, 177.32], [616.4, 177.32], [618.09, 177.68], [591.19, 177.68]]}, {\"category\": \"car\", \"corners_3d\": [[420.46, 267.64], [503.55, 267.76], [443.83, 317.56], [317.33, 317.28], [420.46, 178.9], [503.55, 178.9], [443.83, 182.08], [317.33, 182.06]]}, {\"category\": \"car\", \"corners_3d\": [[56.09, 281.87], [-28.4, 282.13], [94.02, 260.65], [161.7, 260.49], [56.09, 204.2], [-28.4, 204.27], [94.02, 198.1], [161.7, 198.05]]}, {\"category\": \"cyclist\", \"corners_3d\": [[689.51, 236.81], [702.04, 235.6], [760.02, 240.2], [747.74, 241.59], [689.51, 163.94], [702.04, 164.11], [760.02, 163.47], [747.74, 163.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[700.29, 235.17], [697.88, 233.85], [713.54, 233.81], [716.29, 235.13], [700.29, 170.83], [697.88, 170.87], [713.54, 170.87], [716.29, 170.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[682.14, 230.21], [672.98, 229.25], [682.85, 228.71], [692.08, 229.66], [682.14, 175.05], [672.98, 175.01], [682.85, 174.99], [692.08, 175.03]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[387.86, 242.5], [383.35, 244.02], [370.85, 244.04], [375.64, 242.52], [387.86, 177.7], [383.35, 177.81], [370.85, 177.81], [375.64, 177.7]]}, {\"category\": \"cyclist\", \"corners_3d\": [[688.38, 226.93], [697.19, 225.87], [747.67, 228.36], [739.49, 229.53], [688.38, 163.51], [697.19, 163.7], [747.67, 163.27], [739.49, 163.07]]}]\n```" - } - ] - }, - { - "id": 3248, - "data_source": "Mono3DRefer", - "scene_name": "005223", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[468.0, 323.96], [311.34, 327.22], [377.11, 282.15], [487.73, 280.5], [468.0, 181.46], [311.34, 181.65], [377.11, 179.08], [487.73, 178.98]]}, {\"category\": \"car\", \"corners_3d\": [[511.95, 257.9], [439.96, 257.9], [474.81, 240.37], [531.96, 240.36], [511.95, 184.52], [439.96, 184.52], [474.81, 182.12], [531.96, 182.12]]}]\n```" - } - ] - }, - { - "id": 3249, - "data_source": "Mono3DRefer", - "scene_name": "005224", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[314.66, 258.01], [275.32, 267.91], [61.29, 266.17], [121.91, 256.61], [314.66, 197.04], [275.32, 199.86], [61.29, 199.36], [121.91, 196.65]]}]\n```" - } - ] - }, - { - "id": 3250, - "data_source": "Mono3DRefer", - "scene_name": "005225", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[565.2, 206.68], [537.72, 206.68], [543.86, 203.76], [568.97, 203.76], [565.2, 182.54], [537.72, 182.54], [543.86, 181.7], [568.97, 181.7]]}]\n```" - } - ] - }, - { - "id": 3251, - "data_source": "Mono3DRefer", - "scene_name": "005231", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[885.24, 319.24], [865.17, 309.3], [955.06, 309.32], [982.08, 319.27], [885.24, 144.89], [865.17, 147.45], [955.06, 147.44], [982.08, 144.89]]}]\n```" - } - ] - }, - { - "id": 3252, - "data_source": "Mono3DRefer", - "scene_name": "005250", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[536.7, 267.86], [632.29, 267.75], [644.75, 308.03], [508.55, 308.26], [536.7, 177.53], [632.29, 177.52], [644.75, 179.5], [508.55, 179.51]]}, {\"category\": \"car\", \"corners_3d\": [[573.96, 233.21], [637.73, 233.21], [645.78, 250.83], [563.41, 250.82], [573.96, 166.43], [637.73, 166.43], [645.78, 164.55], [563.41, 164.55]]}, {\"category\": \"car\", \"corners_3d\": [[745.62, 275.29], [844.31, 275.3], [945.84, 319.72], [804.35, 319.7], [745.62, 174.1], [844.31, 174.1], [945.84, 174.64], [804.35, 174.63]]}, {\"category\": \"car\", \"corners_3d\": [[680.35, 240.51], [742.98, 240.28], [791.77, 260.62], [710.39, 261.02], [680.35, 177.79], [742.98, 177.77], [791.77, 179.26], [710.39, 179.29]]}, {\"category\": \"car\", \"corners_3d\": [[653.0, 221.95], [700.25, 221.95], [719.78, 232.59], [662.28, 232.58], [653.0, 175.0], [700.25, 175.0], [719.78, 175.47], [662.28, 175.47]]}, {\"category\": \"car\", \"corners_3d\": [[147.0, 213.15], [133.45, 214.92], [64.44, 215.67], [81.18, 213.84], [147.0, 184.29], [133.45, 184.79], [64.44, 185.01], [81.18, 184.49]]}, {\"category\": \"car\", \"corners_3d\": [[401.09, 201.39], [401.36, 202.5], [341.97, 203.17], [343.99, 202.01], [401.09, 177.91], [401.36, 178.11], [341.97, 178.23], [343.99, 178.02]]}, {\"category\": \"car\", \"corners_3d\": [[307.94, 205.85], [303.47, 207.2], [242.71, 207.89], [249.68, 206.49], [307.94, 182.01], [303.47, 182.38], [242.71, 182.57], [249.68, 182.18]]}, {\"category\": \"car\", \"corners_3d\": [[642.26, 207.93], [678.17, 208.02], [681.92, 212.15], [641.79, 212.04], [642.26, 175.18], [678.17, 175.19], [681.92, 175.46], [641.79, 175.46]]}]\n```" - } - ] - }, - { - "id": 3253, - "data_source": "Mono3DRefer", - "scene_name": "005252", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[564.26, 190.87], [548.01, 190.88], [550.58, 189.88], [565.93, 189.88], [564.26, 175.5], [548.01, 175.5], [550.58, 175.36], [565.93, 175.35]]}, {\"category\": \"car\", \"corners_3d\": [[547.85, 202.08], [522.38, 202.08], [529.53, 199.67], [552.9, 199.66], [547.85, 175.13], [522.38, 175.13], [529.53, 174.95], [552.9, 174.95]]}]\n```" - } - ] - }, - { - "id": 3254, - "data_source": "Mono3DRefer", - "scene_name": "005257", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[242.49, 448.8], [21.37, 447.16], [274.21, 331.96], [403.02, 332.51], [242.49, 244.92], [21.37, 244.49], [274.21, 214.41], [403.02, 214.55]]}, {\"category\": \"car\", \"corners_3d\": [[521.55, 247.83], [469.6, 247.73], [493.09, 236.3], [537.12, 236.37], [521.55, 189.88], [469.6, 189.86], [493.09, 187.26], [537.12, 187.28]]}, {\"category\": \"car\", \"corners_3d\": [[559.64, 221.64], [525.16, 221.69], [531.78, 217.03], [562.96, 216.99], [559.64, 187.91], [525.16, 187.93], [531.78, 186.49], [562.96, 186.48]]}, {\"category\": \"car\", \"corners_3d\": [[581.12, 209.55], [558.38, 209.21], [576.06, 206.72], [597.39, 207.01], [581.12, 185.39], [558.38, 185.28], [576.06, 184.42], [597.39, 184.52]]}]\n```" - } - ] - }, - { - "id": 3255, - "data_source": "Mono3DRefer", - "scene_name": "005269", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[668.88, 238.03], [729.88, 237.98], [760.52, 253.7], [684.82, 253.78], [668.88, 185.51], [729.88, 185.5], [760.52, 188.55], [684.82, 188.56]]}, {\"category\": \"car\", \"corners_3d\": [[528.03, 212.38], [496.74, 212.36], [507.39, 208.84], [535.9, 208.85], [528.03, 186.92], [496.74, 186.91], [507.39, 185.66], [535.9, 185.66]]}, {\"category\": \"car\", \"corners_3d\": [[445.73, 248.48], [387.07, 248.43], [425.16, 235.86], [474.09, 235.9], [445.73, 194.04], [387.07, 194.02], [425.16, 190.5], [474.09, 190.51]]}, {\"category\": \"car\", \"corners_3d\": [[477.12, 234.03], [429.49, 233.87], [457.64, 225.63], [498.88, 225.75], [477.12, 191.11], [429.49, 191.07], [457.64, 188.6], [498.88, 188.64]]}, {\"category\": \"car\", \"corners_3d\": [[312.13, 297.55], [214.1, 297.39], [321.64, 264.09], [393.53, 264.17], [312.13, 215.86], [214.1, 215.81], [321.64, 204.32], [393.53, 204.35]]}]\n```" - } - ] - }, - { - "id": 3256, - "data_source": "Mono3DRefer", - "scene_name": "005300", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[982.61, 278.55], [1070.54, 295.93], [789.65, 308.43], [737.27, 287.63], [982.61, 184.79], [1070.54, 186.75], [789.65, 188.16], [737.27, 185.81]]}, {\"category\": \"car\", \"corners_3d\": [[650.22, 266.41], [624.48, 256.44], [807.88, 250.49], [851.73, 259.02], [650.22, 168.05], [624.48, 168.56], [807.88, 168.87], [851.73, 168.43]]}, {\"category\": \"car\", \"corners_3d\": [[-561.08, 406.57], [-360.17, 359.64], [128.61, 336.98], [62.01, 372.15], [-561.08, 222.73], [-360.17, 212.71], [128.61, 207.88], [62.01, 215.38]]}, {\"category\": \"car\", \"corners_3d\": [[9.12, 275.46], [53.9, 264.96], [256.12, 260.01], [235.67, 269.35], [9.12, 198.22], [53.9, 195.63], [256.12, 194.4], [235.67, 196.71]]}, {\"category\": \"car\", \"corners_3d\": [[163.41, 244.57], [181.59, 239.42], [306.19, 236.77], [298.02, 241.5], [163.41, 193.9], [181.59, 192.39], [306.19, 191.61], [298.02, 193.0]]}, {\"category\": \"car\", \"corners_3d\": [[275.21, 219.85], [282.54, 217.76], [363.88, 216.68], [360.44, 218.67], [275.21, 188.14], [282.54, 187.46], [363.88, 187.11], [360.44, 187.76]]}]\n```" - } - ] - }, - { - "id": 3257, - "data_source": "Mono3DRefer", - "scene_name": "005300", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[479.24, 218.59], [494.35, 218.41], [496.03, 221.74], [479.81, 221.94], [479.24, 178.06], [494.35, 178.04], [496.03, 178.42], [479.81, 178.44]]}]\n```" - } - ] - }, - { - "id": 3258, - "data_source": "Mono3DRefer", - "scene_name": "005300", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[982.61, 278.55], [1070.54, 295.93], [789.65, 308.43], [737.27, 287.63], [982.61, 184.79], [1070.54, 186.75], [789.65, 188.16], [737.27, 185.81]]}, {\"category\": \"car\", \"corners_3d\": [[650.22, 266.41], [624.48, 256.44], [807.88, 250.49], [851.73, 259.02], [650.22, 168.05], [624.48, 168.56], [807.88, 168.87], [851.73, 168.43]]}, {\"category\": \"car\", \"corners_3d\": [[-561.08, 406.57], [-360.17, 359.64], [128.61, 336.98], [62.01, 372.15], [-561.08, 222.73], [-360.17, 212.71], [128.61, 207.88], [62.01, 215.38]]}, {\"category\": \"cyclist\", \"corners_3d\": [[479.24, 218.59], [494.35, 218.41], [496.03, 221.74], [479.81, 221.94], [479.24, 178.06], [494.35, 178.04], [496.03, 178.42], [479.81, 178.44]]}, {\"category\": \"car\", \"corners_3d\": [[9.12, 275.46], [53.9, 264.96], [256.12, 260.01], [235.67, 269.35], [9.12, 198.22], [53.9, 195.63], [256.12, 194.4], [235.67, 196.71]]}, {\"category\": \"car\", \"corners_3d\": [[163.41, 244.57], [181.59, 239.42], [306.19, 236.77], [298.02, 241.5], [163.41, 193.9], [181.59, 192.39], [306.19, 191.61], [298.02, 193.0]]}, {\"category\": \"car\", \"corners_3d\": [[275.21, 219.85], [282.54, 217.76], [363.88, 216.68], [360.44, 218.67], [275.21, 188.14], [282.54, 187.46], [363.88, 187.11], [360.44, 187.76]]}]\n```" - } - ] - }, - { - "id": 3259, - "data_source": "Mono3DRefer", - "scene_name": "005302", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[585.47, 193.9], [567.89, 193.9], [569.76, 192.74], [586.37, 192.73], [585.47, 179.18], [567.89, 179.19], [569.76, 178.84], [586.37, 178.83]]}]\n```" - } - ] - }, - { - "id": 3260, - "data_source": "Mono3DRefer", - "scene_name": "005305", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[535.52, 202.41], [501.13, 202.34], [514.43, 200.48], [545.78, 200.54], [535.52, 174.64], [501.13, 174.66], [514.43, 175.27], [545.78, 175.25]]}]\n```" - } - ] - }, - { - "id": 3261, - "data_source": "Mono3DRefer", - "scene_name": "005307", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[699.97, 253.13], [772.97, 252.97], [841.85, 284.1], [740.6, 284.4], [699.97, 187.69], [772.97, 187.66], [841.85, 193.41], [740.6, 193.47]]}]\n```" - } - ] - }, - { - "id": 3262, - "data_source": "Mono3DRefer", - "scene_name": "005315", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[802.78, 344.16], [958.71, 343.07], [1415.2, 552.94], [1070.81, 558.42], [802.78, 186.93], [958.71, 186.84], [1415.2, 204.08], [1070.81, 204.53]]}, {\"category\": \"car\", \"corners_3d\": [[342.69, 219.98], [358.71, 217.37], [443.28, 217.53], [432.16, 220.16], [342.69, 176.61], [358.71, 176.41], [443.28, 176.42], [432.16, 176.63]]}, {\"category\": \"car\", \"corners_3d\": [[736.87, 199.57], [705.2, 199.64], [691.87, 196.94], [720.39, 196.89], [736.87, 166.35], [705.2, 166.34], [691.87, 166.99], [720.39, 167.0]]}, {\"category\": \"car\", \"corners_3d\": [[590.5, 203.16], [563.73, 203.2], [565.23, 200.41], [589.54, 200.38], [590.5, 177.15], [563.73, 177.16], [565.23, 176.76], [589.54, 176.76]]}]\n```" - } - ] - }, - { - "id": 3263, - "data_source": "Mono3DRefer", - "scene_name": "005319", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[501.98, 263.19], [413.52, 263.53], [454.88, 241.95], [522.24, 241.75], [501.98, 167.67], [413.52, 167.65], [454.88, 168.89], [522.24, 168.9]]}]\n```" - } - ] - }, - { - "id": 3264, - "data_source": "Mono3DRefer", - "scene_name": "005319", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[441.09, 329.16], [288.89, 330.82], [392.18, 273.36], [488.7, 272.69], [441.09, 195.68], [288.89, 195.92], [392.18, 187.53], [488.7, 187.43]]}, {\"category\": \"car\", \"corners_3d\": [[535.16, 237.31], [467.46, 236.95], [502.13, 224.93], [557.23, 225.17], [535.16, 173.22], [467.46, 173.22], [502.13, 173.15], [557.23, 173.15]]}, {\"category\": \"car\", \"corners_3d\": [[664.61, 202.04], [691.61, 202.57], [673.25, 204.49], [644.9, 203.88], [664.61, 173.21], [691.61, 173.22], [673.25, 173.24], [644.9, 173.24]]}]\n```" - } - ] - }, - { - "id": 3265, - "data_source": "Mono3DRefer", - "scene_name": "005319", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[441.09, 329.16], [288.89, 330.82], [392.18, 273.36], [488.7, 272.69], [441.09, 195.68], [288.89, 195.92], [392.18, 187.53], [488.7, 187.43]]}, {\"category\": \"van\", \"corners_3d\": [[501.98, 263.19], [413.52, 263.53], [454.88, 241.95], [522.24, 241.75], [501.98, 167.67], [413.52, 167.65], [454.88, 168.89], [522.24, 168.9]]}, {\"category\": \"car\", \"corners_3d\": [[535.16, 237.31], [467.46, 236.95], [502.13, 224.93], [557.23, 225.17], [535.16, 173.22], [467.46, 173.22], [502.13, 173.15], [557.23, 173.15]]}, {\"category\": \"car\", \"corners_3d\": [[664.61, 202.04], [691.61, 202.57], [673.25, 204.49], [644.9, 203.88], [664.61, 173.21], [691.61, 173.22], [673.25, 173.24], [644.9, 173.24]]}]\n```" - } - ] - }, - { - "id": 3266, - "data_source": "Mono3DRefer", - "scene_name": "005331", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[546.09, 219.59], [502.83, 219.56], [516.49, 213.93], [554.53, 213.95], [546.09, 184.1], [502.83, 184.1], [516.49, 182.74], [554.53, 182.75]]}]\n```" - } - ] - }, - { - "id": 3267, - "data_source": "Mono3DRefer", - "scene_name": "005340", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[637.11, 211.03], [674.72, 211.17], [676.59, 216.3], [633.96, 216.12], [637.11, 178.13], [674.72, 178.15], [676.59, 178.86], [633.96, 178.83]]}, {\"category\": \"car\", \"corners_3d\": [[592.0, 202.16], [624.28, 202.29], [616.85, 206.04], [580.52, 205.87], [592.0, 171.31], [624.28, 171.31], [616.85, 171.11], [580.52, 171.12]]}, {\"category\": \"car\", \"corners_3d\": [[653.05, 198.64], [681.99, 198.73], [682.82, 201.93], [650.32, 201.81], [653.05, 171.61], [681.99, 171.61], [682.82, 171.45], [650.32, 171.46]]}, {\"category\": \"car\", \"corners_3d\": [[1040.37, 189.38], [1072.41, 190.28], [988.11, 190.79], [959.69, 189.83], [1040.37, 153.92], [1072.41, 152.88], [988.11, 152.3], [959.69, 153.4]]}, {\"category\": \"car\", \"corners_3d\": [[988.47, 189.1], [1014.6, 189.91], [940.99, 190.31], [917.85, 189.46], [988.47, 155.51], [1014.6, 154.65], [940.99, 154.22], [917.85, 155.12]]}]\n```" - } - ] - }, - { - "id": 3268, - "data_source": "Mono3DRefer", - "scene_name": "005351", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[256.64, 260.77], [311.2, 264.99], [128.4, 291.0], [69.92, 284.15], [256.64, 194.82], [311.2, 195.87], [128.4, 202.37], [69.92, 200.66]]}, {\"category\": \"car\", \"corners_3d\": [[164.96, 255.01], [223.65, 258.99], [-8.51, 287.55], [-71.41, 280.6], [164.96, 179.65], [223.65, 179.98], [-8.51, 182.34], [-71.41, 181.77]]}]\n```" - } - ] - }, - { - "id": 3269, - "data_source": "Mono3DRefer", - "scene_name": "005357", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[697.91, 201.34], [722.19, 201.55], [717.67, 203.84], [691.51, 203.58], [697.91, 178.61], [722.19, 178.65], [717.67, 179.11], [691.51, 179.06]]}, {\"category\": \"car\", \"corners_3d\": [[-236.72, 415.76], [-415.54, 407.37], [30.06, 313.36], [145.09, 316.32], [-236.72, 202.83], [-415.54, 201.8], [30.06, 190.19], [145.09, 190.56]]}, {\"category\": \"car\", \"corners_3d\": [[189.47, 305.04], [81.12, 302.94], [253.23, 265.64], [332.12, 266.7], [189.47, 162.49], [81.12, 162.65], [253.23, 165.58], [332.12, 165.49]]}, {\"category\": \"car\", \"corners_3d\": [[728.87, 243.98], [797.11, 244.12], [844.84, 263.83], [757.65, 263.61], [728.87, 179.75], [797.11, 179.76], [844.84, 181.68], [757.65, 181.65]]}, {\"category\": \"car\", \"corners_3d\": [[647.65, 205.77], [617.98, 205.46], [631.44, 202.72], [658.71, 202.98], [647.65, 177.94], [617.98, 177.89], [631.44, 177.47], [658.71, 177.51]]}, {\"category\": \"car\", \"corners_3d\": [[726.67, 230.17], [775.52, 230.52], [796.11, 242.36], [737.16, 241.86], [726.67, 164.52], [775.52, 164.46], [796.11, 162.74], [737.16, 162.82]]}]\n```" - } - ] - }, - { - "id": 3270, - "data_source": "Mono3DRefer", - "scene_name": "005358", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[2.26, 297.62], [78.45, 281.15], [305.0, 279.78], [263.86, 295.8], [2.26, 197.07], [78.45, 193.88], [305.0, 193.61], [263.86, 196.72]]}]\n```" - } - ] - }, - { - "id": 3271, - "data_source": "Mono3DRefer", - "scene_name": "005358", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[644.42, 220.68], [664.24, 220.87], [658.69, 225.73], [636.89, 225.49], [644.42, 166.86], [664.24, 166.84], [658.69, 166.23], [636.89, 166.26]]}]\n```" - } - ] - }, - { - "id": 3272, - "data_source": "Mono3DRefer", - "scene_name": "005358", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[523.39, 204.07], [531.57, 203.52], [549.14, 203.93], [541.16, 204.5], [523.39, 164.23], [531.57, 164.39], [549.14, 164.27], [541.16, 164.11]]}]\n```" - } - ] - }, - { - "id": 3273, - "data_source": "Mono3DRefer", - "scene_name": "005358", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[2.26, 297.62], [78.45, 281.15], [305.0, 279.78], [263.86, 295.8], [2.26, 197.07], [78.45, 193.88], [305.0, 193.61], [263.86, 196.72]]}, {\"category\": \"cyclist\", \"corners_3d\": [[644.42, 220.68], [664.24, 220.87], [658.69, 225.73], [636.89, 225.49], [644.42, 166.86], [664.24, 166.84], [658.69, 166.23], [636.89, 166.26]]}]\n```" - } - ] - }, - { - "id": 3274, - "data_source": "Mono3DRefer", - "scene_name": "005358", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[2.26, 297.62], [78.45, 281.15], [305.0, 279.78], [263.86, 295.8], [2.26, 197.07], [78.45, 193.88], [305.0, 193.61], [263.86, 196.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[523.39, 204.07], [531.57, 203.52], [549.14, 203.93], [541.16, 204.5], [523.39, 164.23], [531.57, 164.39], [549.14, 164.27], [541.16, 164.11]]}]\n```" - } - ] - }, - { - "id": 3275, - "data_source": "Mono3DRefer", - "scene_name": "005358", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[523.39, 204.07], [531.57, 203.52], [549.14, 203.93], [541.16, 204.5], [523.39, 164.23], [531.57, 164.39], [549.14, 164.27], [541.16, 164.11]]}, {\"category\": \"cyclist\", \"corners_3d\": [[644.42, 220.68], [664.24, 220.87], [658.69, 225.73], [636.89, 225.49], [644.42, 166.86], [664.24, 166.84], [658.69, 166.23], [636.89, 166.26]]}]\n```" - } - ] - }, - { - "id": 3276, - "data_source": "Mono3DRefer", - "scene_name": "005358", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[2.26, 297.62], [78.45, 281.15], [305.0, 279.78], [263.86, 295.8], [2.26, 197.07], [78.45, 193.88], [305.0, 193.61], [263.86, 196.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[523.39, 204.07], [531.57, 203.52], [549.14, 203.93], [541.16, 204.5], [523.39, 164.23], [531.57, 164.39], [549.14, 164.27], [541.16, 164.11]]}, {\"category\": \"cyclist\", \"corners_3d\": [[644.42, 220.68], [664.24, 220.87], [658.69, 225.73], [636.89, 225.49], [644.42, 166.86], [664.24, 166.84], [658.69, 166.23], [636.89, 166.26]]}]\n```" - } - ] - }, - { - "id": 3277, - "data_source": "Mono3DRefer", - "scene_name": "005379", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1097.2, 222.8], [1159.3, 228.89], [927.51, 229.44], [890.27, 223.24], [1097.2, 122.51], [1159.3, 116.37], [927.51, 115.81], [890.27, 122.07]]}]\n```" - } - ] - }, - { - "id": 3278, - "data_source": "Mono3DRefer", - "scene_name": "005379", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[374.17, 324.87], [234.86, 322.17], [372.92, 275.24], [469.61, 276.5], [374.17, 196.08], [234.86, 195.66], [372.92, 188.49], [469.61, 188.69]]}, {\"category\": \"car\", \"corners_3d\": [[482.45, 273.94], [383.59, 272.83], [456.68, 247.03], [530.42, 247.64], [482.45, 168.67], [383.59, 168.72], [456.68, 169.78], [530.42, 169.76]]}, {\"category\": \"car\", \"corners_3d\": [[719.53, 222.06], [769.33, 222.06], [799.34, 231.33], [740.15, 231.33], [719.53, 179.83], [769.33, 179.83], [799.34, 181.15], [740.15, 181.15]]}, {\"category\": \"car\", \"corners_3d\": [[556.92, 233.07], [502.23, 232.71], [527.69, 223.45], [574.01, 223.71], [556.92, 174.86], [502.23, 174.85], [527.69, 174.54], [574.01, 174.55]]}, {\"category\": \"car\", \"corners_3d\": [[699.77, 212.03], [743.76, 212.08], [763.25, 218.5], [712.06, 218.43], [699.77, 170.76], [743.76, 170.76], [763.25, 170.41], [712.06, 170.42]]}, {\"category\": \"car\", \"corners_3d\": [[575.66, 220.85], [535.19, 220.72], [547.18, 215.5], [583.25, 215.6], [575.66, 175.54], [535.19, 175.53], [547.18, 175.24], [583.25, 175.25]]}]\n```" - } - ] - }, - { - "id": 3279, - "data_source": "Mono3DRefer", - "scene_name": "005379", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[374.17, 324.87], [234.86, 322.17], [372.92, 275.24], [469.61, 276.5], [374.17, 196.08], [234.86, 195.66], [372.92, 188.49], [469.61, 188.69]]}, {\"category\": \"car\", \"corners_3d\": [[482.45, 273.94], [383.59, 272.83], [456.68, 247.03], [530.42, 247.64], [482.45, 168.67], [383.59, 168.72], [456.68, 169.78], [530.42, 169.76]]}, {\"category\": \"van\", \"corners_3d\": [[1097.2, 222.8], [1159.3, 228.89], [927.51, 229.44], [890.27, 223.24], [1097.2, 122.51], [1159.3, 116.37], [927.51, 115.81], [890.27, 122.07]]}, {\"category\": \"car\", \"corners_3d\": [[719.53, 222.06], [769.33, 222.06], [799.34, 231.33], [740.15, 231.33], [719.53, 179.83], [769.33, 179.83], [799.34, 181.15], [740.15, 181.15]]}, {\"category\": \"car\", \"corners_3d\": [[556.92, 233.07], [502.23, 232.71], [527.69, 223.45], [574.01, 223.71], [556.92, 174.86], [502.23, 174.85], [527.69, 174.54], [574.01, 174.55]]}, {\"category\": \"car\", \"corners_3d\": [[699.77, 212.03], [743.76, 212.08], [763.25, 218.5], [712.06, 218.43], [699.77, 170.76], [743.76, 170.76], [763.25, 170.41], [712.06, 170.42]]}, {\"category\": \"car\", \"corners_3d\": [[575.66, 220.85], [535.19, 220.72], [547.18, 215.5], [583.25, 215.6], [575.66, 175.54], [535.19, 175.53], [547.18, 175.24], [583.25, 175.25]]}]\n```" - } - ] - }, - { - "id": 3280, - "data_source": "Mono3DRefer", - "scene_name": "005393", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[427.07, 217.83], [387.08, 218.03], [400.57, 214.32], [437.23, 214.15], [427.07, 179.96], [387.08, 179.99], [400.57, 179.41], [437.23, 179.38]]}, {\"category\": \"car\", \"corners_3d\": [[449.52, 195.56], [426.42, 195.64], [433.03, 194.18], [454.62, 194.11], [449.52, 178.59], [426.42, 178.61], [433.03, 178.24], [454.62, 178.22]]}]\n```" - } - ] - }, - { - "id": 3281, - "data_source": "Mono3DRefer", - "scene_name": "005402", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[470.47, 270.57], [379.08, 270.58], [432.79, 247.75], [502.82, 247.75], [470.47, 194.08], [379.08, 194.08], [432.79, 189.13], [502.82, 189.12]]}]\n```" - } - ] - }, - { - "id": 3282, - "data_source": "Mono3DRefer", - "scene_name": "005415", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[263.25, 332.46], [414.71, 333.16], [226.89, 476.63], [-57.42, 474.11], [263.25, 204.75], [414.71, 204.89], [226.89, 233.56], [-57.42, 233.06]]}, {\"category\": \"car\", \"corners_3d\": [[699.65, 234.73], [644.45, 234.92], [634.7, 225.51], [681.58, 225.37], [699.65, 183.5], [644.45, 183.53], [634.7, 181.91], [681.58, 181.89]]}, {\"category\": \"car\", \"corners_3d\": [[423.68, 210.14], [431.19, 208.49], [512.18, 208.41], [508.41, 210.04], [423.68, 180.98], [431.19, 180.62], [512.18, 180.6], [508.41, 180.96]]}]\n```" - } - ] - }, - { - "id": 3283, - "data_source": "Mono3DRefer", - "scene_name": "005415", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[553.14, 208.09], [519.8, 207.96], [533.48, 204.7], [563.75, 204.8], [553.14, 174.13], [519.8, 174.13], [533.48, 174.01], [563.75, 174.01]]}, {\"category\": \"van\", \"corners_3d\": [[627.37, 202.01], [654.73, 202.04], [656.93, 204.84], [626.95, 204.79], [627.37, 172.35], [654.73, 172.34], [656.93, 172.3], [626.95, 172.3]]}]\n```" - } - ] - }, - { - "id": 3284, - "data_source": "Mono3DRefer", - "scene_name": "005415", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[263.25, 332.46], [414.71, 333.16], [226.89, 476.63], [-57.42, 474.11], [263.25, 204.75], [414.71, 204.89], [226.89, 233.56], [-57.42, 233.06]]}, {\"category\": \"car\", \"corners_3d\": [[699.65, 234.73], [644.45, 234.92], [634.7, 225.51], [681.58, 225.37], [699.65, 183.5], [644.45, 183.53], [634.7, 181.91], [681.58, 181.89]]}, {\"category\": \"van\", \"corners_3d\": [[553.14, 208.09], [519.8, 207.96], [533.48, 204.7], [563.75, 204.8], [553.14, 174.13], [519.8, 174.13], [533.48, 174.01], [563.75, 174.01]]}, {\"category\": \"van\", \"corners_3d\": [[627.37, 202.01], [654.73, 202.04], [656.93, 204.84], [626.95, 204.79], [627.37, 172.35], [654.73, 172.34], [656.93, 172.3], [626.95, 172.3]]}, {\"category\": \"car\", \"corners_3d\": [[423.68, 210.14], [431.19, 208.49], [512.18, 208.41], [508.41, 210.04], [423.68, 180.98], [431.19, 180.62], [512.18, 180.6], [508.41, 180.96]]}]\n```" - } - ] - }, - { - "id": 3285, - "data_source": "Mono3DRefer", - "scene_name": "005421", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[393.5, 289.89], [285.33, 289.38], [375.87, 258.82], [455.88, 259.1], [393.5, 182.92], [285.33, 182.88], [375.87, 180.25], [455.88, 180.27]]}, {\"category\": \"car\", \"corners_3d\": [[659.7, 200.56], [689.68, 200.64], [692.2, 203.06], [659.61, 202.97], [659.7, 172.15], [689.68, 172.14], [692.2, 172.08], [659.61, 172.08]]}]\n```" - } - ] - }, - { - "id": 3286, - "data_source": "Mono3DRefer", - "scene_name": "005424", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[534.74, 212.55], [500.66, 212.59], [510.17, 208.57], [540.79, 208.54], [534.74, 178.97], [500.66, 178.98], [510.17, 178.36], [540.79, 178.35]]}]\n```" - } - ] - }, - { - "id": 3287, - "data_source": "Mono3DRefer", - "scene_name": "005433", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[473.3, 190.02], [461.17, 189.96], [483.15, 189.44], [494.75, 189.5], [473.3, 139.74], [461.17, 140.01], [483.15, 142.23], [494.75, 141.99]]}]\n```" - } - ] - }, - { - "id": 3288, - "data_source": "Mono3DRefer", - "scene_name": "005433", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[775.95, 363.11], [736.09, 375.53], [653.55, 355.2], [694.23, 345.17], [775.95, 142.19], [736.09, 139.59], [653.55, 143.85], [694.23, 145.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[734.02, 378.28], [685.18, 399.57], [598.72, 379.38], [651.16, 361.67], [734.02, 147.73], [685.18, 144.2], [598.72, 147.54], [651.16, 150.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[399.59, 312.97], [366.98, 321.84], [292.91, 315.01], [328.48, 306.95], [399.59, 145.23], [366.98, 142.86], [292.91, 144.68], [328.48, 146.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[350.17, 317.93], [289.88, 330.87], [234.28, 322.34], [295.96, 310.77], [350.17, 126.98], [289.88, 121.94], [234.28, 125.26], [295.96, 129.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[532.18, 207.55], [515.98, 207.81], [508.51, 206.84], [524.21, 206.61], [532.18, 145.16], [515.98, 144.83], [508.51, 146.09], [524.21, 146.4]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[535.89, 194.51], [516.51, 194.54], [517.22, 194.11], [536.01, 194.07], [535.89, 134.94], [516.51, 134.82], [517.22, 136.24], [536.01, 136.34]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[562.38, 195.46], [543.99, 195.52], [542.79, 195.19], [560.79, 195.13], [562.38, 143.09], [543.99, 142.93], [542.79, 143.75], [560.79, 143.9]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[590.93, 195.21], [576.57, 195.26], [574.08, 194.84], [588.03, 194.79], [590.93, 145.27], [576.57, 145.16], [574.08, 146.17], [588.03, 146.28]]}]\n```" - } - ] - }, - { - "id": 3289, - "data_source": "Mono3DRefer", - "scene_name": "005433", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[775.95, 363.11], [736.09, 375.53], [653.55, 355.2], [694.23, 345.17], [775.95, 142.19], [736.09, 139.59], [653.55, 143.85], [694.23, 145.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[734.02, 378.28], [685.18, 399.57], [598.72, 379.38], [651.16, 361.67], [734.02, 147.73], [685.18, 144.2], [598.72, 147.54], [651.16, 150.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[399.59, 312.97], [366.98, 321.84], [292.91, 315.01], [328.48, 306.95], [399.59, 145.23], [366.98, 142.86], [292.91, 144.68], [328.48, 146.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[350.17, 317.93], [289.88, 330.87], [234.28, 322.34], [295.96, 310.77], [350.17, 126.98], [289.88, 121.94], [234.28, 125.26], [295.96, 129.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[532.18, 207.55], [515.98, 207.81], [508.51, 206.84], [524.21, 206.61], [532.18, 145.16], [515.98, 144.83], [508.51, 146.09], [524.21, 146.4]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[535.89, 194.51], [516.51, 194.54], [517.22, 194.11], [536.01, 194.07], [535.89, 134.94], [516.51, 134.82], [517.22, 136.24], [536.01, 136.34]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[562.38, 195.46], [543.99, 195.52], [542.79, 195.19], [560.79, 195.13], [562.38, 143.09], [543.99, 142.93], [542.79, 143.75], [560.79, 143.9]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[590.93, 195.21], [576.57, 195.26], [574.08, 194.84], [588.03, 194.79], [590.93, 145.27], [576.57, 145.16], [574.08, 146.17], [588.03, 146.28]]}, {\"category\": \"cyclist\", \"corners_3d\": [[473.3, 190.02], [461.17, 189.96], [483.15, 189.44], [494.75, 189.5], [473.3, 139.74], [461.17, 140.01], [483.15, 142.23], [494.75, 141.99]]}]\n```" - } - ] - }, - { - "id": 3290, - "data_source": "Mono3DRefer", - "scene_name": "005435", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[471.03, 214.61], [460.69, 216.9], [357.75, 216.45], [373.29, 214.21], [471.03, 184.24], [460.69, 184.86], [357.75, 184.74], [373.29, 184.13]]}]\n```" - } - ] - }, - { - "id": 3291, - "data_source": "Mono3DRefer", - "scene_name": "005446", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[378.78, 351.74], [222.78, 351.77], [348.17, 293.68], [453.52, 293.67], [378.78, 206.8], [222.78, 206.81], [348.17, 195.78], [453.52, 195.78]]}]\n```" - } - ] - }, - { - "id": 3292, - "data_source": "Mono3DRefer", - "scene_name": "005448", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[595.83, 196.87], [620.74, 196.86], [622.78, 198.81], [595.84, 198.83], [595.83, 172.25], [620.74, 172.25], [622.78, 172.2], [595.84, 172.2]]}, {\"category\": \"car\", \"corners_3d\": [[333.3, 310.02], [454.42, 310.04], [298.44, 447.46], [56.05, 447.39], [333.3, 180.85], [454.42, 180.85], [298.44, 188.86], [56.05, 188.85]]}]\n```" - } - ] - }, - { - "id": 3293, - "data_source": "Mono3DRefer", - "scene_name": "005448", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[725.28, 248.95], [741.08, 247.04], [824.16, 254.47], [809.1, 256.79], [725.28, 151.25], [741.08, 151.79], [824.16, 149.68], [809.1, 149.02]]}, {\"category\": \"cyclist\", \"corners_3d\": [[715.87, 235.65], [725.01, 234.0], [793.8, 237.33], [786.11, 239.17], [715.87, 154.14], [725.01, 154.63], [793.8, 153.64], [786.11, 153.09]]}]\n```" - } - ] - }, - { - "id": 3294, - "data_source": "Mono3DRefer", - "scene_name": "005448", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[733.26, 247.19], [730.27, 245.13], [750.86, 245.18], [754.44, 247.24], [733.26, 162.82], [730.27, 163.09], [750.86, 163.09], [754.44, 162.81]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[707.89, 237.59], [695.93, 236.21], [708.17, 235.46], [720.24, 236.8], [707.89, 167.71], [695.93, 167.82], [708.17, 167.88], [720.24, 167.78]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[326.55, 256.72], [318.88, 259.12], [302.46, 259.16], [310.6, 256.76], [326.55, 172.31], [318.88, 172.29], [302.46, 172.29], [310.6, 172.31]]}]\n```" - } - ] - }, - { - "id": 3295, - "data_source": "Mono3DRefer", - "scene_name": "005448", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[595.83, 196.87], [620.74, 196.86], [622.78, 198.81], [595.84, 198.83], [595.83, 172.25], [620.74, 172.25], [622.78, 172.2], [595.84, 172.2]]}, {\"category\": \"car\", \"corners_3d\": [[333.3, 310.02], [454.42, 310.04], [298.44, 447.46], [56.05, 447.39], [333.3, 180.85], [454.42, 180.85], [298.44, 188.86], [56.05, 188.85]]}, {\"category\": \"cyclist\", \"corners_3d\": [[725.28, 248.95], [741.08, 247.04], [824.16, 254.47], [809.1, 256.79], [725.28, 151.25], [741.08, 151.79], [824.16, 149.68], [809.1, 149.02]]}, {\"category\": \"cyclist\", \"corners_3d\": [[715.87, 235.65], [725.01, 234.0], [793.8, 237.33], [786.11, 239.17], [715.87, 154.14], [725.01, 154.63], [793.8, 153.64], [786.11, 153.09]]}]\n```" - } - ] - }, - { - "id": 3296, - "data_source": "Mono3DRefer", - "scene_name": "005448", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[595.83, 196.87], [620.74, 196.86], [622.78, 198.81], [595.84, 198.83], [595.83, 172.25], [620.74, 172.25], [622.78, 172.2], [595.84, 172.2]]}, {\"category\": \"car\", \"corners_3d\": [[333.3, 310.02], [454.42, 310.04], [298.44, 447.46], [56.05, 447.39], [333.3, 180.85], [454.42, 180.85], [298.44, 188.86], [56.05, 188.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[733.26, 247.19], [730.27, 245.13], [750.86, 245.18], [754.44, 247.24], [733.26, 162.82], [730.27, 163.09], [750.86, 163.09], [754.44, 162.81]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[707.89, 237.59], [695.93, 236.21], [708.17, 235.46], [720.24, 236.8], [707.89, 167.71], [695.93, 167.82], [708.17, 167.88], [720.24, 167.78]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[326.55, 256.72], [318.88, 259.12], [302.46, 259.16], [310.6, 256.76], [326.55, 172.31], [318.88, 172.29], [302.46, 172.29], [310.6, 172.31]]}]\n```" - } - ] - }, - { - "id": 3297, - "data_source": "Mono3DRefer", - "scene_name": "005448", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[725.28, 248.95], [741.08, 247.04], [824.16, 254.47], [809.1, 256.79], [725.28, 151.25], [741.08, 151.79], [824.16, 149.68], [809.1, 149.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[733.26, 247.19], [730.27, 245.13], [750.86, 245.18], [754.44, 247.24], [733.26, 162.82], [730.27, 163.09], [750.86, 163.09], [754.44, 162.81]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[707.89, 237.59], [695.93, 236.21], [708.17, 235.46], [720.24, 236.8], [707.89, 167.71], [695.93, 167.82], [708.17, 167.88], [720.24, 167.78]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[326.55, 256.72], [318.88, 259.12], [302.46, 259.16], [310.6, 256.76], [326.55, 172.31], [318.88, 172.29], [302.46, 172.29], [310.6, 172.31]]}, {\"category\": \"cyclist\", \"corners_3d\": [[715.87, 235.65], [725.01, 234.0], [793.8, 237.33], [786.11, 239.17], [715.87, 154.14], [725.01, 154.63], [793.8, 153.64], [786.11, 153.09]]}]\n```" - } - ] - }, - { - "id": 3298, - "data_source": "Mono3DRefer", - "scene_name": "005448", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[595.83, 196.87], [620.74, 196.86], [622.78, 198.81], [595.84, 198.83], [595.83, 172.25], [620.74, 172.25], [622.78, 172.2], [595.84, 172.2]]}, {\"category\": \"car\", \"corners_3d\": [[333.3, 310.02], [454.42, 310.04], [298.44, 447.46], [56.05, 447.39], [333.3, 180.85], [454.42, 180.85], [298.44, 188.86], [56.05, 188.85]]}, {\"category\": \"cyclist\", \"corners_3d\": [[725.28, 248.95], [741.08, 247.04], [824.16, 254.47], [809.1, 256.79], [725.28, 151.25], [741.08, 151.79], [824.16, 149.68], [809.1, 149.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[733.26, 247.19], [730.27, 245.13], [750.86, 245.18], [754.44, 247.24], [733.26, 162.82], [730.27, 163.09], [750.86, 163.09], [754.44, 162.81]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[707.89, 237.59], [695.93, 236.21], [708.17, 235.46], [720.24, 236.8], [707.89, 167.71], [695.93, 167.82], [708.17, 167.88], [720.24, 167.78]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[326.55, 256.72], [318.88, 259.12], [302.46, 259.16], [310.6, 256.76], [326.55, 172.31], [318.88, 172.29], [302.46, 172.29], [310.6, 172.31]]}, {\"category\": \"cyclist\", \"corners_3d\": [[715.87, 235.65], [725.01, 234.0], [793.8, 237.33], [786.11, 239.17], [715.87, 154.14], [725.01, 154.63], [793.8, 153.64], [786.11, 153.09]]}]\n```" - } - ] - }, - { - "id": 3299, - "data_source": "Mono3DRefer", - "scene_name": "005449", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[414.36, 210.59], [421.28, 208.86], [513.19, 208.58], [510.72, 210.29], [414.36, 177.17], [421.28, 176.97], [513.19, 176.94], [510.72, 177.13]]}, {\"category\": \"car\", \"corners_3d\": [[652.1, 210.4], [688.84, 210.35], [704.14, 216.06], [661.81, 216.14], [652.1, 173.07], [688.84, 173.07], [704.14, 173.11], [661.81, 173.11]]}, {\"category\": \"car\", \"corners_3d\": [[556.74, 221.83], [512.95, 221.78], [526.67, 215.7], [565.03, 215.74], [556.74, 181.51], [512.95, 181.5], [526.67, 180.42], [565.03, 180.43]]}, {\"category\": \"car\", \"corners_3d\": [[228.8, 246.67], [257.4, 239.96], [434.06, 238.51], [423.39, 244.93], [228.8, 180.04], [257.4, 179.39], [434.06, 179.25], [423.39, 179.87]]}, {\"category\": \"car\", \"corners_3d\": [[402.54, 250.37], [376.37, 260.52], [171.42, 260.77], [221.45, 250.56], [402.54, 180.73], [376.37, 181.76], [171.42, 181.79], [221.45, 180.75]]}, {\"category\": \"car\", \"corners_3d\": [[366.18, 280.62], [322.76, 297.36], [63.94, 295.08], [140.91, 278.91], [366.18, 185.99], [322.76, 188.03], [63.94, 187.75], [140.91, 185.78]]}, {\"category\": \"car\", \"corners_3d\": [[307.06, 307.27], [245.37, 337.57], [-185.35, 341.43], [-41.74, 309.83], [307.06, 180.83], [245.37, 182.63], [-185.35, 182.85], [-41.74, 180.98]]}, {\"category\": \"car\", \"corners_3d\": [[726.44, 262.22], [813.1, 262.22], [902.71, 301.68], [777.78, 301.67], [726.44, 192.57], [813.1, 192.58], [902.71, 201.29], [777.78, 201.28]]}]\n```" - } - ] - }, - { - "id": 3300, - "data_source": "Mono3DRefer", - "scene_name": "005451", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[566.85, 216.89], [608.81, 216.54], [627.44, 224.65], [577.8, 225.15], [566.85, 178.17], [608.81, 178.13], [627.44, 179.11], [577.8, 179.17]]}, {\"category\": \"car\", \"corners_3d\": [[619.9, 258.79], [707.78, 257.64], [770.36, 287.57], [652.06, 289.68], [619.9, 172.29], [707.78, 172.3], [770.36, 172.11], [652.06, 172.09]]}]\n```" - } - ] - }, - { - "id": 3301, - "data_source": "Mono3DRefer", - "scene_name": "005455", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[674.94, 226.09], [732.37, 226.6], [739.9, 238.08], [670.25, 237.33], [674.94, 172.5], [732.37, 172.49], [739.9, 172.42], [670.25, 172.42]]}, {\"category\": \"car\", \"corners_3d\": [[492.65, 241.03], [434.98, 239.95], [485.38, 229.14], [534.31, 229.9], [492.65, 172.45], [434.98, 172.46], [485.38, 172.52], [534.31, 172.52]]}, {\"category\": \"car\", \"corners_3d\": [[114.15, 285.58], [31.97, 282.93], [210.27, 255.22], [274.59, 256.7], [114.15, 191.52], [31.97, 191.08], [210.27, 186.49], [274.59, 186.74]]}, {\"category\": \"car\", \"corners_3d\": [[438.1, 226.96], [395.8, 225.82], [462.36, 218.03], [499.53, 218.85], [438.1, 184.97], [395.8, 184.71], [462.36, 182.97], [499.53, 183.15]]}, {\"category\": \"car\", \"corners_3d\": [[820.57, 192.61], [841.0, 193.28], [792.86, 194.18], [773.2, 193.46], [820.57, 167.25], [841.0, 167.06], [792.86, 166.8], [773.2, 167.01]]}, {\"category\": \"car\", \"corners_3d\": [[-198.41, 296.81], [-273.92, 294.07], [8.96, 260.01], [67.48, 261.42], [-198.41, 203.18], [-273.92, 202.51], [8.96, 194.18], [67.48, 194.52]]}]\n```" - } - ] - }, - { - "id": 3302, - "data_source": "Mono3DRefer", - "scene_name": "005456", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[439.46, 305.03], [325.48, 305.05], [399.19, 270.68], [483.53, 270.67], [439.46, 199.12], [325.48, 199.13], [399.19, 192.3], [483.53, 192.29]]}]\n```" - } - ] - }, - { - "id": 3303, - "data_source": "Mono3DRefer", - "scene_name": "005466", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[731.55, 281.41], [852.72, 281.42], [1033.45, 362.31], [821.98, 362.27], [731.55, 171.35], [852.72, 171.35], [1033.45, 170.23], [821.98, 170.23]]}, {\"category\": \"car\", \"corners_3d\": [[-112.7, 295.34], [-17.01, 279.46], [190.59, 280.13], [125.37, 296.23], [-112.7, 195.29], [-17.01, 192.38], [190.59, 192.5], [125.37, 195.45]]}, {\"category\": \"car\", \"corners_3d\": [[493.49, 220.17], [548.17, 219.93], [545.9, 229.34], [480.21, 229.69], [493.49, 170.35], [548.17, 170.36], [545.9, 169.87], [480.21, 169.85]]}, {\"category\": \"car\", \"corners_3d\": [[662.05, 221.98], [715.1, 221.73], [749.62, 232.57], [684.94, 232.95], [662.05, 169.89], [715.1, 169.91], [749.62, 169.26], [684.94, 169.23]]}, {\"category\": \"car\", \"corners_3d\": [[651.21, 210.05], [696.57, 210.1], [709.97, 217.04], [656.16, 216.97], [651.21, 172.34], [696.57, 172.34], [709.97, 172.25], [656.16, 172.25]]}, {\"category\": \"car\", \"corners_3d\": [[537.74, 203.43], [568.45, 203.47], [562.35, 206.42], [528.69, 206.37], [537.74, 174.62], [568.45, 174.62], [562.35, 174.8], [528.69, 174.79]]}, {\"category\": \"car\", \"corners_3d\": [[638.5, 198.39], [670.72, 198.42], [675.79, 201.75], [639.37, 201.71], [638.5, 164.76], [670.72, 164.75], [675.79, 163.7], [639.37, 163.71]]}]\n```" - } - ] - }, - { - "id": 3304, - "data_source": "Mono3DRefer", - "scene_name": "005478", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[511.62, 219.15], [569.38, 219.49], [540.16, 232.37], [466.87, 231.81], [511.62, 160.18], [569.38, 160.09], [540.16, 156.56], [466.87, 156.72]]}, {\"category\": \"car\", \"corners_3d\": [[1352.17, 253.24], [1503.63, 269.63], [1074.74, 269.63], [995.94, 253.24], [1352.17, 135.03], [1503.63, 127.32], [1074.74, 127.32], [995.94, 135.03]]}, {\"category\": \"car\", \"corners_3d\": [[594.45, 198.02], [551.51, 197.95], [562.98, 194.87], [600.67, 194.93], [594.45, 156.92], [551.51, 156.96], [562.98, 158.91], [600.67, 158.87]]}, {\"category\": \"car\", \"corners_3d\": [[629.31, 173.48], [608.29, 173.48], [611.73, 173.44], [631.34, 173.44], [629.31, 155.46], [608.29, 155.5], [611.73, 156.66], [631.34, 156.63]]}]\n```" - } - ] - }, - { - "id": 3305, - "data_source": "Mono3DRefer", - "scene_name": "005485", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[496.97, 223.91], [431.43, 223.87], [474.65, 211.93], [524.88, 211.96], [496.97, 143.24], [431.43, 143.27], [474.65, 150.19], [524.88, 150.18]]}]\n```" - } - ] - }, - { - "id": 3306, - "data_source": "Mono3DRefer", - "scene_name": "005485", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[700.24, 195.24], [700.02, 194.96], [729.74, 195.05], [730.34, 195.34], [700.24, 164.99], [700.02, 165.09], [729.74, 165.06], [730.34, 164.96]]}]\n```" - } - ] - }, - { - "id": 3307, - "data_source": "Mono3DRefer", - "scene_name": "005485", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[540.59, 208.3], [500.99, 208.0], [528.55, 203.54], [563.29, 203.76], [540.59, 176.76], [500.99, 176.73], [528.55, 176.23], [563.29, 176.26]]}]\n```" - } - ] - }, - { - "id": 3308, - "data_source": "Mono3DRefer", - "scene_name": "005485", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[496.97, 223.91], [431.43, 223.87], [474.65, 211.93], [524.88, 211.96], [496.97, 143.24], [431.43, 143.27], [474.65, 150.19], [524.88, 150.18]]}, {\"category\": \"cyclist\", \"corners_3d\": [[700.24, 195.24], [700.02, 194.96], [729.74, 195.05], [730.34, 195.34], [700.24, 164.99], [700.02, 165.09], [729.74, 165.06], [730.34, 164.96]]}]\n```" - } - ] - }, - { - "id": 3309, - "data_source": "Mono3DRefer", - "scene_name": "005485", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[496.97, 223.91], [431.43, 223.87], [474.65, 211.93], [524.88, 211.96], [496.97, 143.24], [431.43, 143.27], [474.65, 150.19], [524.88, 150.18]]}, {\"category\": \"car\", \"corners_3d\": [[540.59, 208.3], [500.99, 208.0], [528.55, 203.54], [563.29, 203.76], [540.59, 176.76], [500.99, 176.73], [528.55, 176.23], [563.29, 176.26]]}]\n```" - } - ] - }, - { - "id": 3310, - "data_source": "Mono3DRefer", - "scene_name": "005485", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[540.59, 208.3], [500.99, 208.0], [528.55, 203.54], [563.29, 203.76], [540.59, 176.76], [500.99, 176.73], [528.55, 176.23], [563.29, 176.26]]}, {\"category\": \"cyclist\", \"corners_3d\": [[700.24, 195.24], [700.02, 194.96], [729.74, 195.05], [730.34, 195.34], [700.24, 164.99], [700.02, 165.09], [729.74, 165.06], [730.34, 164.96]]}]\n```" - } - ] - }, - { - "id": 3311, - "data_source": "Mono3DRefer", - "scene_name": "005485", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[496.97, 223.91], [431.43, 223.87], [474.65, 211.93], [524.88, 211.96], [496.97, 143.24], [431.43, 143.27], [474.65, 150.19], [524.88, 150.18]]}, {\"category\": \"car\", \"corners_3d\": [[540.59, 208.3], [500.99, 208.0], [528.55, 203.54], [563.29, 203.76], [540.59, 176.76], [500.99, 176.73], [528.55, 176.23], [563.29, 176.26]]}, {\"category\": \"cyclist\", \"corners_3d\": [[700.24, 195.24], [700.02, 194.96], [729.74, 195.05], [730.34, 195.34], [700.24, 164.99], [700.02, 165.09], [729.74, 165.06], [730.34, 164.96]]}]\n```" - } - ] - }, - { - "id": 3312, - "data_source": "Mono3DRefer", - "scene_name": "005486", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[251.03, 405.89], [8.73, 404.37], [271.1, 305.55], [410.57, 306.05], [251.03, 177.2], [8.73, 177.17], [271.1, 175.33], [410.57, 175.34]]}, {\"category\": \"car\", \"corners_3d\": [[1257.25, 488.42], [963.26, 488.52], [776.19, 321.84], [914.99, 321.81], [1257.25, 230.18], [963.26, 230.2], [776.19, 199.92], [914.99, 199.91]]}, {\"category\": \"car\", \"corners_3d\": [[485.04, 251.97], [405.98, 252.15], [441.77, 237.1], [505.79, 236.98], [485.04, 175.96], [405.98, 175.97], [441.77, 175.38], [505.79, 175.37]]}, {\"category\": \"car\", \"corners_3d\": [[762.68, 235.75], [699.3, 235.7], [683.71, 223.91], [735.17, 223.94], [762.68, 180.8], [699.3, 180.8], [683.71, 179.31], [735.17, 179.31]]}, {\"category\": \"car\", \"corners_3d\": [[534.06, 222.99], [474.9, 223.11], [492.12, 215.42], [542.2, 215.33], [534.06, 174.93], [474.9, 174.94], [492.12, 174.62], [542.2, 174.61]]}, {\"category\": \"car\", \"corners_3d\": [[652.62, 199.58], [679.23, 199.6], [684.63, 202.24], [655.39, 202.21], [652.62, 173.87], [679.23, 173.87], [684.63, 173.97], [655.39, 173.97]]}]\n```" - } - ] - }, - { - "id": 3313, - "data_source": "Mono3DRefer", - "scene_name": "005486", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[641.25, 191.16], [666.99, 191.17], [672.58, 193.23], [643.94, 193.22], [641.25, 160.25], [666.99, 160.25], [672.58, 158.83], [643.94, 158.83]]}]\n```" - } - ] - }, - { - "id": 3314, - "data_source": "Mono3DRefer", - "scene_name": "005486", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[251.03, 405.89], [8.73, 404.37], [271.1, 305.55], [410.57, 306.05], [251.03, 177.2], [8.73, 177.17], [271.1, 175.33], [410.57, 175.34]]}, {\"category\": \"car\", \"corners_3d\": [[1257.25, 488.42], [963.26, 488.52], [776.19, 321.84], [914.99, 321.81], [1257.25, 230.18], [963.26, 230.2], [776.19, 199.92], [914.99, 199.91]]}, {\"category\": \"car\", \"corners_3d\": [[485.04, 251.97], [405.98, 252.15], [441.77, 237.1], [505.79, 236.98], [485.04, 175.96], [405.98, 175.97], [441.77, 175.38], [505.79, 175.37]]}, {\"category\": \"car\", \"corners_3d\": [[762.68, 235.75], [699.3, 235.7], [683.71, 223.91], [735.17, 223.94], [762.68, 180.8], [699.3, 180.8], [683.71, 179.31], [735.17, 179.31]]}, {\"category\": \"car\", \"corners_3d\": [[534.06, 222.99], [474.9, 223.11], [492.12, 215.42], [542.2, 215.33], [534.06, 174.93], [474.9, 174.94], [492.12, 174.62], [542.2, 174.61]]}, {\"category\": \"car\", \"corners_3d\": [[652.62, 199.58], [679.23, 199.6], [684.63, 202.24], [655.39, 202.21], [652.62, 173.87], [679.23, 173.87], [684.63, 173.97], [655.39, 173.97]]}, {\"category\": \"van\", \"corners_3d\": [[641.25, 191.16], [666.99, 191.17], [672.58, 193.23], [643.94, 193.22], [641.25, 160.25], [666.99, 160.25], [672.58, 158.83], [643.94, 158.83]]}]\n```" - } - ] - }, - { - "id": 3315, - "data_source": "Mono3DRefer", - "scene_name": "005493", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[132.06, 558.69], [-259.75, 570.95], [198.63, 349.87], [369.5, 347.4], [132.06, 181.75], [-259.75, 182.03], [198.63, 176.93], [369.5, 176.88]]}, {\"category\": \"car\", \"corners_3d\": [[389.16, 211.96], [361.41, 212.6], [355.25, 210.17], [381.43, 209.61], [389.16, 187.04], [361.41, 187.27], [355.25, 186.4], [381.43, 186.19]]}, {\"category\": \"car\", \"corners_3d\": [[472.55, 255.36], [395.68, 256.49], [420.86, 238.79], [481.37, 238.09], [472.55, 190.84], [395.68, 191.08], [420.86, 187.22], [481.37, 187.07]]}]\n```" - } - ] - }, - { - "id": 3316, - "data_source": "Mono3DRefer", - "scene_name": "005520", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[56.01, 235.8], [7.61, 235.44], [106.22, 226.08], [147.82, 226.34], [56.01, 182.74], [7.61, 182.69], [106.22, 181.22], [147.82, 181.26]]}]\n```" - } - ] - }, - { - "id": 3317, - "data_source": "Mono3DRefer", - "scene_name": "005522", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[499.68, 250.32], [427.85, 250.4], [461.42, 235.43], [519.37, 235.37], [499.68, 190.33], [427.85, 190.35], [461.42, 186.97], [519.37, 186.96]]}]\n```" - } - ] - }, - { - "id": 3318, - "data_source": "Mono3DRefer", - "scene_name": "005526", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[296.56, 297.07], [278.51, 304.25], [179.41, 304.59], [203.28, 297.38], [296.56, 144.88], [278.51, 142.69], [179.41, 142.59], [203.28, 144.79]]}]\n```" - } - ] - }, - { - "id": 3319, - "data_source": "Mono3DRefer", - "scene_name": "005545", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[613.29, 189.68], [582.76, 189.67], [587.72, 188.9], [615.4, 188.92], [613.29, 162.54], [582.76, 162.58], [587.72, 164.35], [615.4, 164.32]]}]\n```" - } - ] - }, - { - "id": 3320, - "data_source": "Mono3DRefer", - "scene_name": "005545", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[411.53, 329.67], [329.71, 328.31], [394.19, 299.42], [460.27, 300.3], [411.53, 158.47], [329.71, 158.68], [394.19, 163.18], [460.27, 163.04]]}, {\"category\": \"cyclist\", \"corners_3d\": [[209.65, 427.83], [104.79, 426.76], [262.2, 352.03], [335.48, 352.55], [209.65, 148.83], [104.79, 148.97], [262.2, 158.89], [335.48, 158.82]]}]\n```" - } - ] - }, - { - "id": 3321, - "data_source": "Mono3DRefer", - "scene_name": "005545", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[637.58, 192.22], [680.28, 192.25], [690.57, 194.43], [639.17, 194.39], [637.58, 139.34], [680.28, 139.24], [690.57, 130.64], [639.17, 130.79]]}]\n```" - } - ] - }, - { - "id": 3322, - "data_source": "Mono3DRefer", - "scene_name": "005545", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[824.97, 264.12], [885.66, 264.26], [910.44, 271.85], [844.14, 271.67], [824.97, 157.59], [885.66, 157.55], [910.44, 155.36], [844.14, 155.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[718.21, 220.69], [751.72, 220.79], [754.76, 221.82], [720.35, 221.73], [718.21, 158.44], [751.72, 158.39], [754.76, 157.78], [720.35, 157.83]]}]\n```" - } - ] - }, - { - "id": 3323, - "data_source": "Mono3DRefer", - "scene_name": "005545", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[411.53, 329.67], [329.71, 328.31], [394.19, 299.42], [460.27, 300.3], [411.53, 158.47], [329.71, 158.68], [394.19, 163.18], [460.27, 163.04]]}, {\"category\": \"car\", \"corners_3d\": [[613.29, 189.68], [582.76, 189.67], [587.72, 188.9], [615.4, 188.92], [613.29, 162.54], [582.76, 162.58], [587.72, 164.35], [615.4, 164.32]]}, {\"category\": \"cyclist\", \"corners_3d\": [[209.65, 427.83], [104.79, 426.76], [262.2, 352.03], [335.48, 352.55], [209.65, 148.83], [104.79, 148.97], [262.2, 158.89], [335.48, 158.82]]}]\n```" - } - ] - }, - { - "id": 3324, - "data_source": "Mono3DRefer", - "scene_name": "005545", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[613.29, 189.68], [582.76, 189.67], [587.72, 188.9], [615.4, 188.92], [613.29, 162.54], [582.76, 162.58], [587.72, 164.35], [615.4, 164.32]]}, {\"category\": \"van\", \"corners_3d\": [[637.58, 192.22], [680.28, 192.25], [690.57, 194.43], [639.17, 194.39], [637.58, 139.34], [680.28, 139.24], [690.57, 130.64], [639.17, 130.79]]}]\n```" - } - ] - }, - { - "id": 3325, - "data_source": "Mono3DRefer", - "scene_name": "005545", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[824.97, 264.12], [885.66, 264.26], [910.44, 271.85], [844.14, 271.67], [824.97, 157.59], [885.66, 157.55], [910.44, 155.36], [844.14, 155.41]]}, {\"category\": \"car\", \"corners_3d\": [[613.29, 189.68], [582.76, 189.67], [587.72, 188.9], [615.4, 188.92], [613.29, 162.54], [582.76, 162.58], [587.72, 164.35], [615.4, 164.32]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[718.21, 220.69], [751.72, 220.79], [754.76, 221.82], [720.35, 221.73], [718.21, 158.44], [751.72, 158.39], [754.76, 157.78], [720.35, 157.83]]}]\n```" - } - ] - }, - { - "id": 3326, - "data_source": "Mono3DRefer", - "scene_name": "005545", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[411.53, 329.67], [329.71, 328.31], [394.19, 299.42], [460.27, 300.3], [411.53, 158.47], [329.71, 158.68], [394.19, 163.18], [460.27, 163.04]]}, {\"category\": \"cyclist\", \"corners_3d\": [[209.65, 427.83], [104.79, 426.76], [262.2, 352.03], [335.48, 352.55], [209.65, 148.83], [104.79, 148.97], [262.2, 158.89], [335.48, 158.82]]}, {\"category\": \"van\", \"corners_3d\": [[637.58, 192.22], [680.28, 192.25], [690.57, 194.43], [639.17, 194.39], [637.58, 139.34], [680.28, 139.24], [690.57, 130.64], [639.17, 130.79]]}]\n```" - } - ] - }, - { - "id": 3327, - "data_source": "Mono3DRefer", - "scene_name": "005545", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[824.97, 264.12], [885.66, 264.26], [910.44, 271.85], [844.14, 271.67], [824.97, 157.59], [885.66, 157.55], [910.44, 155.36], [844.14, 155.41]]}, {\"category\": \"cyclist\", \"corners_3d\": [[411.53, 329.67], [329.71, 328.31], [394.19, 299.42], [460.27, 300.3], [411.53, 158.47], [329.71, 158.68], [394.19, 163.18], [460.27, 163.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[718.21, 220.69], [751.72, 220.79], [754.76, 221.82], [720.35, 221.73], [718.21, 158.44], [751.72, 158.39], [754.76, 157.78], [720.35, 157.83]]}, {\"category\": \"cyclist\", \"corners_3d\": [[209.65, 427.83], [104.79, 426.76], [262.2, 352.03], [335.48, 352.55], [209.65, 148.83], [104.79, 148.97], [262.2, 158.89], [335.48, 158.82]]}]\n```" - } - ] - }, - { - "id": 3328, - "data_source": "Mono3DRefer", - "scene_name": "005545", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[824.97, 264.12], [885.66, 264.26], [910.44, 271.85], [844.14, 271.67], [824.97, 157.59], [885.66, 157.55], [910.44, 155.36], [844.14, 155.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[718.21, 220.69], [751.72, 220.79], [754.76, 221.82], [720.35, 221.73], [718.21, 158.44], [751.72, 158.39], [754.76, 157.78], [720.35, 157.83]]}, {\"category\": \"van\", \"corners_3d\": [[637.58, 192.22], [680.28, 192.25], [690.57, 194.43], [639.17, 194.39], [637.58, 139.34], [680.28, 139.24], [690.57, 130.64], [639.17, 130.79]]}]\n```" - } - ] - }, - { - "id": 3329, - "data_source": "Mono3DRefer", - "scene_name": "005545", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[411.53, 329.67], [329.71, 328.31], [394.19, 299.42], [460.27, 300.3], [411.53, 158.47], [329.71, 158.68], [394.19, 163.18], [460.27, 163.04]]}, {\"category\": \"car\", \"corners_3d\": [[613.29, 189.68], [582.76, 189.67], [587.72, 188.9], [615.4, 188.92], [613.29, 162.54], [582.76, 162.58], [587.72, 164.35], [615.4, 164.32]]}, {\"category\": \"cyclist\", \"corners_3d\": [[209.65, 427.83], [104.79, 426.76], [262.2, 352.03], [335.48, 352.55], [209.65, 148.83], [104.79, 148.97], [262.2, 158.89], [335.48, 158.82]]}, {\"category\": \"van\", \"corners_3d\": [[637.58, 192.22], [680.28, 192.25], [690.57, 194.43], [639.17, 194.39], [637.58, 139.34], [680.28, 139.24], [690.57, 130.64], [639.17, 130.79]]}]\n```" - } - ] - }, - { - "id": 3330, - "data_source": "Mono3DRefer", - "scene_name": "005545", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[824.97, 264.12], [885.66, 264.26], [910.44, 271.85], [844.14, 271.67], [824.97, 157.59], [885.66, 157.55], [910.44, 155.36], [844.14, 155.41]]}, {\"category\": \"cyclist\", \"corners_3d\": [[411.53, 329.67], [329.71, 328.31], [394.19, 299.42], [460.27, 300.3], [411.53, 158.47], [329.71, 158.68], [394.19, 163.18], [460.27, 163.04]]}, {\"category\": \"car\", \"corners_3d\": [[613.29, 189.68], [582.76, 189.67], [587.72, 188.9], [615.4, 188.92], [613.29, 162.54], [582.76, 162.58], [587.72, 164.35], [615.4, 164.32]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[718.21, 220.69], [751.72, 220.79], [754.76, 221.82], [720.35, 221.73], [718.21, 158.44], [751.72, 158.39], [754.76, 157.78], [720.35, 157.83]]}, {\"category\": \"cyclist\", \"corners_3d\": [[209.65, 427.83], [104.79, 426.76], [262.2, 352.03], [335.48, 352.55], [209.65, 148.83], [104.79, 148.97], [262.2, 158.89], [335.48, 158.82]]}]\n```" - } - ] - }, - { - "id": 3331, - "data_source": "Mono3DRefer", - "scene_name": "005545", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[824.97, 264.12], [885.66, 264.26], [910.44, 271.85], [844.14, 271.67], [824.97, 157.59], [885.66, 157.55], [910.44, 155.36], [844.14, 155.41]]}, {\"category\": \"car\", \"corners_3d\": [[613.29, 189.68], [582.76, 189.67], [587.72, 188.9], [615.4, 188.92], [613.29, 162.54], [582.76, 162.58], [587.72, 164.35], [615.4, 164.32]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[718.21, 220.69], [751.72, 220.79], [754.76, 221.82], [720.35, 221.73], [718.21, 158.44], [751.72, 158.39], [754.76, 157.78], [720.35, 157.83]]}, {\"category\": \"van\", \"corners_3d\": [[637.58, 192.22], [680.28, 192.25], [690.57, 194.43], [639.17, 194.39], [637.58, 139.34], [680.28, 139.24], [690.57, 130.64], [639.17, 130.79]]}]\n```" - } - ] - }, - { - "id": 3332, - "data_source": "Mono3DRefer", - "scene_name": "005545", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[824.97, 264.12], [885.66, 264.26], [910.44, 271.85], [844.14, 271.67], [824.97, 157.59], [885.66, 157.55], [910.44, 155.36], [844.14, 155.41]]}, {\"category\": \"cyclist\", \"corners_3d\": [[411.53, 329.67], [329.71, 328.31], [394.19, 299.42], [460.27, 300.3], [411.53, 158.47], [329.71, 158.68], [394.19, 163.18], [460.27, 163.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[718.21, 220.69], [751.72, 220.79], [754.76, 221.82], [720.35, 221.73], [718.21, 158.44], [751.72, 158.39], [754.76, 157.78], [720.35, 157.83]]}, {\"category\": \"cyclist\", \"corners_3d\": [[209.65, 427.83], [104.79, 426.76], [262.2, 352.03], [335.48, 352.55], [209.65, 148.83], [104.79, 148.97], [262.2, 158.89], [335.48, 158.82]]}, {\"category\": \"van\", \"corners_3d\": [[637.58, 192.22], [680.28, 192.25], [690.57, 194.43], [639.17, 194.39], [637.58, 139.34], [680.28, 139.24], [690.57, 130.64], [639.17, 130.79]]}]\n```" - } - ] - }, - { - "id": 3333, - "data_source": "Mono3DRefer", - "scene_name": "005545", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[824.97, 264.12], [885.66, 264.26], [910.44, 271.85], [844.14, 271.67], [824.97, 157.59], [885.66, 157.55], [910.44, 155.36], [844.14, 155.41]]}, {\"category\": \"cyclist\", \"corners_3d\": [[411.53, 329.67], [329.71, 328.31], [394.19, 299.42], [460.27, 300.3], [411.53, 158.47], [329.71, 158.68], [394.19, 163.18], [460.27, 163.04]]}, {\"category\": \"car\", \"corners_3d\": [[613.29, 189.68], [582.76, 189.67], [587.72, 188.9], [615.4, 188.92], [613.29, 162.54], [582.76, 162.58], [587.72, 164.35], [615.4, 164.32]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[718.21, 220.69], [751.72, 220.79], [754.76, 221.82], [720.35, 221.73], [718.21, 158.44], [751.72, 158.39], [754.76, 157.78], [720.35, 157.83]]}, {\"category\": \"cyclist\", \"corners_3d\": [[209.65, 427.83], [104.79, 426.76], [262.2, 352.03], [335.48, 352.55], [209.65, 148.83], [104.79, 148.97], [262.2, 158.89], [335.48, 158.82]]}, {\"category\": \"van\", \"corners_3d\": [[637.58, 192.22], [680.28, 192.25], [690.57, 194.43], [639.17, 194.39], [637.58, 139.34], [680.28, 139.24], [690.57, 130.64], [639.17, 130.79]]}]\n```" - } - ] - }, - { - "id": 3334, - "data_source": "Mono3DRefer", - "scene_name": "005550", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[566.81, 201.62], [541.88, 201.66], [544.79, 199.46], [567.82, 199.43], [566.81, 177.04], [541.88, 177.05], [544.79, 176.73], [567.82, 176.72]]}]\n```" - } - ] - }, - { - "id": 3335, - "data_source": "Mono3DRefer", - "scene_name": "005558", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[642.03, 198.97], [673.04, 198.96], [681.33, 200.38], [647.09, 200.4], [642.03, 170.72], [673.04, 170.73], [681.33, 169.22], [647.09, 169.21]]}, {\"category\": \"car\", \"corners_3d\": [[575.87, 207.7], [612.44, 207.69], [613.95, 210.53], [572.74, 210.55], [575.87, 175.11], [612.44, 175.11], [613.95, 173.83], [572.74, 173.83]]}, {\"category\": \"car\", \"corners_3d\": [[569.89, 197.63], [598.14, 197.6], [601.41, 198.81], [570.41, 198.84], [569.89, 173.61], [598.14, 173.64], [601.41, 172.51], [570.41, 172.48]]}, {\"category\": \"car\", \"corners_3d\": [[612.88, 185.62], [629.05, 185.62], [630.75, 185.64], [613.81, 185.64], [612.88, 169.86], [629.05, 169.87], [630.75, 169.13], [613.81, 169.13]]}]\n```" - } - ] - }, - { - "id": 3336, - "data_source": "Mono3DRefer", - "scene_name": "005563", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[551.93, 217.31], [516.8, 217.57], [517.34, 212.94], [548.85, 212.73], [551.93, 181.4], [516.8, 181.45], [517.34, 180.56], [548.85, 180.52]]}, {\"category\": \"car\", \"corners_3d\": [[521.95, 203.84], [496.97, 204.12], [489.99, 201.63], [513.05, 201.39], [521.95, 182.35], [496.97, 182.44], [489.99, 181.67], [513.05, 181.6]]}]\n```" - } - ] - }, - { - "id": 3337, - "data_source": "Mono3DRefer", - "scene_name": "005571", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[696.06, 229.22], [705.07, 228.01], [760.57, 230.73], [752.39, 232.07], [696.06, 160.76], [705.07, 161.02], [760.57, 160.43], [752.39, 160.15]]}, {\"category\": \"cyclist\", \"corners_3d\": [[699.2, 239.94], [712.67, 238.56], [777.17, 243.85], [764.06, 245.46], [699.2, 160.34], [712.67, 160.6], [777.17, 159.61], [764.06, 159.31]]}]\n```" - } - ] - }, - { - "id": 3338, - "data_source": "Mono3DRefer", - "scene_name": "005571", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[689.56, 232.15], [679.59, 231.08], [690.11, 230.48], [700.17, 231.53], [689.56, 172.84], [679.59, 172.84], [690.11, 172.84], [700.17, 172.84]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[371.0, 244.84], [365.7, 246.54], [352.13, 246.57], [357.74, 244.86], [371.0, 174.59], [365.7, 174.64], [352.13, 174.64], [357.74, 174.59]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[709.51, 237.01], [707.23, 235.53], [724.3, 235.54], [726.98, 237.01], [709.51, 167.16], [707.23, 167.29], [724.3, 167.29], [726.98, 167.16]]}]\n```" - } - ] - }, - { - "id": 3339, - "data_source": "Mono3DRefer", - "scene_name": "005571", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[592.87, 198.34], [617.75, 198.32], [619.54, 200.39], [592.64, 200.41], [592.87, 173.74], [617.75, 173.74], [619.54, 173.81], [592.64, 173.82]]}, {\"category\": \"car\", \"corners_3d\": [[397.02, 278.69], [489.94, 278.7], [415.05, 344.66], [264.25, 344.63], [397.02, 179.6], [489.94, 179.6], [415.05, 183.81], [264.25, 183.81]]}, {\"category\": \"car\", \"corners_3d\": [[-57.2, 295.65], [-158.94, 295.84], [14.93, 267.73], [93.24, 267.62], [-57.2, 201.49], [-158.94, 201.53], [14.93, 194.98], [93.24, 194.95]]}]\n```" - } - ] - }, - { - "id": 3340, - "data_source": "Mono3DRefer", - "scene_name": "005571", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[689.56, 232.15], [679.59, 231.08], [690.11, 230.48], [700.17, 231.53], [689.56, 172.84], [679.59, 172.84], [690.11, 172.84], [700.17, 172.84]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[371.0, 244.84], [365.7, 246.54], [352.13, 246.57], [357.74, 244.86], [371.0, 174.59], [365.7, 174.64], [352.13, 174.64], [357.74, 174.59]]}, {\"category\": \"cyclist\", \"corners_3d\": [[696.06, 229.22], [705.07, 228.01], [760.57, 230.73], [752.39, 232.07], [696.06, 160.76], [705.07, 161.02], [760.57, 160.43], [752.39, 160.15]]}, {\"category\": \"cyclist\", \"corners_3d\": [[699.2, 239.94], [712.67, 238.56], [777.17, 243.85], [764.06, 245.46], [699.2, 160.34], [712.67, 160.6], [777.17, 159.61], [764.06, 159.31]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[709.51, 237.01], [707.23, 235.53], [724.3, 235.54], [726.98, 237.01], [709.51, 167.16], [707.23, 167.29], [724.3, 167.29], [726.98, 167.16]]}]\n```" - } - ] - }, - { - "id": 3341, - "data_source": "Mono3DRefer", - "scene_name": "005571", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[696.06, 229.22], [705.07, 228.01], [760.57, 230.73], [752.39, 232.07], [696.06, 160.76], [705.07, 161.02], [760.57, 160.43], [752.39, 160.15]]}, {\"category\": \"car\", \"corners_3d\": [[592.87, 198.34], [617.75, 198.32], [619.54, 200.39], [592.64, 200.41], [592.87, 173.74], [617.75, 173.74], [619.54, 173.81], [592.64, 173.82]]}, {\"category\": \"car\", \"corners_3d\": [[397.02, 278.69], [489.94, 278.7], [415.05, 344.66], [264.25, 344.63], [397.02, 179.6], [489.94, 179.6], [415.05, 183.81], [264.25, 183.81]]}, {\"category\": \"car\", \"corners_3d\": [[-57.2, 295.65], [-158.94, 295.84], [14.93, 267.73], [93.24, 267.62], [-57.2, 201.49], [-158.94, 201.53], [14.93, 194.98], [93.24, 194.95]]}, {\"category\": \"cyclist\", \"corners_3d\": [[699.2, 239.94], [712.67, 238.56], [777.17, 243.85], [764.06, 245.46], [699.2, 160.34], [712.67, 160.6], [777.17, 159.61], [764.06, 159.31]]}]\n```" - } - ] - }, - { - "id": 3342, - "data_source": "Mono3DRefer", - "scene_name": "005571", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[689.56, 232.15], [679.59, 231.08], [690.11, 230.48], [700.17, 231.53], [689.56, 172.84], [679.59, 172.84], [690.11, 172.84], [700.17, 172.84]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[371.0, 244.84], [365.7, 246.54], [352.13, 246.57], [357.74, 244.86], [371.0, 174.59], [365.7, 174.64], [352.13, 174.64], [357.74, 174.59]]}, {\"category\": \"car\", \"corners_3d\": [[592.87, 198.34], [617.75, 198.32], [619.54, 200.39], [592.64, 200.41], [592.87, 173.74], [617.75, 173.74], [619.54, 173.81], [592.64, 173.82]]}, {\"category\": \"car\", \"corners_3d\": [[397.02, 278.69], [489.94, 278.7], [415.05, 344.66], [264.25, 344.63], [397.02, 179.6], [489.94, 179.6], [415.05, 183.81], [264.25, 183.81]]}, {\"category\": \"car\", \"corners_3d\": [[-57.2, 295.65], [-158.94, 295.84], [14.93, 267.73], [93.24, 267.62], [-57.2, 201.49], [-158.94, 201.53], [14.93, 194.98], [93.24, 194.95]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[709.51, 237.01], [707.23, 235.53], [724.3, 235.54], [726.98, 237.01], [709.51, 167.16], [707.23, 167.29], [724.3, 167.29], [726.98, 167.16]]}]\n```" - } - ] - }, - { - "id": 3343, - "data_source": "Mono3DRefer", - "scene_name": "005571", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[689.56, 232.15], [679.59, 231.08], [690.11, 230.48], [700.17, 231.53], [689.56, 172.84], [679.59, 172.84], [690.11, 172.84], [700.17, 172.84]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[371.0, 244.84], [365.7, 246.54], [352.13, 246.57], [357.74, 244.86], [371.0, 174.59], [365.7, 174.64], [352.13, 174.64], [357.74, 174.59]]}, {\"category\": \"cyclist\", \"corners_3d\": [[696.06, 229.22], [705.07, 228.01], [760.57, 230.73], [752.39, 232.07], [696.06, 160.76], [705.07, 161.02], [760.57, 160.43], [752.39, 160.15]]}, {\"category\": \"car\", \"corners_3d\": [[592.87, 198.34], [617.75, 198.32], [619.54, 200.39], [592.64, 200.41], [592.87, 173.74], [617.75, 173.74], [619.54, 173.81], [592.64, 173.82]]}, {\"category\": \"car\", \"corners_3d\": [[397.02, 278.69], [489.94, 278.7], [415.05, 344.66], [264.25, 344.63], [397.02, 179.6], [489.94, 179.6], [415.05, 183.81], [264.25, 183.81]]}, {\"category\": \"car\", \"corners_3d\": [[-57.2, 295.65], [-158.94, 295.84], [14.93, 267.73], [93.24, 267.62], [-57.2, 201.49], [-158.94, 201.53], [14.93, 194.98], [93.24, 194.95]]}, {\"category\": \"cyclist\", \"corners_3d\": [[699.2, 239.94], [712.67, 238.56], [777.17, 243.85], [764.06, 245.46], [699.2, 160.34], [712.67, 160.6], [777.17, 159.61], [764.06, 159.31]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[709.51, 237.01], [707.23, 235.53], [724.3, 235.54], [726.98, 237.01], [709.51, 167.16], [707.23, 167.29], [724.3, 167.29], [726.98, 167.16]]}]\n```" - } - ] - }, - { - "id": 3344, - "data_source": "Mono3DRefer", - "scene_name": "005585", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[443.81, 279.48], [406.09, 296.9], [114.31, 292.25], [190.46, 276.03], [443.81, 194.9], [406.09, 198.5], [114.31, 197.54], [190.46, 194.18]]}, {\"category\": \"car\", \"corners_3d\": [[1278.18, 316.82], [1487.01, 364.9], [930.96, 356.62], [863.62, 312.11], [1278.18, 172.82], [1487.01, 172.81], [930.96, 172.81], [863.62, 172.82]]}, {\"category\": \"car\", \"corners_3d\": [[1074.81, 277.64], [1163.12, 296.65], [821.66, 298.45], [785.33, 278.93], [1074.81, 161.48], [1163.12, 159.42], [821.66, 159.22], [785.33, 161.34]]}, {\"category\": \"car\", \"corners_3d\": [[358.56, 236.04], [381.84, 231.41], [515.87, 232.43], [502.97, 237.24], [358.56, 185.55], [381.84, 184.62], [515.87, 184.83], [502.97, 185.79]]}, {\"category\": \"car\", \"corners_3d\": [[412.5, 220.44], [420.83, 217.57], [535.7, 217.0], [534.76, 219.8], [412.5, 182.03], [420.83, 181.48], [535.7, 181.37], [534.76, 181.91]]}]\n```" - } - ] - }, - { - "id": 3345, - "data_source": "Mono3DRefer", - "scene_name": "005587", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[235.28, 234.83], [234.87, 235.57], [179.75, 237.38], [180.84, 236.6], [235.28, 185.13], [234.87, 185.28], [179.75, 185.64], [180.84, 185.49]]}]\n```" - } - ] - }, - { - "id": 3346, - "data_source": "Mono3DRefer", - "scene_name": "005587", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[151.88, 240.19], [146.79, 241.48], [131.98, 241.8], [137.37, 240.49], [151.88, 189.37], [146.79, 189.68], [131.98, 189.76], [137.37, 189.44]]}]\n```" - } - ] - }, - { - "id": 3347, - "data_source": "Mono3DRefer", - "scene_name": "005587", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[596.61, 201.89], [624.81, 201.92], [624.83, 204.62], [594.0, 204.6], [596.61, 174.03], [624.81, 174.03], [624.83, 174.14], [594.0, 174.14]]}]\n```" - } - ] - }, - { - "id": 3348, - "data_source": "Mono3DRefer", - "scene_name": "005587", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[151.88, 240.19], [146.79, 241.48], [131.98, 241.8], [137.37, 240.49], [151.88, 189.37], [146.79, 189.68], [131.98, 189.76], [137.37, 189.44]]}, {\"category\": \"cyclist\", \"corners_3d\": [[235.28, 234.83], [234.87, 235.57], [179.75, 237.38], [180.84, 236.6], [235.28, 185.13], [234.87, 185.28], [179.75, 185.64], [180.84, 185.49]]}]\n```" - } - ] - }, - { - "id": 3349, - "data_source": "Mono3DRefer", - "scene_name": "005587", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[596.61, 201.89], [624.81, 201.92], [624.83, 204.62], [594.0, 204.6], [596.61, 174.03], [624.81, 174.03], [624.83, 174.14], [594.0, 174.14]]}, {\"category\": \"cyclist\", \"corners_3d\": [[235.28, 234.83], [234.87, 235.57], [179.75, 237.38], [180.84, 236.6], [235.28, 185.13], [234.87, 185.28], [179.75, 185.64], [180.84, 185.49]]}]\n```" - } - ] - }, - { - "id": 3350, - "data_source": "Mono3DRefer", - "scene_name": "005587", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[596.61, 201.89], [624.81, 201.92], [624.83, 204.62], [594.0, 204.6], [596.61, 174.03], [624.81, 174.03], [624.83, 174.14], [594.0, 174.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[151.88, 240.19], [146.79, 241.48], [131.98, 241.8], [137.37, 240.49], [151.88, 189.37], [146.79, 189.68], [131.98, 189.76], [137.37, 189.44]]}]\n```" - } - ] - }, - { - "id": 3351, - "data_source": "Mono3DRefer", - "scene_name": "005587", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[596.61, 201.89], [624.81, 201.92], [624.83, 204.62], [594.0, 204.6], [596.61, 174.03], [624.81, 174.03], [624.83, 174.14], [594.0, 174.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[151.88, 240.19], [146.79, 241.48], [131.98, 241.8], [137.37, 240.49], [151.88, 189.37], [146.79, 189.68], [131.98, 189.76], [137.37, 189.44]]}, {\"category\": \"cyclist\", \"corners_3d\": [[235.28, 234.83], [234.87, 235.57], [179.75, 237.38], [180.84, 236.6], [235.28, 185.13], [234.87, 185.28], [179.75, 185.64], [180.84, 185.49]]}]\n```" - } - ] - }, - { - "id": 3352, - "data_source": "Mono3DRefer", - "scene_name": "005597", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[562.9, 200.2], [536.37, 200.25], [539.31, 198.04], [563.7, 198.0], [562.9, 174.08], [536.37, 174.08], [539.31, 173.98], [563.7, 173.98]]}]\n```" - } - ] - }, - { - "id": 3353, - "data_source": "Mono3DRefer", - "scene_name": "005600", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-320.63, 806.07], [-1087.75, 871.58], [50.77, 379.62], [260.0, 373.48], [-320.63, 140.34], [-1087.75, 136.97], [50.77, 162.24], [260.0, 162.55]]}]\n```" - } - ] - }, - { - "id": 3354, - "data_source": "Mono3DRefer", - "scene_name": "005600", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[317.98, 361.17], [131.57, 361.21], [316.36, 288.31], [430.6, 288.29], [317.98, 187.15], [131.57, 187.15], [316.36, 181.62], [430.6, 181.62]]}, {\"category\": \"car\", \"corners_3d\": [[668.77, 199.57], [700.79, 199.8], [696.97, 201.93], [662.49, 201.67], [668.77, 167.54], [700.79, 167.5], [696.97, 167.07], [662.49, 167.13]]}, {\"category\": \"car\", \"corners_3d\": [[637.77, 189.78], [612.85, 189.64], [626.43, 188.32], [649.48, 188.44], [637.77, 167.97], [612.85, 168.01], [626.43, 168.39], [649.48, 168.35]]}]\n```" - } - ] - }, - { - "id": 3355, - "data_source": "Mono3DRefer", - "scene_name": "005600", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-320.63, 806.07], [-1087.75, 871.58], [50.77, 379.62], [260.0, 373.48], [-320.63, 140.34], [-1087.75, 136.97], [50.77, 162.24], [260.0, 162.55]]}, {\"category\": \"car\", \"corners_3d\": [[317.98, 361.17], [131.57, 361.21], [316.36, 288.31], [430.6, 288.29], [317.98, 187.15], [131.57, 187.15], [316.36, 181.62], [430.6, 181.62]]}, {\"category\": \"car\", \"corners_3d\": [[668.77, 199.57], [700.79, 199.8], [696.97, 201.93], [662.49, 201.67], [668.77, 167.54], [700.79, 167.5], [696.97, 167.07], [662.49, 167.13]]}, {\"category\": \"car\", \"corners_3d\": [[637.77, 189.78], [612.85, 189.64], [626.43, 188.32], [649.48, 188.44], [637.77, 167.97], [612.85, 168.01], [626.43, 168.39], [649.48, 168.35]]}]\n```" - } - ] - }, - { - "id": 3356, - "data_source": "Mono3DRefer", - "scene_name": "005614", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[949.16, 208.34], [983.79, 209.04], [984.54, 212.96], [946.24, 212.1], [949.16, 179.99], [983.79, 180.13], [984.54, 180.91], [946.24, 180.74]]}]\n```" - } - ] - }, - { - "id": 3357, - "data_source": "Mono3DRefer", - "scene_name": "005618", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-91.52, 322.14], [36.95, 300.13], [247.73, 310.92], [146.44, 337.21], [-91.52, 202.85], [36.95, 198.43], [247.73, 200.6], [146.44, 205.88]]}, {\"category\": \"car\", \"corners_3d\": [[623.97, 233.54], [687.92, 234.3], [682.07, 249.32], [602.82, 248.15], [623.97, 175.39], [687.92, 175.42], [682.07, 176.04], [602.82, 175.99]]}, {\"category\": \"car\", \"corners_3d\": [[826.64, 237.14], [895.08, 237.56], [964.38, 256.69], [875.28, 255.98], [826.64, 172.43], [895.08, 172.43], [964.38, 172.3], [875.28, 172.31]]}, {\"category\": \"car\", \"corners_3d\": [[806.26, 218.8], [861.71, 219.28], [895.74, 229.82], [827.42, 229.09], [806.26, 174.33], [861.71, 174.34], [895.74, 174.68], [827.42, 174.66]]}, {\"category\": \"car\", \"corners_3d\": [[673.43, 211.05], [707.73, 211.41], [702.55, 215.51], [664.7, 215.07], [673.43, 179.18], [707.73, 179.24], [702.55, 179.92], [664.7, 179.84]]}, {\"category\": \"car\", \"corners_3d\": [[787.2, 205.89], [825.04, 206.25], [834.07, 211.18], [790.6, 210.71], [787.2, 167.86], [825.04, 167.81], [834.07, 167.06], [790.6, 167.13]]}, {\"category\": \"car\", \"corners_3d\": [[1016.13, 354.73], [1253.18, 364.64], [2850.33, 963.21], [1716.32, 818.37], [1016.13, 166.49], [1253.18, 166.15], [2850.33, 145.21], [1716.32, 150.27]]}]\n```" - } - ] - }, - { - "id": 3358, - "data_source": "Mono3DRefer", - "scene_name": "005635", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[747.28, 219.56], [791.87, 220.14], [788.54, 225.27], [737.59, 224.51], [747.28, 180.28], [791.87, 180.19], [788.54, 179.46], [737.59, 179.57]]}, {\"category\": \"car\", \"corners_3d\": [[95.22, 244.1], [140.66, 243.03], [111.03, 251.39], [58.23, 252.81], [95.22, 212.05], [140.66, 211.56], [111.03, 215.38], [58.23, 216.02]]}, {\"category\": \"car\", \"corners_3d\": [[157.65, 233.36], [195.91, 232.58], [183.85, 236.63], [142.05, 237.55], [157.65, 200.87], [195.91, 200.61], [183.85, 201.93], [142.05, 202.23]]}, {\"category\": \"car\", \"corners_3d\": [[805.81, 198.9], [827.34, 199.06], [820.98, 199.93], [798.19, 199.75], [805.81, 180.41], [827.34, 180.36], [820.98, 180.05], [798.19, 180.12]]}, {\"category\": \"car\", \"corners_3d\": [[875.76, 189.55], [892.84, 189.59], [887.38, 189.8], [869.58, 189.74], [875.76, 175.95], [892.84, 175.84], [887.38, 175.41], [869.58, 175.52]]}]\n```" - } - ] - }, - { - "id": 3359, - "data_source": "Mono3DRefer", - "scene_name": "005655", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[552.11, 194.4], [531.71, 194.41], [536.42, 192.96], [555.44, 192.95], [552.11, 172.85], [531.71, 172.85], [536.42, 172.85], [555.44, 172.85]]}]\n```" - } - ] - }, - { - "id": 3360, - "data_source": "Mono3DRefer", - "scene_name": "005657", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[540.27, 224.42], [502.67, 224.75], [504.79, 219.05], [538.27, 218.8], [540.27, 186.08], [502.67, 186.17], [504.79, 184.71], [538.27, 184.64]]}, {\"category\": \"car\", \"corners_3d\": [[512.42, 210.57], [485.36, 210.96], [478.21, 207.71], [503.05, 207.38], [512.42, 187.37], [485.36, 187.52], [478.21, 186.27], [503.05, 186.14]]}]\n```" - } - ] - }, - { - "id": 3361, - "data_source": "Mono3DRefer", - "scene_name": "005670", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1016.29, 349.08], [978.5, 354.54], [904.2, 335.82], [940.12, 331.46], [1016.29, 146.54], [978.5, 145.44], [904.2, 149.21], [940.12, 150.09]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[653.47, 263.86], [611.06, 264.94], [600.5, 259.31], [640.24, 258.36], [653.47, 129.34], [611.06, 128.67], [600.5, 132.13], [640.24, 132.71]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[704.35, 262.23], [665.19, 263.28], [654.81, 259.37], [692.26, 258.42], [704.35, 145.6], [665.19, 145.16], [654.81, 146.83], [692.26, 147.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[790.91, 250.52], [766.29, 251.54], [743.42, 247.75], [767.05, 246.84], [790.91, 150.97], [766.29, 150.54], [743.42, 152.14], [767.05, 152.52]]}]\n```" - } - ] - }, - { - "id": 3362, - "data_source": "Mono3DRefer", - "scene_name": "005678", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[393.99, 330.51], [249.53, 329.9], [373.85, 277.68], [470.48, 277.95], [393.99, 202.34], [249.53, 202.22], [373.85, 192.46], [470.48, 192.51]]}, {\"category\": \"car\", \"corners_3d\": [[552.58, 218.18], [514.22, 218.11], [526.17, 213.46], [560.61, 213.52], [552.58, 185.54], [514.22, 185.52], [526.17, 184.22], [560.61, 184.23]]}]\n```" - } - ] - }, - { - "id": 3363, - "data_source": "Mono3DRefer", - "scene_name": "005690", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[92.08, 276.61], [147.65, 264.78], [342.89, 263.77], [312.82, 275.33], [92.08, 191.54], [147.65, 189.41], [342.89, 189.23], [312.82, 191.31]]}]\n```" - } - ] - }, - { - "id": 3364, - "data_source": "Mono3DRefer", - "scene_name": "005690", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[627.99, 221.2], [647.42, 221.26], [648.26, 226.07], [626.89, 226.01], [627.99, 168.53], [647.42, 168.53], [648.26, 168.1], [626.89, 168.1]]}]\n```" - } - ] - }, - { - "id": 3365, - "data_source": "Mono3DRefer", - "scene_name": "005690", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[533.77, 204.09], [541.36, 203.56], [558.15, 203.96], [550.75, 204.49], [533.77, 166.31], [541.36, 166.42], [558.15, 166.34], [550.75, 166.23]]}]\n```" - } - ] - }, - { - "id": 3366, - "data_source": "Mono3DRefer", - "scene_name": "005690", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[627.99, 221.2], [647.42, 221.26], [648.26, 226.07], [626.89, 226.01], [627.99, 168.53], [647.42, 168.53], [648.26, 168.1], [626.89, 168.1]]}, {\"category\": \"car\", \"corners_3d\": [[92.08, 276.61], [147.65, 264.78], [342.89, 263.77], [312.82, 275.33], [92.08, 191.54], [147.65, 189.41], [342.89, 189.23], [312.82, 191.31]]}]\n```" - } - ] - }, - { - "id": 3367, - "data_source": "Mono3DRefer", - "scene_name": "005690", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[92.08, 276.61], [147.65, 264.78], [342.89, 263.77], [312.82, 275.33], [92.08, 191.54], [147.65, 189.41], [342.89, 189.23], [312.82, 191.31]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[533.77, 204.09], [541.36, 203.56], [558.15, 203.96], [550.75, 204.49], [533.77, 166.31], [541.36, 166.42], [558.15, 166.34], [550.75, 166.23]]}]\n```" - } - ] - }, - { - "id": 3368, - "data_source": "Mono3DRefer", - "scene_name": "005690", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[627.99, 221.2], [647.42, 221.26], [648.26, 226.07], [626.89, 226.01], [627.99, 168.53], [647.42, 168.53], [648.26, 168.1], [626.89, 168.1]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[533.77, 204.09], [541.36, 203.56], [558.15, 203.96], [550.75, 204.49], [533.77, 166.31], [541.36, 166.42], [558.15, 166.34], [550.75, 166.23]]}]\n```" - } - ] - }, - { - "id": 3369, - "data_source": "Mono3DRefer", - "scene_name": "005690", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[627.99, 221.2], [647.42, 221.26], [648.26, 226.07], [626.89, 226.01], [627.99, 168.53], [647.42, 168.53], [648.26, 168.1], [626.89, 168.1]]}, {\"category\": \"car\", \"corners_3d\": [[92.08, 276.61], [147.65, 264.78], [342.89, 263.77], [312.82, 275.33], [92.08, 191.54], [147.65, 189.41], [342.89, 189.23], [312.82, 191.31]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[533.77, 204.09], [541.36, 203.56], [558.15, 203.96], [550.75, 204.49], [533.77, 166.31], [541.36, 166.42], [558.15, 166.34], [550.75, 166.23]]}]\n```" - } - ] - }, - { - "id": 3370, - "data_source": "Mono3DRefer", - "scene_name": "005701", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[562.19, 199.47], [541.6, 199.48], [545.75, 197.64], [564.92, 197.63], [562.19, 179.97], [541.6, 179.98], [545.75, 179.48], [564.92, 179.48]]}, {\"category\": \"car\", \"corners_3d\": [[530.43, 218.67], [493.06, 218.65], [507.44, 213.3], [540.45, 213.31], [530.43, 179.08], [493.06, 179.08], [507.44, 178.35], [540.45, 178.36]]}]\n```" - } - ] - }, - { - "id": 3371, - "data_source": "Mono3DRefer", - "scene_name": "005710", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[562.76, 205.48], [530.21, 205.53], [535.83, 202.31], [565.18, 202.28], [562.76, 171.29], [530.21, 171.28], [535.83, 171.44], [565.18, 171.44]]}, {\"category\": \"car\", \"corners_3d\": [[574.55, 192.2], [557.84, 192.2], [561.43, 191.14], [577.22, 191.15], [574.55, 173.62], [557.84, 173.62], [561.43, 173.57], [577.22, 173.57]]}]\n```" - } - ] - }, - { - "id": 3372, - "data_source": "Mono3DRefer", - "scene_name": "005713", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[597.77, 204.52], [625.62, 204.55], [625.72, 207.46], [595.31, 207.43], [597.77, 177.02], [625.62, 177.02], [625.72, 177.4], [595.31, 177.4]]}]\n```" - } - ] - }, - { - "id": 3373, - "data_source": "Mono3DRefer", - "scene_name": "005713", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[53.14, 270.71], [50.32, 272.41], [-43.89, 277.4], [-39.25, 275.53], [53.14, 196.1], [50.32, 196.51], [-43.89, 197.69], [-39.25, 197.25]]}]\n```" - } - ] - }, - { - "id": 3374, - "data_source": "Mono3DRefer", - "scene_name": "005713", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[597.77, 204.52], [625.62, 204.55], [625.72, 207.46], [595.31, 207.43], [597.77, 177.02], [625.62, 177.02], [625.72, 177.4], [595.31, 177.4]]}, {\"category\": \"cyclist\", \"corners_3d\": [[53.14, 270.71], [50.32, 272.41], [-43.89, 277.4], [-39.25, 275.53], [53.14, 196.1], [50.32, 196.51], [-43.89, 197.69], [-39.25, 197.25]]}]\n```" - } - ] - }, - { - "id": 3375, - "data_source": "Mono3DRefer", - "scene_name": "005716", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[582.47, 193.74], [565.41, 193.6], [578.12, 192.52], [594.37, 192.64], [582.47, 175.74], [565.41, 175.72], [578.12, 175.57], [594.37, 175.58]]}, {\"category\": \"car\", \"corners_3d\": [[467.0, 281.77], [359.19, 281.78], [425.57, 252.84], [504.73, 252.83], [467.0, 168.22], [359.19, 168.22], [425.57, 169.45], [504.73, 169.45]]}, {\"category\": \"car\", \"corners_3d\": [[510.1, 240.05], [454.64, 239.9], [481.55, 229.64], [528.57, 229.75], [510.1, 189.39], [454.64, 189.35], [481.55, 186.83], [528.57, 186.85]]}, {\"category\": \"car\", \"corners_3d\": [[550.73, 211.96], [520.99, 211.95], [529.88, 208.3], [556.85, 208.31], [550.73, 178.88], [520.99, 178.88], [529.88, 178.32], [556.85, 178.32]]}]\n```" - } - ] - }, - { - "id": 3376, - "data_source": "Mono3DRefer", - "scene_name": "005720", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[652.94, 199.53], [675.45, 199.61], [675.47, 201.61], [651.27, 201.53], [652.94, 178.35], [675.45, 178.37], [675.47, 178.78], [651.27, 178.76]]}, {\"category\": \"car\", \"corners_3d\": [[747.74, 251.3], [817.13, 251.53], [871.23, 274.5], [781.45, 274.11], [747.74, 168.89], [817.13, 168.88], [871.23, 167.72], [781.45, 167.74]]}, {\"category\": \"car\", \"corners_3d\": [[806.37, 287.33], [915.37, 289.69], [1022.11, 350.96], [854.44, 345.53], [806.37, 187.48], [915.37, 187.78], [1022.11, 195.6], [854.44, 194.91]]}, {\"category\": \"car\", \"corners_3d\": [[712.45, 223.65], [757.35, 223.81], [775.95, 232.34], [723.49, 232.13], [712.45, 176.96], [757.35, 176.98], [775.95, 177.67], [723.49, 177.65]]}, {\"category\": \"car\", \"corners_3d\": [[701.07, 209.03], [730.69, 209.17], [736.47, 212.95], [703.75, 212.79], [701.07, 182.23], [730.69, 182.26], [736.47, 183.24], [703.75, 183.2]]}]\n```" - } - ] - }, - { - "id": 3377, - "data_source": "Mono3DRefer", - "scene_name": "005724", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[540.65, 308.26], [485.95, 308.27], [500.59, 292.36], [548.48, 292.35], [540.65, 160.06], [485.95, 160.06], [500.59, 162.61], [548.48, 162.61]]}]\n```" - } - ] - }, - { - "id": 3378, - "data_source": "Mono3DRefer", - "scene_name": "005734", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[96.25, 219.43], [134.87, 219.45], [72.75, 225.45], [29.2, 225.42], [96.25, 180.53], [134.87, 180.53], [72.75, 181.52], [29.2, 181.51]]}, {\"category\": \"car\", \"corners_3d\": [[151.16, 217.82], [123.38, 217.75], [199.04, 211.15], [222.83, 211.2], [151.16, 180.46], [123.38, 180.45], [199.04, 179.33], [222.83, 179.34]]}, {\"category\": \"car\", \"corners_3d\": [[522.29, 225.35], [472.06, 225.39], [491.79, 217.4], [534.37, 217.37], [522.29, 183.52], [472.06, 183.53], [491.79, 181.91], [534.37, 181.9]]}, {\"category\": \"car\", \"corners_3d\": [[-950.79, 1113.2], [-1898.03, 1126.22], [25.58, 392.6], [240.11, 391.9], [-950.79, 167.16], [-1898.03, 167.08], [25.58, 171.52], [240.11, 171.53]]}, {\"category\": \"car\", \"corners_3d\": [[592.72, 208.44], [626.06, 208.44], [628.39, 213.64], [590.17, 213.64], [592.72, 175.3], [626.06, 175.3], [628.39, 175.66], [590.17, 175.66]]}]\n```" - } - ] - }, - { - "id": 3379, - "data_source": "Mono3DRefer", - "scene_name": "005752", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[518.54, 237.34], [433.63, 237.42], [483.97, 218.09], [543.44, 218.05], [518.54, 118.98], [433.63, 118.91], [483.97, 135.06], [543.44, 135.09]]}]\n```" - } - ] - }, - { - "id": 3380, - "data_source": "Mono3DRefer", - "scene_name": "005752", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[567.97, 202.04], [546.19, 202.05], [550.36, 199.86], [570.51, 199.85], [567.97, 180.41], [546.19, 180.41], [550.36, 179.84], [570.51, 179.84]]}, {\"category\": \"car\", \"corners_3d\": [[369.73, 200.92], [389.2, 200.92], [377.94, 202.35], [357.49, 202.35], [369.73, 183.26], [389.2, 183.26], [377.94, 183.79], [357.49, 183.79]]}, {\"category\": \"car\", \"corners_3d\": [[587.77, 211.91], [620.26, 211.91], [621.69, 217.46], [584.59, 217.45], [587.77, 179.62], [620.26, 179.62], [621.69, 180.58], [584.59, 180.58]]}]\n```" - } - ] - }, - { - "id": 3381, - "data_source": "Mono3DRefer", - "scene_name": "005752", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[518.54, 237.34], [433.63, 237.42], [483.97, 218.09], [543.44, 218.05], [518.54, 118.98], [433.63, 118.91], [483.97, 135.06], [543.44, 135.09]]}, {\"category\": \"car\", \"corners_3d\": [[567.97, 202.04], [546.19, 202.05], [550.36, 199.86], [570.51, 199.85], [567.97, 180.41], [546.19, 180.41], [550.36, 179.84], [570.51, 179.84]]}, {\"category\": \"car\", \"corners_3d\": [[369.73, 200.92], [389.2, 200.92], [377.94, 202.35], [357.49, 202.35], [369.73, 183.26], [389.2, 183.26], [377.94, 183.79], [357.49, 183.79]]}, {\"category\": \"car\", \"corners_3d\": [[587.77, 211.91], [620.26, 211.91], [621.69, 217.46], [584.59, 217.45], [587.77, 179.62], [620.26, 179.62], [621.69, 180.58], [584.59, 180.58]]}]\n```" - } - ] - }, - { - "id": 3382, - "data_source": "Mono3DRefer", - "scene_name": "005753", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[641.17, 195.64], [618.92, 195.52], [627.36, 193.84], [647.99, 193.94], [641.17, 176.33], [618.92, 176.32], [627.36, 176.06], [647.99, 176.08]]}]\n```" - } - ] - }, - { - "id": 3383, - "data_source": "Mono3DRefer", - "scene_name": "005768", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[963.78, 295.1], [1061.62, 319.74], [676.12, 335.85], [639.96, 306.06], [963.78, 188.94], [1061.62, 192.18], [676.12, 194.3], [639.96, 190.38]]}]\n```" - } - ] - }, - { - "id": 3384, - "data_source": "Mono3DRefer", - "scene_name": "005774", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[381.98, 310.31], [244.57, 310.59], [365.26, 264.05], [456.14, 263.93], [381.98, 174.46], [244.57, 174.46], [365.26, 173.92], [456.14, 173.92]]}, {\"category\": \"car\", \"corners_3d\": [[203.32, 209.92], [236.52, 209.92], [202.84, 213.26], [166.65, 213.26], [203.32, 179.8], [236.52, 179.8], [202.84, 180.42], [166.65, 180.42]]}, {\"category\": \"car\", \"corners_3d\": [[527.47, 215.1], [487.3, 215.08], [503.57, 209.75], [538.67, 209.77], [527.47, 178.16], [487.3, 178.16], [503.57, 177.49], [538.67, 177.49]]}, {\"category\": \"car\", \"corners_3d\": [[337.83, 196.34], [357.98, 196.36], [340.58, 197.86], [319.17, 197.84], [337.83, 175.96], [357.98, 175.97], [340.58, 176.17], [319.17, 176.16]]}, {\"category\": \"car\", \"corners_3d\": [[356.91, 191.07], [341.64, 191.05], [367.01, 189.49], [380.99, 189.51], [356.91, 170.78], [341.64, 170.78], [367.01, 170.96], [380.99, 170.96]]}, {\"category\": \"car\", \"corners_3d\": [[575.84, 190.93], [558.56, 190.93], [561.07, 189.88], [577.34, 189.88], [575.84, 176.53], [558.56, 176.53], [561.07, 176.31], [577.34, 176.31]]}, {\"category\": \"car\", \"corners_3d\": [[559.6, 200.14], [534.26, 200.15], [539.91, 197.87], [563.12, 197.86], [559.6, 173.92], [534.26, 173.92], [539.91, 173.83], [563.12, 173.83]]}, {\"category\": \"car\", \"corners_3d\": [[590.13, 206.61], [622.51, 206.64], [622.22, 211.42], [585.26, 211.38], [590.13, 174.44], [622.51, 174.44], [622.22, 174.66], [585.26, 174.66]]}]\n```" - } - ] - }, - { - "id": 3385, - "data_source": "Mono3DRefer", - "scene_name": "005778", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[576.83, 216.17], [592.25, 216.13], [593.66, 219.47], [577.05, 219.53], [576.83, 174.31], [592.25, 174.31], [593.66, 174.43], [577.05, 174.43]]}]\n```" - } - ] - }, - { - "id": 3386, - "data_source": "Mono3DRefer", - "scene_name": "005778", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[823.61, 307.88], [777.73, 285.78], [1065.13, 281.56], [1162.73, 301.89], [823.61, 151.72], [777.73, 155.18], [1065.13, 155.84], [1162.73, 152.66]]}, {\"category\": \"car\", \"corners_3d\": [[-35.87, 302.72], [43.94, 285.47], [286.98, 283.24], [245.36, 299.76], [-35.87, 201.22], [43.94, 197.45], [286.98, 196.96], [245.36, 200.57]]}, {\"category\": \"car\", \"corners_3d\": [[214.81, 251.63], [241.0, 244.9], [379.36, 243.31], [366.41, 249.73], [214.81, 192.12], [241.0, 190.47], [379.36, 190.08], [366.41, 191.65]]}, {\"category\": \"car\", \"corners_3d\": [[361.94, 219.74], [370.61, 217.4], [456.42, 216.88], [452.3, 219.16], [361.94, 184.94], [370.61, 184.34], [456.42, 184.2], [452.3, 184.79]]}, {\"category\": \"car\", \"corners_3d\": [[1556.46, 371.24], [1968.85, 453.86], [1256.39, 472.66], [1044.08, 380.43], [1556.46, 180.49], [1968.85, 183.67], [1256.39, 184.39], [1044.08, 180.84]]}]\n```" - } - ] - }, - { - "id": 3387, - "data_source": "Mono3DRefer", - "scene_name": "005778", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[823.61, 307.88], [777.73, 285.78], [1065.13, 281.56], [1162.73, 301.89], [823.61, 151.72], [777.73, 155.18], [1065.13, 155.84], [1162.73, 152.66]]}, {\"category\": \"cyclist\", \"corners_3d\": [[576.83, 216.17], [592.25, 216.13], [593.66, 219.47], [577.05, 219.53], [576.83, 174.31], [592.25, 174.31], [593.66, 174.43], [577.05, 174.43]]}, {\"category\": \"car\", \"corners_3d\": [[-35.87, 302.72], [43.94, 285.47], [286.98, 283.24], [245.36, 299.76], [-35.87, 201.22], [43.94, 197.45], [286.98, 196.96], [245.36, 200.57]]}, {\"category\": \"car\", \"corners_3d\": [[214.81, 251.63], [241.0, 244.9], [379.36, 243.31], [366.41, 249.73], [214.81, 192.12], [241.0, 190.47], [379.36, 190.08], [366.41, 191.65]]}, {\"category\": \"car\", \"corners_3d\": [[361.94, 219.74], [370.61, 217.4], [456.42, 216.88], [452.3, 219.16], [361.94, 184.94], [370.61, 184.34], [456.42, 184.2], [452.3, 184.79]]}, {\"category\": \"car\", \"corners_3d\": [[1556.46, 371.24], [1968.85, 453.86], [1256.39, 472.66], [1044.08, 380.43], [1556.46, 180.49], [1968.85, 183.67], [1256.39, 184.39], [1044.08, 180.84]]}]\n```" - } - ] - }, - { - "id": 3388, - "data_source": "Mono3DRefer", - "scene_name": "005788", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[695.0, 206.12], [719.54, 206.37], [715.35, 209.07], [688.88, 208.78], [695.0, 183.14], [719.54, 183.22], [715.35, 184.06], [688.88, 183.97]]}, {\"category\": \"car\", \"corners_3d\": [[204.64, 307.14], [107.84, 304.61], [257.54, 271.49], [331.8, 272.9], [204.64, 201.85], [107.84, 201.31], [257.54, 194.15], [331.8, 194.46]]}, {\"category\": \"car\", \"corners_3d\": [[350.06, 269.47], [279.9, 268.09], [368.31, 248.54], [424.93, 249.42], [350.06, 177.6], [279.9, 177.53], [368.31, 176.57], [424.93, 176.61]]}, {\"category\": \"car\", \"corners_3d\": [[716.23, 232.07], [768.99, 232.25], [795.07, 244.16], [731.66, 243.91], [716.23, 182.56], [768.99, 182.59], [795.07, 184.54], [731.66, 184.5]]}, {\"category\": \"car\", \"corners_3d\": [[673.84, 205.26], [651.0, 204.96], [662.3, 202.87], [683.74, 203.13], [673.84, 183.79], [651.0, 183.69], [662.3, 182.99], [683.74, 183.07]]}, {\"category\": \"car\", \"corners_3d\": [[722.54, 224.32], [762.31, 224.71], [770.37, 232.99], [724.24, 232.46], [722.54, 171.02], [762.31, 171.01], [770.37, 170.72], [724.24, 170.74]]}]\n```" - } - ] - }, - { - "id": 3389, - "data_source": "Mono3DRefer", - "scene_name": "005805", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[652.27, 205.17], [676.71, 205.17], [683.57, 208.51], [656.61, 208.5], [652.27, 183.37], [676.71, 183.37], [683.57, 184.45], [656.61, 184.45]]}]\n```" - } - ] - }, - { - "id": 3390, - "data_source": "Mono3DRefer", - "scene_name": "005809", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[335.79, 375.83], [137.73, 374.75], [341.46, 290.06], [456.82, 290.42], [335.79, 203.86], [137.73, 203.7], [341.46, 190.76], [456.82, 190.81]]}]\n```" - } - ] - }, - { - "id": 3391, - "data_source": "Mono3DRefer", - "scene_name": "005817", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[958.17, 324.54], [935.53, 311.66], [1024.38, 313.86], [1056.34, 327.2], [958.17, 143.79], [935.53, 147.07], [1024.38, 146.51], [1056.34, 143.11]]}]\n```" - } - ] - }, - { - "id": 3392, - "data_source": "Mono3DRefer", - "scene_name": "005821", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[198.48, 430.99], [-23.69, 434.16], [237.91, 320.94], [362.68, 319.92], [198.48, 229.37], [-23.69, 230.06], [237.91, 205.28], [362.68, 205.05]]}, {\"category\": \"car\", \"corners_3d\": [[545.71, 222.44], [499.78, 222.44], [514.34, 215.83], [554.14, 215.83], [545.71, 184.13], [499.78, 184.13], [514.34, 182.62], [554.14, 182.62]]}]\n```" - } - ] - }, - { - "id": 3393, - "data_source": "Mono3DRefer", - "scene_name": "005821", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[418.24, 293.91], [305.45, 294.11], [387.09, 260.71], [468.75, 260.6], [418.24, 171.48], [305.45, 171.47], [387.09, 171.85], [468.75, 171.86]]}]\n```" - } - ] - }, - { - "id": 3394, - "data_source": "Mono3DRefer", - "scene_name": "005821", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[198.48, 430.99], [-23.69, 434.16], [237.91, 320.94], [362.68, 319.92], [198.48, 229.37], [-23.69, 230.06], [237.91, 205.28], [362.68, 205.05]]}, {\"category\": \"van\", \"corners_3d\": [[418.24, 293.91], [305.45, 294.11], [387.09, 260.71], [468.75, 260.6], [418.24, 171.48], [305.45, 171.47], [387.09, 171.85], [468.75, 171.86]]}, {\"category\": \"car\", \"corners_3d\": [[545.71, 222.44], [499.78, 222.44], [514.34, 215.83], [554.14, 215.83], [545.71, 184.13], [499.78, 184.13], [514.34, 182.62], [554.14, 182.62]]}]\n```" - } - ] - }, - { - "id": 3395, - "data_source": "Mono3DRefer", - "scene_name": "005823", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[550.28, 211.13], [520.05, 211.15], [528.07, 207.39], [555.34, 207.38], [550.28, 182.52], [520.05, 182.52], [528.07, 181.57], [555.34, 181.57]]}, {\"category\": \"car\", \"corners_3d\": [[443.28, 275.3], [357.93, 275.06], [419.69, 251.29], [485.25, 251.43], [443.28, 184.4], [357.93, 184.38], [419.69, 181.7], [485.25, 181.71]]}]\n```" - } - ] - }, - { - "id": 3396, - "data_source": "Mono3DRefer", - "scene_name": "005823", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[410.53, 214.25], [400.31, 214.2], [410.61, 212.6], [420.45, 212.64], [410.53, 183.15], [400.31, 183.14], [410.61, 182.74], [420.45, 182.75]]}]\n```" - } - ] - }, - { - "id": 3397, - "data_source": "Mono3DRefer", - "scene_name": "005823", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[550.28, 211.13], [520.05, 211.15], [528.07, 207.39], [555.34, 207.38], [550.28, 182.52], [520.05, 182.52], [528.07, 181.57], [555.34, 181.57]]}, {\"category\": \"car\", \"corners_3d\": [[443.28, 275.3], [357.93, 275.06], [419.69, 251.29], [485.25, 251.43], [443.28, 184.4], [357.93, 184.38], [419.69, 181.7], [485.25, 181.71]]}, {\"category\": \"cyclist\", \"corners_3d\": [[410.53, 214.25], [400.31, 214.2], [410.61, 212.6], [420.45, 212.64], [410.53, 183.15], [400.31, 183.14], [410.61, 182.74], [420.45, 182.75]]}]\n```" - } - ] - }, - { - "id": 3398, - "data_source": "Mono3DRefer", - "scene_name": "005829", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1003.93, 296.75], [860.55, 298.83], [753.92, 255.42], [849.59, 254.52], [1003.93, 142.7], [860.55, 142.2], [753.92, 152.76], [849.59, 152.98]]}, {\"category\": \"car\", \"corners_3d\": [[362.16, 314.41], [245.35, 315.55], [344.57, 272.32], [425.63, 271.76], [362.16, 203.44], [245.35, 203.69], [344.57, 194.35], [425.63, 194.22]]}, {\"category\": \"car\", \"corners_3d\": [[476.27, 242.98], [413.51, 243.41], [437.62, 231.67], [489.86, 231.37], [476.27, 192.08], [413.51, 192.2], [437.62, 188.98], [489.86, 188.9]]}, {\"category\": \"car\", \"corners_3d\": [[503.64, 225.4], [453.89, 225.58], [470.49, 218.23], [513.27, 218.09], [503.64, 176.73], [453.89, 176.74], [470.49, 176.2], [513.27, 176.19]]}, {\"category\": \"car\", \"corners_3d\": [[592.09, 198.52], [617.54, 198.47], [621.53, 200.48], [594.09, 200.55], [592.09, 174.18], [617.54, 174.18], [621.53, 174.28], [594.09, 174.29]]}]\n```" - } - ] - }, - { - "id": 3399, - "data_source": "Mono3DRefer", - "scene_name": "005829", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[586.04, 198.15], [592.86, 197.96], [597.27, 198.16], [590.43, 198.35], [586.04, 169.23], [592.86, 169.26], [597.27, 169.23], [590.43, 169.2]]}]\n```" - } - ] - }, - { - "id": 3400, - "data_source": "Mono3DRefer", - "scene_name": "005829", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1003.93, 296.75], [860.55, 298.83], [753.92, 255.42], [849.59, 254.52], [1003.93, 142.7], [860.55, 142.2], [753.92, 152.76], [849.59, 152.98]]}, {\"category\": \"car\", \"corners_3d\": [[362.16, 314.41], [245.35, 315.55], [344.57, 272.32], [425.63, 271.76], [362.16, 203.44], [245.35, 203.69], [344.57, 194.35], [425.63, 194.22]]}, {\"category\": \"car\", \"corners_3d\": [[476.27, 242.98], [413.51, 243.41], [437.62, 231.67], [489.86, 231.37], [476.27, 192.08], [413.51, 192.2], [437.62, 188.98], [489.86, 188.9]]}, {\"category\": \"car\", \"corners_3d\": [[503.64, 225.4], [453.89, 225.58], [470.49, 218.23], [513.27, 218.09], [503.64, 176.73], [453.89, 176.74], [470.49, 176.2], [513.27, 176.19]]}, {\"category\": \"car\", \"corners_3d\": [[592.09, 198.52], [617.54, 198.47], [621.53, 200.48], [594.09, 200.55], [592.09, 174.18], [617.54, 174.18], [621.53, 174.28], [594.09, 174.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[586.04, 198.15], [592.86, 197.96], [597.27, 198.16], [590.43, 198.35], [586.04, 169.23], [592.86, 169.26], [597.27, 169.23], [590.43, 169.2]]}]\n```" - } - ] - }, - { - "id": 3401, - "data_source": "Mono3DRefer", - "scene_name": "005833", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[377.32, 197.38], [394.57, 197.4], [381.8, 198.72], [363.63, 198.7], [377.32, 179.95], [394.57, 179.96], [381.8, 180.34], [363.63, 180.34]]}, {\"category\": \"car\", \"corners_3d\": [[398.2, 195.71], [385.18, 195.71], [402.1, 194.03], [414.17, 194.04], [398.2, 178.62], [385.18, 178.62], [402.1, 178.2], [414.17, 178.2]]}, {\"category\": \"car\", \"corners_3d\": [[572.15, 194.89], [553.47, 194.89], [556.52, 193.5], [574.02, 193.5], [572.15, 175.55], [553.47, 175.55], [556.52, 175.38], [574.02, 175.38]]}, {\"category\": \"car\", \"corners_3d\": [[589.67, 208.14], [621.98, 208.14], [623.66, 213.12], [586.78, 213.12], [589.67, 176.02], [621.98, 176.02], [623.66, 176.47], [586.78, 176.47]]}, {\"category\": \"car\", \"corners_3d\": [[527.83, 223.23], [480.2, 223.27], [498.51, 215.68], [538.96, 215.65], [527.83, 175.98], [480.2, 175.98], [498.51, 175.51], [538.96, 175.51]]}, {\"category\": \"car\", \"corners_3d\": [[285.48, 205.91], [312.04, 205.91], [290.93, 208.25], [262.48, 208.25], [285.48, 181.81], [312.04, 181.81], [290.93, 182.44], [262.48, 182.44]]}, {\"category\": \"car\", \"corners_3d\": [[556.78, 202.95], [531.48, 202.94], [538.54, 200.43], [561.73, 200.44], [556.78, 179.7], [531.48, 179.69], [538.54, 179.12], [561.73, 179.13]]}]\n```" - } - ] - }, - { - "id": 3402, - "data_source": "Mono3DRefer", - "scene_name": "005835", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[341.38, 308.83], [314.13, 321.46], [225.73, 321.1], [260.82, 308.53], [341.38, 152.95], [314.13, 150.24], [225.73, 150.31], [260.82, 153.01]]}]\n```" - } - ] - }, - { - "id": 3403, - "data_source": "Mono3DRefer", - "scene_name": "005842", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1170.16, 237.22], [1245.93, 246.27], [946.14, 245.38], [907.74, 236.54], [1170.16, 151.17], [1245.93, 148.13], [946.14, 148.43], [907.74, 151.41]]}, {\"category\": \"car\", \"corners_3d\": [[616.2, 205.66], [578.17, 205.52], [587.91, 201.93], [621.79, 202.04], [616.2, 169.2], [578.17, 169.21], [587.91, 169.62], [621.79, 169.6]]}, {\"category\": \"car\", \"corners_3d\": [[648.46, 183.85], [628.65, 183.82], [631.99, 183.13], [650.55, 183.16], [648.46, 166.87], [628.65, 166.89], [631.99, 167.26], [650.55, 167.25]]}, {\"category\": \"car\", \"corners_3d\": [[545.94, 222.16], [595.42, 222.54], [574.01, 233.81], [513.63, 233.23], [545.94, 171.7], [595.42, 171.69], [574.01, 171.43], [513.63, 171.45]]}, {\"category\": \"car\", \"corners_3d\": [[-7.04, 309.09], [60.84, 300.12], [125.27, 311.72], [55.71, 322.47], [-7.04, 236.54], [60.84, 232.34], [125.27, 237.77], [55.71, 242.79]]}, {\"category\": \"car\", \"corners_3d\": [[1443.97, 260.09], [1619.4, 278.92], [1229.33, 277.27], [1124.51, 258.97], [1443.97, 143.74], [1619.4, 137.45], [1229.33, 138.0], [1124.51, 144.11]]}]\n```" - } - ] - }, - { - "id": 3404, - "data_source": "Mono3DRefer", - "scene_name": "005845", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[198.54, 228.64], [234.85, 228.45], [191.46, 235.88], [150.12, 236.12], [198.54, 199.3], [234.85, 199.21], [191.46, 202.73], [150.12, 202.84]]}, {\"category\": \"car\", \"corners_3d\": [[139.57, 229.81], [176.51, 229.62], [125.39, 237.2], [83.3, 237.45], [139.57, 198.21], [176.51, 198.13], [125.39, 201.5], [83.3, 201.61]]}]\n```" - } - ] - }, - { - "id": 3405, - "data_source": "Mono3DRefer", - "scene_name": "005847", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[599.97, 216.45], [616.52, 216.44], [617.66, 220.08], [599.73, 220.09], [599.97, 171.53], [616.52, 171.53], [617.66, 171.42], [599.73, 171.42]]}]\n```" - } - ] - }, - { - "id": 3406, - "data_source": "Mono3DRefer", - "scene_name": "005847", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1296.46, 527.25], [1042.13, 401.53], [1663.17, 393.13], [2225.67, 507.47], [1296.46, 90.06], [1042.13, 119.43], [1663.17, 121.39], [2225.67, 94.68]]}, {\"category\": \"car\", \"corners_3d\": [[-593.32, 426.18], [-319.83, 367.83], [96.63, 365.46], [-49.87, 422.2], [-593.32, 228.19], [-319.83, 215.44], [96.63, 214.92], [-49.87, 227.32]]}, {\"category\": \"car\", \"corners_3d\": [[92.77, 286.39], [147.16, 273.11], [334.47, 271.33], [305.49, 284.11], [92.77, 202.4], [147.16, 198.95], [334.47, 198.48], [305.49, 201.81]]}, {\"category\": \"car\", \"corners_3d\": [[344.19, 228.37], [358.21, 225.07], [459.52, 224.77], [451.97, 228.03], [344.19, 186.5], [358.21, 185.69], [459.52, 185.62], [451.97, 186.42]]}]\n```" - } - ] - }, - { - "id": 3407, - "data_source": "Mono3DRefer", - "scene_name": "005847", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1296.46, 527.25], [1042.13, 401.53], [1663.17, 393.13], [2225.67, 507.47], [1296.46, 90.06], [1042.13, 119.43], [1663.17, 121.39], [2225.67, 94.68]]}, {\"category\": \"cyclist\", \"corners_3d\": [[599.97, 216.45], [616.52, 216.44], [617.66, 220.08], [599.73, 220.09], [599.97, 171.53], [616.52, 171.53], [617.66, 171.42], [599.73, 171.42]]}, {\"category\": \"car\", \"corners_3d\": [[-593.32, 426.18], [-319.83, 367.83], [96.63, 365.46], [-49.87, 422.2], [-593.32, 228.19], [-319.83, 215.44], [96.63, 214.92], [-49.87, 227.32]]}, {\"category\": \"car\", \"corners_3d\": [[92.77, 286.39], [147.16, 273.11], [334.47, 271.33], [305.49, 284.11], [92.77, 202.4], [147.16, 198.95], [334.47, 198.48], [305.49, 201.81]]}, {\"category\": \"car\", \"corners_3d\": [[344.19, 228.37], [358.21, 225.07], [459.52, 224.77], [451.97, 228.03], [344.19, 186.5], [358.21, 185.69], [459.52, 185.62], [451.97, 186.42]]}]\n```" - } - ] - }, - { - "id": 3408, - "data_source": "Mono3DRefer", - "scene_name": "005851", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[429.52, 229.52], [390.7, 230.4], [390.26, 225.32], [425.71, 224.59], [429.52, 194.5], [390.7, 194.83], [390.26, 192.89], [425.71, 192.61]]}]\n```" - } - ] - }, - { - "id": 3409, - "data_source": "Mono3DRefer", - "scene_name": "005854", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[598.39, 230.9], [622.35, 230.76], [627.84, 235.82], [601.79, 235.98], [598.39, 203.79], [622.35, 203.72], [627.84, 206.41], [601.79, 206.5]]}]\n```" - } - ] - }, - { - "id": 3410, - "data_source": "Mono3DRefer", - "scene_name": "005854", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[469.85, 303.8], [342.97, 304.28], [417.8, 265.19], [506.85, 264.95], [469.85, 198.74], [342.97, 198.83], [417.8, 191.1], [506.85, 191.06]]}]\n```" - } - ] - }, - { - "id": 3411, - "data_source": "Mono3DRefer", - "scene_name": "005854", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[469.85, 303.8], [342.97, 304.28], [417.8, 265.19], [506.85, 264.95], [469.85, 198.74], [342.97, 198.83], [417.8, 191.1], [506.85, 191.06]]}, {\"category\": \"van\", \"corners_3d\": [[598.39, 230.9], [622.35, 230.76], [627.84, 235.82], [601.79, 235.98], [598.39, 203.79], [622.35, 203.72], [627.84, 206.41], [601.79, 206.5]]}]\n```" - } - ] - }, - { - "id": 3412, - "data_source": "Mono3DRefer", - "scene_name": "005860", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[662.81, 200.16], [685.52, 200.27], [684.09, 202.34], [659.68, 202.21], [662.81, 178.81], [685.52, 178.83], [684.09, 179.29], [659.68, 179.26]]}, {\"category\": \"car\", \"corners_3d\": [[740.7, 241.22], [800.61, 241.45], [839.69, 258.04], [765.19, 257.69], [740.7, 170.2], [800.61, 170.19], [839.69, 169.54], [765.19, 169.56]]}, {\"category\": \"car\", \"corners_3d\": [[778.69, 264.7], [865.33, 266.22], [923.53, 301.61], [803.47, 298.75], [778.69, 184.59], [865.33, 184.78], [923.53, 189.3], [803.47, 188.93]]}, {\"category\": \"car\", \"corners_3d\": [[716.34, 219.56], [757.29, 219.74], [771.77, 226.77], [724.64, 226.53], [716.34, 177.09], [757.29, 177.11], [771.77, 177.75], [724.64, 177.72]]}, {\"category\": \"car\", \"corners_3d\": [[708.18, 207.24], [735.94, 207.38], [740.4, 210.72], [709.95, 210.55], [708.18, 182.16], [735.94, 182.2], [740.4, 183.1], [709.95, 183.06]]}]\n```" - } - ] - }, - { - "id": 3413, - "data_source": "Mono3DRefer", - "scene_name": "005871", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[695.86, 208.14], [720.36, 208.4], [716.24, 211.26], [689.83, 210.95], [695.86, 185.21], [720.36, 185.3], [716.24, 186.3], [689.83, 186.19]]}, {\"category\": \"car\", \"corners_3d\": [[117.02, 331.36], [3.41, 327.8], [203.73, 283.45], [287.56, 285.25], [117.02, 205.18], [3.41, 204.45], [203.73, 195.41], [287.56, 195.78]]}, {\"category\": \"car\", \"corners_3d\": [[307.56, 282.16], [228.09, 280.5], [335.99, 256.06], [398.48, 257.05], [307.56, 177.71], [228.09, 177.63], [335.99, 176.55], [398.48, 176.59]]}, {\"category\": \"car\", \"corners_3d\": [[719.18, 238.61], [776.28, 238.82], [806.58, 253.38], [736.81, 253.06], [719.18, 185.05], [776.28, 185.08], [806.58, 187.78], [736.81, 187.73]]}, {\"category\": \"car\", \"corners_3d\": [[665.68, 210.85], [640.87, 210.51], [652.21, 207.84], [675.35, 208.13], [665.68, 187.59], [640.87, 187.46], [652.21, 186.43], [675.35, 186.54]]}, {\"category\": \"car\", \"corners_3d\": [[721.93, 229.92], [764.26, 230.31], [775.82, 240.22], [726.17, 239.67], [721.93, 173.12], [764.26, 173.12], [775.82, 173.17], [726.17, 173.17]]}]\n```" - } - ] - }, - { - "id": 3414, - "data_source": "Mono3DRefer", - "scene_name": "005884", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[581.84, 210.24], [535.94, 210.12], [550.14, 205.27], [590.09, 205.36], [581.84, 166.25], [535.94, 166.27], [550.14, 167.12], [590.09, 167.11]]}, {\"category\": \"car\", \"corners_3d\": [[619.22, 184.52], [597.52, 184.5], [601.31, 183.69], [621.51, 183.71], [619.22, 165.9], [597.52, 165.91], [601.31, 166.39], [621.51, 166.38]]}, {\"category\": \"car\", \"corners_3d\": [[492.45, 232.6], [555.62, 233.03], [520.77, 251.7], [438.5, 250.96], [492.45, 168.1], [555.62, 168.07], [520.77, 166.59], [438.5, 166.64]]}, {\"category\": \"car\", \"corners_3d\": [[1513.77, 277.92], [1756.26, 305.66], [1210.42, 307.73], [1080.34, 279.22], [1513.77, 132.17], [1756.26, 121.42], [1210.42, 120.62], [1080.34, 131.66]]}]\n```" - } - ] - }, - { - "id": 3415, - "data_source": "Mono3DRefer", - "scene_name": "005885", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[667.66, 229.26], [726.88, 229.17], [766.81, 246.31], [689.61, 246.47], [667.66, 157.84], [726.88, 157.86], [766.81, 153.3], [689.61, 153.26]]}]\n```" - } - ] - }, - { - "id": 3416, - "data_source": "Mono3DRefer", - "scene_name": "005885", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[765.54, 292.43], [870.39, 292.45], [1013.53, 358.22], [851.0, 358.19], [765.54, 190.97], [870.39, 190.97], [1013.53, 200.94], [851.0, 200.94]]}, {\"category\": \"car\", \"corners_3d\": [[707.95, 252.15], [783.29, 251.99], [842.1, 276.8], [743.23, 277.08], [707.95, 181.4], [783.29, 181.38], [842.1, 184.06], [743.23, 184.09]]}, {\"category\": \"car\", \"corners_3d\": [[541.09, 226.95], [496.36, 227.18], [504.83, 219.77], [543.44, 219.59], [541.09, 182.01], [496.36, 182.05], [504.83, 180.79], [543.44, 180.76]]}, {\"category\": \"car\", \"corners_3d\": [[654.96, 212.97], [688.46, 212.93], [699.64, 217.77], [662.11, 217.81], [654.96, 182.77], [688.46, 182.76], [699.64, 183.96], [662.11, 183.97]]}, {\"category\": \"car\", \"corners_3d\": [[557.76, 207.94], [526.84, 207.98], [532.87, 204.48], [560.7, 204.44], [557.76, 182.93], [526.84, 182.94], [532.87, 181.94], [560.7, 181.92]]}]\n```" - } - ] - }, - { - "id": 3417, - "data_source": "Mono3DRefer", - "scene_name": "005885", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[765.54, 292.43], [870.39, 292.45], [1013.53, 358.22], [851.0, 358.19], [765.54, 190.97], [870.39, 190.97], [1013.53, 200.94], [851.0, 200.94]]}, {\"category\": \"car\", \"corners_3d\": [[707.95, 252.15], [783.29, 251.99], [842.1, 276.8], [743.23, 277.08], [707.95, 181.4], [783.29, 181.38], [842.1, 184.06], [743.23, 184.09]]}, {\"category\": \"van\", \"corners_3d\": [[667.66, 229.26], [726.88, 229.17], [766.81, 246.31], [689.61, 246.47], [667.66, 157.84], [726.88, 157.86], [766.81, 153.3], [689.61, 153.26]]}, {\"category\": \"car\", \"corners_3d\": [[541.09, 226.95], [496.36, 227.18], [504.83, 219.77], [543.44, 219.59], [541.09, 182.01], [496.36, 182.05], [504.83, 180.79], [543.44, 180.76]]}, {\"category\": \"car\", \"corners_3d\": [[654.96, 212.97], [688.46, 212.93], [699.64, 217.77], [662.11, 217.81], [654.96, 182.77], [688.46, 182.76], [699.64, 183.96], [662.11, 183.97]]}, {\"category\": \"car\", \"corners_3d\": [[557.76, 207.94], [526.84, 207.98], [532.87, 204.48], [560.7, 204.44], [557.76, 182.93], [526.84, 182.94], [532.87, 181.94], [560.7, 181.92]]}]\n```" - } - ] - }, - { - "id": 3418, - "data_source": "Mono3DRefer", - "scene_name": "005888", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[861.44, 318.36], [839.43, 306.18], [920.8, 306.35], [950.74, 318.56], [861.44, 141.73], [839.43, 145.15], [920.8, 145.11], [950.74, 141.67]]}]\n```" - } - ] - }, - { - "id": 3419, - "data_source": "Mono3DRefer", - "scene_name": "005899", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[559.41, 329.58], [622.83, 329.72], [625.08, 359.95], [548.83, 359.74], [559.41, 149.12], [622.83, 149.09], [625.08, 142.73], [548.83, 142.77]]}]\n```" - } - ] - }, - { - "id": 3420, - "data_source": "Mono3DRefer", - "scene_name": "005901", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[706.46, 258.69], [779.45, 257.72], [855.76, 283.73], [761.21, 285.39], [706.46, 188.31], [779.45, 188.14], [855.76, 192.82], [761.21, 193.12]]}, {\"category\": \"car\", \"corners_3d\": [[-147.77, 330.03], [-16.12, 303.25], [281.45, 304.44], [210.11, 331.75], [-147.77, 199.93], [-16.12, 195.31], [281.45, 195.52], [210.11, 200.22]]}, {\"category\": \"car\", \"corners_3d\": [[337.81, 280.3], [289.96, 296.25], [51.16, 293.29], [128.24, 278.05], [337.81, 188.01], [289.96, 190.26], [51.16, 189.84], [128.24, 187.69]]}, {\"category\": \"car\", \"corners_3d\": [[116.9, 275.75], [172.48, 264.11], [365.33, 264.07], [334.36, 275.7], [116.9, 193.53], [172.48, 191.19], [365.33, 191.19], [334.36, 193.52]]}, {\"category\": \"car\", \"corners_3d\": [[208.68, 260.34], [250.14, 251.93], [407.2, 252.66], [382.21, 261.24], [208.68, 188.79], [250.14, 187.26], [407.2, 187.39], [382.21, 188.95]]}, {\"category\": \"car\", \"corners_3d\": [[652.16, 225.31], [699.41, 225.07], [727.53, 235.57], [670.88, 235.92], [652.16, 183.69], [699.41, 183.64], [727.53, 185.81], [670.88, 185.88]]}, {\"category\": \"car\", \"corners_3d\": [[503.87, 223.67], [495.58, 226.98], [378.53, 226.71], [393.91, 223.43], [503.87, 184.35], [495.58, 185.09], [378.53, 185.03], [393.91, 184.29]]}, {\"category\": \"car\", \"corners_3d\": [[644.71, 209.6], [676.73, 209.5], [690.9, 214.22], [654.77, 214.35], [644.71, 182.04], [676.73, 182.01], [690.9, 183.19], [654.77, 183.22]]}, {\"category\": \"car\", \"corners_3d\": [[391.93, 222.5], [402.19, 219.98], [490.65, 219.86], [485.14, 222.36], [391.93, 183.14], [402.19, 182.61], [490.65, 182.59], [485.14, 183.11]]}, {\"category\": \"car\", \"corners_3d\": [[409.15, 215.23], [418.84, 213.1], [491.56, 213.05], [485.74, 215.18], [409.15, 174.75], [418.84, 174.65], [491.56, 174.65], [485.74, 174.75]]}, {\"category\": \"car\", \"corners_3d\": [[630.32, 198.19], [652.72, 198.16], [658.24, 200.34], [633.92, 200.36], [630.32, 175.08], [652.72, 175.08], [658.24, 175.27], [633.92, 175.27]]}]\n```" - } - ] - }, - { - "id": 3421, - "data_source": "Mono3DRefer", - "scene_name": "005920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[597.49, 220.96], [644.5, 220.96], [649.65, 228.17], [595.6, 228.16], [597.49, 183.13], [644.5, 183.14], [649.65, 184.68], [595.6, 184.67]]}, {\"category\": \"car\", \"corners_3d\": [[447.17, 203.59], [423.18, 203.55], [438.21, 201.4], [460.54, 201.43], [447.17, 181.17], [423.18, 181.16], [438.21, 180.58], [460.54, 180.59]]}, {\"category\": \"car\", \"corners_3d\": [[584.96, 210.6], [619.04, 210.62], [619.23, 214.92], [581.27, 214.89], [584.96, 179.58], [619.04, 179.58], [619.23, 180.35], [581.27, 180.35]]}]\n```" - } - ] - }, - { - "id": 3422, - "data_source": "Mono3DRefer", - "scene_name": "005920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[728.09, 206.93], [726.54, 206.28], [759.14, 206.36], [761.34, 207.01], [728.09, 179.4], [726.54, 179.28], [759.14, 179.29], [761.34, 179.42]]}]\n```" - } - ] - }, - { - "id": 3423, - "data_source": "Mono3DRefer", - "scene_name": "005920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[632.7, 202.23], [657.35, 202.26], [659.61, 204.85], [632.78, 204.82], [632.7, 174.17], [657.35, 174.17], [659.61, 174.29], [632.78, 174.29]]}]\n```" - } - ] - }, - { - "id": 3424, - "data_source": "Mono3DRefer", - "scene_name": "005920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[801.56, 215.05], [788.97, 215.06], [783.78, 213.94], [796.04, 213.92], [801.56, 175.73], [788.97, 175.73], [783.78, 175.66], [796.04, 175.66]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.12, 202.38], [553.75, 202.79], [543.37, 202.85], [542.88, 202.44], [553.12, 173.01], [553.75, 173.01], [543.37, 173.01], [542.88, 173.01]]}]\n```" - } - ] - }, - { - "id": 3425, - "data_source": "Mono3DRefer", - "scene_name": "005920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[597.49, 220.96], [644.5, 220.96], [649.65, 228.17], [595.6, 228.16], [597.49, 183.13], [644.5, 183.14], [649.65, 184.68], [595.6, 184.67]]}, {\"category\": \"car\", \"corners_3d\": [[447.17, 203.59], [423.18, 203.55], [438.21, 201.4], [460.54, 201.43], [447.17, 181.17], [423.18, 181.16], [438.21, 180.58], [460.54, 180.59]]}, {\"category\": \"cyclist\", \"corners_3d\": [[728.09, 206.93], [726.54, 206.28], [759.14, 206.36], [761.34, 207.01], [728.09, 179.4], [726.54, 179.28], [759.14, 179.29], [761.34, 179.42]]}, {\"category\": \"car\", \"corners_3d\": [[584.96, 210.6], [619.04, 210.62], [619.23, 214.92], [581.27, 214.89], [584.96, 179.58], [619.04, 179.58], [619.23, 180.35], [581.27, 180.35]]}]\n```" - } - ] - }, - { - "id": 3426, - "data_source": "Mono3DRefer", - "scene_name": "005920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[597.49, 220.96], [644.5, 220.96], [649.65, 228.17], [595.6, 228.16], [597.49, 183.13], [644.5, 183.14], [649.65, 184.68], [595.6, 184.67]]}, {\"category\": \"car\", \"corners_3d\": [[447.17, 203.59], [423.18, 203.55], [438.21, 201.4], [460.54, 201.43], [447.17, 181.17], [423.18, 181.16], [438.21, 180.58], [460.54, 180.59]]}, {\"category\": \"van\", \"corners_3d\": [[632.7, 202.23], [657.35, 202.26], [659.61, 204.85], [632.78, 204.82], [632.7, 174.17], [657.35, 174.17], [659.61, 174.29], [632.78, 174.29]]}, {\"category\": \"car\", \"corners_3d\": [[584.96, 210.6], [619.04, 210.62], [619.23, 214.92], [581.27, 214.89], [584.96, 179.58], [619.04, 179.58], [619.23, 180.35], [581.27, 180.35]]}]\n```" - } - ] - }, - { - "id": 3427, - "data_source": "Mono3DRefer", - "scene_name": "005920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[597.49, 220.96], [644.5, 220.96], [649.65, 228.17], [595.6, 228.16], [597.49, 183.13], [644.5, 183.14], [649.65, 184.68], [595.6, 184.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[801.56, 215.05], [788.97, 215.06], [783.78, 213.94], [796.04, 213.92], [801.56, 175.73], [788.97, 175.73], [783.78, 175.66], [796.04, 175.66]]}, {\"category\": \"car\", \"corners_3d\": [[447.17, 203.59], [423.18, 203.55], [438.21, 201.4], [460.54, 201.43], [447.17, 181.17], [423.18, 181.16], [438.21, 180.58], [460.54, 180.59]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.12, 202.38], [553.75, 202.79], [543.37, 202.85], [542.88, 202.44], [553.12, 173.01], [553.75, 173.01], [543.37, 173.01], [542.88, 173.01]]}, {\"category\": \"car\", \"corners_3d\": [[584.96, 210.6], [619.04, 210.62], [619.23, 214.92], [581.27, 214.89], [584.96, 179.58], [619.04, 179.58], [619.23, 180.35], [581.27, 180.35]]}]\n```" - } - ] - }, - { - "id": 3428, - "data_source": "Mono3DRefer", - "scene_name": "005920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[728.09, 206.93], [726.54, 206.28], [759.14, 206.36], [761.34, 207.01], [728.09, 179.4], [726.54, 179.28], [759.14, 179.29], [761.34, 179.42]]}, {\"category\": \"van\", \"corners_3d\": [[632.7, 202.23], [657.35, 202.26], [659.61, 204.85], [632.78, 204.82], [632.7, 174.17], [657.35, 174.17], [659.61, 174.29], [632.78, 174.29]]}]\n```" - } - ] - }, - { - "id": 3429, - "data_source": "Mono3DRefer", - "scene_name": "005920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[801.56, 215.05], [788.97, 215.06], [783.78, 213.94], [796.04, 213.92], [801.56, 175.73], [788.97, 175.73], [783.78, 175.66], [796.04, 175.66]]}, {\"category\": \"cyclist\", \"corners_3d\": [[728.09, 206.93], [726.54, 206.28], [759.14, 206.36], [761.34, 207.01], [728.09, 179.4], [726.54, 179.28], [759.14, 179.29], [761.34, 179.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.12, 202.38], [553.75, 202.79], [543.37, 202.85], [542.88, 202.44], [553.12, 173.01], [553.75, 173.01], [543.37, 173.01], [542.88, 173.01]]}]\n```" - } - ] - }, - { - "id": 3430, - "data_source": "Mono3DRefer", - "scene_name": "005920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[801.56, 215.05], [788.97, 215.06], [783.78, 213.94], [796.04, 213.92], [801.56, 175.73], [788.97, 175.73], [783.78, 175.66], [796.04, 175.66]]}, {\"category\": \"van\", \"corners_3d\": [[632.7, 202.23], [657.35, 202.26], [659.61, 204.85], [632.78, 204.82], [632.7, 174.17], [657.35, 174.17], [659.61, 174.29], [632.78, 174.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.12, 202.38], [553.75, 202.79], [543.37, 202.85], [542.88, 202.44], [553.12, 173.01], [553.75, 173.01], [543.37, 173.01], [542.88, 173.01]]}]\n```" - } - ] - }, - { - "id": 3431, - "data_source": "Mono3DRefer", - "scene_name": "005920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[597.49, 220.96], [644.5, 220.96], [649.65, 228.17], [595.6, 228.16], [597.49, 183.13], [644.5, 183.14], [649.65, 184.68], [595.6, 184.67]]}, {\"category\": \"car\", \"corners_3d\": [[447.17, 203.59], [423.18, 203.55], [438.21, 201.4], [460.54, 201.43], [447.17, 181.17], [423.18, 181.16], [438.21, 180.58], [460.54, 180.59]]}, {\"category\": \"cyclist\", \"corners_3d\": [[728.09, 206.93], [726.54, 206.28], [759.14, 206.36], [761.34, 207.01], [728.09, 179.4], [726.54, 179.28], [759.14, 179.29], [761.34, 179.42]]}, {\"category\": \"van\", \"corners_3d\": [[632.7, 202.23], [657.35, 202.26], [659.61, 204.85], [632.78, 204.82], [632.7, 174.17], [657.35, 174.17], [659.61, 174.29], [632.78, 174.29]]}, {\"category\": \"car\", \"corners_3d\": [[584.96, 210.6], [619.04, 210.62], [619.23, 214.92], [581.27, 214.89], [584.96, 179.58], [619.04, 179.58], [619.23, 180.35], [581.27, 180.35]]}]\n```" - } - ] - }, - { - "id": 3432, - "data_source": "Mono3DRefer", - "scene_name": "005920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[597.49, 220.96], [644.5, 220.96], [649.65, 228.17], [595.6, 228.16], [597.49, 183.13], [644.5, 183.14], [649.65, 184.68], [595.6, 184.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[801.56, 215.05], [788.97, 215.06], [783.78, 213.94], [796.04, 213.92], [801.56, 175.73], [788.97, 175.73], [783.78, 175.66], [796.04, 175.66]]}, {\"category\": \"car\", \"corners_3d\": [[447.17, 203.59], [423.18, 203.55], [438.21, 201.4], [460.54, 201.43], [447.17, 181.17], [423.18, 181.16], [438.21, 180.58], [460.54, 180.59]]}, {\"category\": \"cyclist\", \"corners_3d\": [[728.09, 206.93], [726.54, 206.28], [759.14, 206.36], [761.34, 207.01], [728.09, 179.4], [726.54, 179.28], [759.14, 179.29], [761.34, 179.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.12, 202.38], [553.75, 202.79], [543.37, 202.85], [542.88, 202.44], [553.12, 173.01], [553.75, 173.01], [543.37, 173.01], [542.88, 173.01]]}, {\"category\": \"car\", \"corners_3d\": [[584.96, 210.6], [619.04, 210.62], [619.23, 214.92], [581.27, 214.89], [584.96, 179.58], [619.04, 179.58], [619.23, 180.35], [581.27, 180.35]]}]\n```" - } - ] - }, - { - "id": 3433, - "data_source": "Mono3DRefer", - "scene_name": "005920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[597.49, 220.96], [644.5, 220.96], [649.65, 228.17], [595.6, 228.16], [597.49, 183.13], [644.5, 183.14], [649.65, 184.68], [595.6, 184.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[801.56, 215.05], [788.97, 215.06], [783.78, 213.94], [796.04, 213.92], [801.56, 175.73], [788.97, 175.73], [783.78, 175.66], [796.04, 175.66]]}, {\"category\": \"car\", \"corners_3d\": [[447.17, 203.59], [423.18, 203.55], [438.21, 201.4], [460.54, 201.43], [447.17, 181.17], [423.18, 181.16], [438.21, 180.58], [460.54, 180.59]]}, {\"category\": \"van\", \"corners_3d\": [[632.7, 202.23], [657.35, 202.26], [659.61, 204.85], [632.78, 204.82], [632.7, 174.17], [657.35, 174.17], [659.61, 174.29], [632.78, 174.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.12, 202.38], [553.75, 202.79], [543.37, 202.85], [542.88, 202.44], [553.12, 173.01], [553.75, 173.01], [543.37, 173.01], [542.88, 173.01]]}, {\"category\": \"car\", \"corners_3d\": [[584.96, 210.6], [619.04, 210.62], [619.23, 214.92], [581.27, 214.89], [584.96, 179.58], [619.04, 179.58], [619.23, 180.35], [581.27, 180.35]]}]\n```" - } - ] - }, - { - "id": 3434, - "data_source": "Mono3DRefer", - "scene_name": "005920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[801.56, 215.05], [788.97, 215.06], [783.78, 213.94], [796.04, 213.92], [801.56, 175.73], [788.97, 175.73], [783.78, 175.66], [796.04, 175.66]]}, {\"category\": \"cyclist\", \"corners_3d\": [[728.09, 206.93], [726.54, 206.28], [759.14, 206.36], [761.34, 207.01], [728.09, 179.4], [726.54, 179.28], [759.14, 179.29], [761.34, 179.42]]}, {\"category\": \"van\", \"corners_3d\": [[632.7, 202.23], [657.35, 202.26], [659.61, 204.85], [632.78, 204.82], [632.7, 174.17], [657.35, 174.17], [659.61, 174.29], [632.78, 174.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.12, 202.38], [553.75, 202.79], [543.37, 202.85], [542.88, 202.44], [553.12, 173.01], [553.75, 173.01], [543.37, 173.01], [542.88, 173.01]]}]\n```" - } - ] - }, - { - "id": 3435, - "data_source": "Mono3DRefer", - "scene_name": "005920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[597.49, 220.96], [644.5, 220.96], [649.65, 228.17], [595.6, 228.16], [597.49, 183.13], [644.5, 183.14], [649.65, 184.68], [595.6, 184.67]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[801.56, 215.05], [788.97, 215.06], [783.78, 213.94], [796.04, 213.92], [801.56, 175.73], [788.97, 175.73], [783.78, 175.66], [796.04, 175.66]]}, {\"category\": \"car\", \"corners_3d\": [[447.17, 203.59], [423.18, 203.55], [438.21, 201.4], [460.54, 201.43], [447.17, 181.17], [423.18, 181.16], [438.21, 180.58], [460.54, 180.59]]}, {\"category\": \"cyclist\", \"corners_3d\": [[728.09, 206.93], [726.54, 206.28], [759.14, 206.36], [761.34, 207.01], [728.09, 179.4], [726.54, 179.28], [759.14, 179.29], [761.34, 179.42]]}, {\"category\": \"van\", \"corners_3d\": [[632.7, 202.23], [657.35, 202.26], [659.61, 204.85], [632.78, 204.82], [632.7, 174.17], [657.35, 174.17], [659.61, 174.29], [632.78, 174.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.12, 202.38], [553.75, 202.79], [543.37, 202.85], [542.88, 202.44], [553.12, 173.01], [553.75, 173.01], [543.37, 173.01], [542.88, 173.01]]}, {\"category\": \"car\", \"corners_3d\": [[584.96, 210.6], [619.04, 210.62], [619.23, 214.92], [581.27, 214.89], [584.96, 179.58], [619.04, 179.58], [619.23, 180.35], [581.27, 180.35]]}]\n```" - } - ] - }, - { - "id": 3436, - "data_source": "Mono3DRefer", - "scene_name": "005930", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[504.76, 239.48], [452.83, 240.1], [460.27, 232.01], [505.94, 231.53], [504.76, 191.47], [452.83, 191.64], [460.27, 189.38], [505.94, 189.24]]}]\n```" - } - ] - }, - { - "id": 3437, - "data_source": "Mono3DRefer", - "scene_name": "005937", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[164.89, 302.03], [190.66, 295.45], [286.11, 296.27], [265.65, 302.95], [164.89, 136.49], [190.66, 138.88], [286.11, 138.58], [265.65, 136.16]]}]\n```" - } - ] - }, - { - "id": 3438, - "data_source": "Mono3DRefer", - "scene_name": "005941", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[556.54, 212.85], [523.1, 212.86], [531.93, 208.74], [561.93, 208.74], [556.54, 183.48], [523.1, 183.48], [531.93, 182.38], [561.93, 182.38]]}]\n```" - } - ] - }, - { - "id": 3439, - "data_source": "Mono3DRefer", - "scene_name": "005954", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[285.45, 324.83], [261.84, 336.28], [153.81, 336.99], [185.49, 325.43], [285.45, 145.66], [261.84, 142.89], [153.81, 142.72], [185.49, 145.51]]}]\n```" - } - ] - }, - { - "id": 3440, - "data_source": "Mono3DRefer", - "scene_name": "005955", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[317.26, 306.17], [443.87, 306.66], [326.28, 393.72], [118.5, 392.39], [317.26, 199.49], [443.87, 199.59], [326.28, 216.99], [118.5, 216.72]]}, {\"category\": \"car\", \"corners_3d\": [[691.0, 231.97], [639.91, 232.1], [632.46, 223.7], [676.34, 223.61], [691.0, 184.6], [639.91, 184.63], [632.46, 182.96], [676.34, 182.94]]}]\n```" - } - ] - }, - { - "id": 3441, - "data_source": "Mono3DRefer", - "scene_name": "005955", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[555.62, 207.9], [523.5, 207.81], [535.07, 204.67], [564.33, 204.75], [555.62, 175.31], [523.5, 175.31], [535.07, 175.09], [564.33, 175.09]]}, {\"category\": \"van\", \"corners_3d\": [[626.05, 202.5], [652.36, 202.54], [654.24, 205.26], [625.52, 205.22], [626.05, 173.98], [652.36, 173.98], [654.24, 174.08], [625.52, 174.08]]}]\n```" - } - ] - }, - { - "id": 3442, - "data_source": "Mono3DRefer", - "scene_name": "005955", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[317.26, 306.17], [443.87, 306.66], [326.28, 393.72], [118.5, 392.39], [317.26, 199.49], [443.87, 199.59], [326.28, 216.99], [118.5, 216.72]]}, {\"category\": \"car\", \"corners_3d\": [[691.0, 231.97], [639.91, 232.1], [632.46, 223.7], [676.34, 223.61], [691.0, 184.6], [639.91, 184.63], [632.46, 182.96], [676.34, 182.94]]}, {\"category\": \"van\", \"corners_3d\": [[555.62, 207.9], [523.5, 207.81], [535.07, 204.67], [564.33, 204.75], [555.62, 175.31], [523.5, 175.31], [535.07, 175.09], [564.33, 175.09]]}, {\"category\": \"van\", \"corners_3d\": [[626.05, 202.5], [652.36, 202.54], [654.24, 205.26], [625.52, 205.22], [626.05, 173.98], [652.36, 173.98], [654.24, 174.08], [625.52, 174.08]]}]\n```" - } - ] - }, - { - "id": 3443, - "data_source": "Mono3DRefer", - "scene_name": "005956", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[897.83, 333.63], [737.32, 328.24], [733.34, 276.34], [838.93, 278.71], [897.83, 190.18], [737.32, 189.6], [733.34, 184.01], [838.93, 184.26]]}]\n```" - } - ] - }, - { - "id": 3444, - "data_source": "Mono3DRefer", - "scene_name": "005958", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[757.78, 305.28], [745.41, 295.74], [836.22, 295.57], [856.09, 305.08], [757.78, 152.4], [745.41, 154.55], [836.22, 154.59], [856.09, 152.44]]}]\n```" - } - ] - }, - { - "id": 3445, - "data_source": "Mono3DRefer", - "scene_name": "005967", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[290.44, 247.31], [318.99, 240.96], [452.75, 241.24], [436.6, 247.64], [290.44, 173.24], [318.99, 173.21], [452.75, 173.21], [436.6, 173.24]]}, {\"category\": \"car\", \"corners_3d\": [[343.14, 232.33], [362.19, 228.21], [482.44, 228.31], [472.32, 232.45], [343.14, 178.89], [362.19, 178.47], [482.44, 178.48], [472.32, 178.9]]}, {\"category\": \"car\", \"corners_3d\": [[797.89, 198.01], [834.71, 198.08], [860.89, 201.58], [818.9, 201.5], [797.89, 160.81], [834.71, 160.78], [860.89, 159.11], [818.9, 159.15]]}, {\"category\": \"car\", \"corners_3d\": [[389.11, 218.92], [400.8, 216.46], [502.35, 216.45], [496.38, 218.91], [389.11, 180.39], [400.8, 179.99], [502.35, 179.99], [496.38, 180.39]]}, {\"category\": \"car\", \"corners_3d\": [[416.5, 213.06], [426.7, 211.03], [521.9, 211.08], [516.77, 213.13], [416.5, 174.82], [426.7, 174.72], [521.9, 174.73], [516.77, 174.83]]}]\n```" - } - ] - }, - { - "id": 3446, - "data_source": "Mono3DRefer", - "scene_name": "005990", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-991.53, 1411.46], [-1946.66, 1412.76], [-36.87, 486.2], [204.19, 486.12], [-991.53, 279.18], [-1946.66, 279.29], [-36.87, 199.75], [204.19, 199.75]]}, {\"category\": \"car\", \"corners_3d\": [[551.95, 206.16], [526.89, 206.32], [525.4, 203.78], [548.58, 203.64], [551.95, 180.49], [526.89, 180.53], [525.4, 179.94], [548.58, 179.91]]}]\n```" - } - ] - }, - { - "id": 3447, - "data_source": "Mono3DRefer", - "scene_name": "005993", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[572.05, 217.76], [543.56, 217.96], [542.33, 214.09], [568.39, 213.92], [572.05, 188.52], [543.56, 188.58], [542.33, 187.23], [568.39, 187.18]]}]\n```" - } - ] - }, - { - "id": 3448, - "data_source": "Mono3DRefer", - "scene_name": "006003", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[798.68, 309.25], [950.8, 309.27], [1235.57, 423.31], [956.23, 423.23], [798.68, 178.04], [950.8, 178.04], [1235.57, 182.37], [956.23, 182.37]]}, {\"category\": \"car\", \"corners_3d\": [[31.79, 378.32], [-130.84, 377.89], [150.61, 300.64], [252.26, 300.81], [31.79, 230.44], [-130.84, 230.32], [150.61, 208.67], [252.26, 208.72]]}, {\"category\": \"car\", \"corners_3d\": [[358.43, 265.11], [277.84, 265.12], [349.9, 245.04], [412.95, 245.04], [358.43, 180.16], [277.84, 180.16], [349.9, 178.57], [412.95, 178.57]]}, {\"category\": \"car\", \"corners_3d\": [[563.11, 204.54], [532.7, 204.51], [541.35, 201.49], [568.86, 201.51], [563.11, 180.12], [532.7, 180.12], [541.35, 179.42], [568.86, 179.43]]}, {\"category\": \"car\", \"corners_3d\": [[427.4, 235.83], [372.03, 235.79], [410.6, 225.85], [457.25, 225.88], [427.4, 181.31], [372.03, 181.3], [410.6, 179.97], [457.25, 179.97]]}, {\"category\": \"car\", \"corners_3d\": [[486.74, 216.18], [448.76, 216.18], [465.39, 211.68], [499.42, 211.68], [486.74, 184.24], [448.76, 184.24], [465.39, 183.06], [499.42, 183.06]]}, {\"category\": \"car\", \"corners_3d\": [[635.57, 193.99], [656.21, 193.99], [659.76, 195.62], [637.52, 195.62], [635.57, 174.71], [656.21, 174.71], [659.76, 174.86], [637.52, 174.86]]}, {\"category\": \"car\", \"corners_3d\": [[701.04, 190.5], [697.61, 189.89], [765.17, 189.87], [770.99, 190.48], [701.04, 168.32], [697.61, 168.48], [765.17, 168.48], [770.99, 168.33]]}]\n```" - } - ] - }, - { - "id": 3449, - "data_source": "Mono3DRefer", - "scene_name": "006003", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[913.34, 178.94], [925.12, 179.16], [863.45, 179.18], [853.8, 178.96], [913.34, 153.85], [925.12, 153.16], [863.45, 153.11], [853.8, 153.8]]}]\n```" - } - ] - }, - { - "id": 3450, - "data_source": "Mono3DRefer", - "scene_name": "006003", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[798.68, 309.25], [950.8, 309.27], [1235.57, 423.31], [956.23, 423.23], [798.68, 178.04], [950.8, 178.04], [1235.57, 182.37], [956.23, 182.37]]}, {\"category\": \"car\", \"corners_3d\": [[31.79, 378.32], [-130.84, 377.89], [150.61, 300.64], [252.26, 300.81], [31.79, 230.44], [-130.84, 230.32], [150.61, 208.67], [252.26, 208.72]]}, {\"category\": \"car\", \"corners_3d\": [[358.43, 265.11], [277.84, 265.12], [349.9, 245.04], [412.95, 245.04], [358.43, 180.16], [277.84, 180.16], [349.9, 178.57], [412.95, 178.57]]}, {\"category\": \"car\", \"corners_3d\": [[563.11, 204.54], [532.7, 204.51], [541.35, 201.49], [568.86, 201.51], [563.11, 180.12], [532.7, 180.12], [541.35, 179.42], [568.86, 179.43]]}, {\"category\": \"car\", \"corners_3d\": [[427.4, 235.83], [372.03, 235.79], [410.6, 225.85], [457.25, 225.88], [427.4, 181.31], [372.03, 181.3], [410.6, 179.97], [457.25, 179.97]]}, {\"category\": \"car\", \"corners_3d\": [[486.74, 216.18], [448.76, 216.18], [465.39, 211.68], [499.42, 211.68], [486.74, 184.24], [448.76, 184.24], [465.39, 183.06], [499.42, 183.06]]}, {\"category\": \"car\", \"corners_3d\": [[635.57, 193.99], [656.21, 193.99], [659.76, 195.62], [637.52, 195.62], [635.57, 174.71], [656.21, 174.71], [659.76, 174.86], [637.52, 174.86]]}, {\"category\": \"car\", \"corners_3d\": [[701.04, 190.5], [697.61, 189.89], [765.17, 189.87], [770.99, 190.48], [701.04, 168.32], [697.61, 168.48], [765.17, 168.48], [770.99, 168.33]]}, {\"category\": \"van\", \"corners_3d\": [[913.34, 178.94], [925.12, 179.16], [863.45, 179.18], [853.8, 178.96], [913.34, 153.85], [925.12, 153.16], [863.45, 153.11], [853.8, 153.8]]}]\n```" - } - ] - }, - { - "id": 3451, - "data_source": "Mono3DRefer", - "scene_name": "006021", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[807.87, 231.31], [844.01, 235.65], [718.63, 240.66], [689.13, 235.63], [807.87, 172.11], [844.01, 172.05], [718.63, 171.99], [689.13, 172.06]]}, {\"category\": \"car\", \"corners_3d\": [[1277.16, 364.14], [1560.96, 418.51], [1242.91, 544.14], [953.15, 432.57], [1277.16, 202.89], [1560.96, 211.43], [1242.91, 231.16], [953.15, 213.64]]}, {\"category\": \"car\", \"corners_3d\": [[513.27, 326.5], [387.74, 325.99], [460.2, 278.9], [547.21, 279.15], [513.27, 204.62], [387.74, 204.51], [460.2, 194.78], [547.21, 194.83]]}]\n```" - } - ] - }, - { - "id": 3452, - "data_source": "Mono3DRefer", - "scene_name": "006021", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[581.94, 266.59], [559.55, 268.42], [517.67, 258.73], [538.55, 257.26], [581.94, 169.58], [559.55, 169.52], [517.67, 169.86], [538.55, 169.91]]}]\n```" - } - ] - }, - { - "id": 3453, - "data_source": "Mono3DRefer", - "scene_name": "006021", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[581.94, 266.59], [559.55, 268.42], [517.67, 258.73], [538.55, 257.26], [581.94, 169.58], [559.55, 169.52], [517.67, 169.86], [538.55, 169.91]]}, {\"category\": \"car\", \"corners_3d\": [[807.87, 231.31], [844.01, 235.65], [718.63, 240.66], [689.13, 235.63], [807.87, 172.11], [844.01, 172.05], [718.63, 171.99], [689.13, 172.06]]}, {\"category\": \"car\", \"corners_3d\": [[1277.16, 364.14], [1560.96, 418.51], [1242.91, 544.14], [953.15, 432.57], [1277.16, 202.89], [1560.96, 211.43], [1242.91, 231.16], [953.15, 213.64]]}, {\"category\": \"car\", \"corners_3d\": [[513.27, 326.5], [387.74, 325.99], [460.2, 278.9], [547.21, 279.15], [513.27, 204.62], [387.74, 204.51], [460.2, 194.78], [547.21, 194.83]]}]\n```" - } - ] - }, - { - "id": 3454, - "data_source": "Mono3DRefer", - "scene_name": "006022", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[489.53, 264.94], [498.1, 253.38], [725.21, 251.58], [748.62, 262.59], [489.53, 178.7], [498.1, 177.97], [725.21, 177.85], [748.62, 178.55]]}, {\"category\": \"car\", \"corners_3d\": [[543.67, 205.78], [566.76, 205.66], [570.97, 208.82], [545.65, 208.97], [543.67, 184.11], [566.76, 184.07], [570.97, 185.15], [545.65, 185.2]]}, {\"category\": \"car\", \"corners_3d\": [[476.4, 212.8], [448.67, 212.94], [457.51, 209.21], [482.65, 209.1], [476.4, 189.45], [448.67, 189.51], [457.51, 187.96], [482.65, 187.92]]}]\n```" - } - ] - }, - { - "id": 3455, - "data_source": "Mono3DRefer", - "scene_name": "006029", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[444.84, 313.88], [459.24, 305.58], [542.37, 307.11], [533.38, 315.62], [444.84, 155.42], [459.24, 156.99], [542.37, 156.7], [533.38, 155.1]]}]\n```" - } - ] - }, - { - "id": 3456, - "data_source": "Mono3DRefer", - "scene_name": "006030", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[931.03, 235.13], [935.73, 231.19], [1090.29, 236.68], [1096.63, 241.43], [931.03, 184.11], [935.73, 183.39], [1090.29, 184.39], [1096.63, 185.25]]}, {\"category\": \"car\", \"corners_3d\": [[761.7, 226.29], [783.28, 228.07], [725.66, 231.69], [704.71, 229.66], [761.7, 201.47], [783.28, 202.43], [725.66, 204.36], [704.71, 203.28]]}]\n```" - } - ] - }, - { - "id": 3457, - "data_source": "Mono3DRefer", - "scene_name": "006030", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[630.26, 223.72], [664.22, 223.67], [673.04, 230.22], [634.72, 230.27], [630.26, 185.37], [664.22, 185.36], [673.04, 186.97], [634.72, 186.99]]}, {\"category\": \"van\", \"corners_3d\": [[555.17, 206.66], [574.64, 206.56], [578.02, 208.77], [557.28, 208.89], [555.17, 182.6], [574.64, 182.57], [578.02, 183.21], [557.28, 183.24]]}]\n```" - } - ] - }, - { - "id": 3458, - "data_source": "Mono3DRefer", - "scene_name": "006030", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[931.03, 235.13], [935.73, 231.19], [1090.29, 236.68], [1096.63, 241.43], [931.03, 184.11], [935.73, 183.39], [1090.29, 184.39], [1096.63, 185.25]]}, {\"category\": \"van\", \"corners_3d\": [[630.26, 223.72], [664.22, 223.67], [673.04, 230.22], [634.72, 230.27], [630.26, 185.37], [664.22, 185.36], [673.04, 186.97], [634.72, 186.99]]}, {\"category\": \"car\", \"corners_3d\": [[761.7, 226.29], [783.28, 228.07], [725.66, 231.69], [704.71, 229.66], [761.7, 201.47], [783.28, 202.43], [725.66, 204.36], [704.71, 203.28]]}, {\"category\": \"van\", \"corners_3d\": [[555.17, 206.66], [574.64, 206.56], [578.02, 208.77], [557.28, 208.89], [555.17, 182.6], [574.64, 182.57], [578.02, 183.21], [557.28, 183.24]]}]\n```" - } - ] - }, - { - "id": 3459, - "data_source": "Mono3DRefer", - "scene_name": "006035", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[241.11, 313.5], [208.09, 326.9], [99.37, 328.0], [142.64, 314.4], [241.11, 162.83], [208.09, 161.05], [99.37, 160.91], [142.64, 162.71]]}]\n```" - } - ] - }, - { - "id": 3460, - "data_source": "Mono3DRefer", - "scene_name": "006040", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[724.76, 226.12], [782.88, 226.76], [796.47, 238.26], [725.93, 237.33], [724.76, 172.5], [782.88, 172.49], [796.47, 172.42], [725.93, 172.42]]}, {\"category\": \"car\", \"corners_3d\": [[248.12, 350.06], [114.64, 343.12], [318.97, 288.16], [413.76, 291.31], [248.12, 175.79], [114.64, 175.67], [318.97, 174.76], [413.76, 174.82]]}, {\"category\": \"car\", \"corners_3d\": [[-90.18, 343.3], [-205.17, 337.33], [109.98, 282.51], [193.25, 285.13], [-90.18, 201.84], [-205.17, 200.82], [109.98, 191.5], [193.25, 191.95]]}, {\"category\": \"car\", \"corners_3d\": [[442.82, 237.33], [391.21, 235.86], [465.88, 224.99], [509.86, 226.0], [442.82, 187.01], [391.21, 186.69], [465.88, 184.3], [509.86, 184.53]]}, {\"category\": \"car\", \"corners_3d\": [[949.95, 190.56], [969.9, 191.31], [905.1, 191.64], [887.48, 190.86], [949.95, 162.44], [969.9, 162.0], [905.1, 161.8], [887.48, 162.26]]}]\n```" - } - ] - }, - { - "id": 3461, - "data_source": "Mono3DRefer", - "scene_name": "006050", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1166.82, 237.07], [1254.1, 246.52], [978.0, 247.99], [925.22, 238.19], [1166.82, 118.66], [1254.1, 110.68], [978.0, 109.44], [925.22, 117.71]]}]\n```" - } - ] - }, - { - "id": 3462, - "data_source": "Mono3DRefer", - "scene_name": "006050", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[188.05, 435.73], [-47.33, 430.46], [259.01, 317.67], [393.65, 319.32], [188.05, 216.87], [-47.33, 215.99], [259.01, 197.1], [393.65, 197.38]]}, {\"category\": \"car\", \"corners_3d\": [[413.12, 318.75], [276.35, 317.65], [393.52, 270.72], [486.33, 271.22], [413.12, 174.46], [276.35, 174.44], [393.52, 173.93], [486.33, 173.93]]}, {\"category\": \"car\", \"corners_3d\": [[713.26, 233.58], [770.07, 233.49], [808.72, 246.93], [739.36, 247.07], [713.26, 185.19], [770.07, 185.17], [808.72, 187.9], [739.36, 187.93]]}, {\"category\": \"car\", \"corners_3d\": [[526.52, 249.36], [461.51, 249.02], [494.3, 235.41], [547.78, 235.64], [526.52, 180.41], [461.51, 180.38], [494.3, 179.03], [547.78, 179.06]]}, {\"category\": \"car\", \"corners_3d\": [[688.92, 221.03], [737.71, 220.97], [764.1, 229.91], [706.27, 230.0], [688.92, 174.92], [737.71, 174.91], [764.1, 175.3], [706.27, 175.3]]}, {\"category\": \"car\", \"corners_3d\": [[551.26, 233.4], [504.75, 233.32], [519.38, 225.87], [560.17, 225.92], [551.26, 181.49], [504.75, 181.48], [519.38, 180.42], [560.17, 180.43]]}]\n```" - } - ] - }, - { - "id": 3463, - "data_source": "Mono3DRefer", - "scene_name": "006050", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[188.05, 435.73], [-47.33, 430.46], [259.01, 317.67], [393.65, 319.32], [188.05, 216.87], [-47.33, 215.99], [259.01, 197.1], [393.65, 197.38]]}, {\"category\": \"car\", \"corners_3d\": [[413.12, 318.75], [276.35, 317.65], [393.52, 270.72], [486.33, 271.22], [413.12, 174.46], [276.35, 174.44], [393.52, 173.93], [486.33, 173.93]]}, {\"category\": \"van\", \"corners_3d\": [[1166.82, 237.07], [1254.1, 246.52], [978.0, 247.99], [925.22, 238.19], [1166.82, 118.66], [1254.1, 110.68], [978.0, 109.44], [925.22, 117.71]]}, {\"category\": \"car\", \"corners_3d\": [[713.26, 233.58], [770.07, 233.49], [808.72, 246.93], [739.36, 247.07], [713.26, 185.19], [770.07, 185.17], [808.72, 187.9], [739.36, 187.93]]}, {\"category\": \"car\", \"corners_3d\": [[526.52, 249.36], [461.51, 249.02], [494.3, 235.41], [547.78, 235.64], [526.52, 180.41], [461.51, 180.38], [494.3, 179.03], [547.78, 179.06]]}, {\"category\": \"car\", \"corners_3d\": [[688.92, 221.03], [737.71, 220.97], [764.1, 229.91], [706.27, 230.0], [688.92, 174.92], [737.71, 174.91], [764.1, 175.3], [706.27, 175.3]]}, {\"category\": \"car\", \"corners_3d\": [[551.26, 233.4], [504.75, 233.32], [519.38, 225.87], [560.17, 225.92], [551.26, 181.49], [504.75, 181.48], [519.38, 180.42], [560.17, 180.43]]}]\n```" - } - ] - }, - { - "id": 3464, - "data_source": "Mono3DRefer", - "scene_name": "006076", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[818.93, 313.98], [798.83, 302.99], [887.25, 302.35], [915.12, 313.21], [818.93, 143.8], [798.83, 146.82], [887.25, 147.0], [915.12, 144.01]]}]\n```" - } - ] - }, - { - "id": 3465, - "data_source": "Mono3DRefer", - "scene_name": "006104", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[562.12, 207.54], [530.99, 207.52], [538.63, 204.41], [566.98, 204.43], [562.12, 182.0], [530.99, 182.0], [538.63, 181.18], [566.98, 181.18]]}]\n```" - } - ] - }, - { - "id": 3466, - "data_source": "Mono3DRefer", - "scene_name": "006105", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[548.29, 216.08], [512.38, 216.34], [513.39, 211.75], [545.53, 211.55], [548.29, 179.4], [512.38, 179.44], [513.39, 178.74], [545.53, 178.71]]}, {\"category\": \"car\", \"corners_3d\": [[519.12, 205.02], [492.91, 205.33], [485.95, 202.63], [510.07, 202.37], [519.12, 182.5], [492.91, 182.59], [485.95, 181.78], [510.07, 181.7]]}]\n```" - } - ] - }, - { - "id": 3467, - "data_source": "Mono3DRefer", - "scene_name": "006120", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[379.16, 300.78], [259.77, 301.43], [351.93, 265.25], [437.52, 264.92], [379.16, 184.46], [259.77, 184.52], [351.93, 181.24], [437.52, 181.21]]}]\n```" - } - ] - }, - { - "id": 3468, - "data_source": "Mono3DRefer", - "scene_name": "006124", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[710.02, 182.76], [726.37, 182.78], [728.64, 183.4], [711.27, 183.37], [710.02, 168.16], [726.37, 168.15], [728.64, 167.85], [711.27, 167.87]]}]\n```" - } - ] - }, - { - "id": 3469, - "data_source": "Mono3DRefer", - "scene_name": "006124", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[189.52, 230.99], [153.09, 239.18], [-349.35, 248.22], [-239.31, 237.82], [189.52, 114.7], [153.09, 106.51], [-349.35, 97.45], [-239.31, 107.86]]}]\n```" - } - ] - }, - { - "id": 3470, - "data_source": "Mono3DRefer", - "scene_name": "006124", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[189.52, 230.99], [153.09, 239.18], [-349.35, 248.22], [-239.31, 237.82], [189.52, 114.7], [153.09, 106.51], [-349.35, 97.45], [-239.31, 107.86]]}, {\"category\": \"car\", \"corners_3d\": [[710.02, 182.76], [726.37, 182.78], [728.64, 183.4], [711.27, 183.37], [710.02, 168.16], [726.37, 168.15], [728.64, 167.85], [711.27, 167.87]]}]\n```" - } - ] - }, - { - "id": 3471, - "data_source": "Mono3DRefer", - "scene_name": "006126", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[384.52, 216.15], [395.66, 214.15], [466.65, 214.23], [458.94, 216.25], [384.52, 180.21], [395.66, 179.87], [466.65, 179.89], [458.94, 180.23]]}, {\"category\": \"car\", \"corners_3d\": [[714.91, 197.19], [687.22, 197.25], [676.38, 195.08], [701.62, 195.03], [714.91, 168.14], [687.22, 168.13], [676.38, 168.55], [701.62, 168.56]]}, {\"category\": \"car\", \"corners_3d\": [[587.51, 201.14], [564.04, 201.18], [565.35, 198.87], [586.91, 198.84], [587.51, 178.35], [564.04, 178.36], [565.35, 177.91], [586.91, 177.9]]}, {\"category\": \"car\", \"corners_3d\": [[820.19, 350.6], [977.75, 349.46], [1565.76, 616.23], [1175.96, 623.5], [820.19, 183.27], [977.75, 183.21], [1565.76, 198.85], [1175.96, 199.27]]}, {\"category\": \"car\", \"corners_3d\": [[712.08, 271.53], [800.15, 271.06], [899.01, 315.41], [771.59, 316.41], [712.08, 183.04], [800.15, 182.99], [899.01, 187.57], [771.59, 187.68]]}]\n```" - } - ] - }, - { - "id": 3472, - "data_source": "Mono3DRefer", - "scene_name": "006128", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[497.45, 266.21], [412.67, 266.22], [456.16, 245.54], [522.16, 245.53], [497.45, 176.92], [412.67, 176.92], [456.16, 176.02], [522.16, 176.02]]}, {\"category\": \"car\", \"corners_3d\": [[553.74, 216.29], [526.46, 216.26], [534.83, 212.51], [559.77, 212.54], [553.74, 185.9], [526.46, 185.89], [534.83, 184.76], [559.77, 184.77]]}]\n```" - } - ] - }, - { - "id": 3473, - "data_source": "Mono3DRefer", - "scene_name": "006130", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[528.82, 259.37], [493.28, 259.68], [498.92, 250.68], [530.36, 250.43], [528.82, 169.16], [493.28, 169.11], [498.92, 170.53], [530.36, 170.57]]}]\n```" - } - ] - }, - { - "id": 3474, - "data_source": "Mono3DRefer", - "scene_name": "006130", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[415.6, 262.2], [425.37, 259.14], [462.2, 259.56], [453.83, 262.65], [415.6, 191.1], [425.37, 190.73], [462.2, 190.78], [453.83, 191.15]]}]\n```" - } - ] - }, - { - "id": 3475, - "data_source": "Mono3DRefer", - "scene_name": "006130", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[873.27, 289.39], [945.24, 291.33], [975.98, 307.71], [892.86, 305.15], [873.27, 146.21], [945.24, 145.57], [975.98, 140.21], [892.86, 141.05]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[587.61, 264.86], [633.04, 268.68], [600.49, 274.57], [553.65, 270.23], [587.61, 163.01], [633.04, 162.17], [600.49, 160.86], [553.65, 161.82]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[210.35, 294.87], [221.92, 287.68], [285.57, 285.32], [278.49, 292.19], [210.35, 164.0], [221.92, 165.11], [285.57, 165.48], [278.49, 164.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[475.13, 278.32], [424.47, 277.98], [439.48, 271.13], [486.58, 271.42], [475.13, 161.52], [424.47, 161.59], [439.48, 163.01], [486.58, 162.95]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[828.23, 278.58], [772.53, 278.81], [754.32, 269.69], [804.86, 269.5], [828.23, 143.26], [772.53, 143.16], [754.32, 146.76], [804.86, 146.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.31, 220.45], [576.1, 220.45], [575.0, 222.19], [551.19, 222.19], [553.31, 167.67], [576.1, 167.67], [575.0, 167.06], [551.19, 167.06]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[527.95, 217.29], [548.41, 217.18], [549.31, 218.37], [528.16, 218.49], [527.95, 166.64], [548.41, 166.69], [549.31, 166.19], [528.16, 166.14]]}]\n```" - } - ] - }, - { - "id": 3476, - "data_source": "Mono3DRefer", - "scene_name": "006130", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[528.82, 259.37], [493.28, 259.68], [498.92, 250.68], [530.36, 250.43], [528.82, 169.16], [493.28, 169.11], [498.92, 170.53], [530.36, 170.57]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[415.6, 262.2], [425.37, 259.14], [462.2, 259.56], [453.83, 262.65], [415.6, 191.1], [425.37, 190.73], [462.2, 190.78], [453.83, 191.15]]}]\n```" - } - ] - }, - { - "id": 3477, - "data_source": "Mono3DRefer", - "scene_name": "006130", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[873.27, 289.39], [945.24, 291.33], [975.98, 307.71], [892.86, 305.15], [873.27, 146.21], [945.24, 145.57], [975.98, 140.21], [892.86, 141.05]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[587.61, 264.86], [633.04, 268.68], [600.49, 274.57], [553.65, 270.23], [587.61, 163.01], [633.04, 162.17], [600.49, 160.86], [553.65, 161.82]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[210.35, 294.87], [221.92, 287.68], [285.57, 285.32], [278.49, 292.19], [210.35, 164.0], [221.92, 165.11], [285.57, 165.48], [278.49, 164.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[475.13, 278.32], [424.47, 277.98], [439.48, 271.13], [486.58, 271.42], [475.13, 161.52], [424.47, 161.59], [439.48, 163.01], [486.58, 162.95]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[828.23, 278.58], [772.53, 278.81], [754.32, 269.69], [804.86, 269.5], [828.23, 143.26], [772.53, 143.16], [754.32, 146.76], [804.86, 146.83]]}, {\"category\": \"cyclist\", \"corners_3d\": [[528.82, 259.37], [493.28, 259.68], [498.92, 250.68], [530.36, 250.43], [528.82, 169.16], [493.28, 169.11], [498.92, 170.53], [530.36, 170.57]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.31, 220.45], [576.1, 220.45], [575.0, 222.19], [551.19, 222.19], [553.31, 167.67], [576.1, 167.67], [575.0, 167.06], [551.19, 167.06]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[527.95, 217.29], [548.41, 217.18], [549.31, 218.37], [528.16, 218.49], [527.95, 166.64], [548.41, 166.69], [549.31, 166.19], [528.16, 166.14]]}]\n```" - } - ] - }, - { - "id": 3478, - "data_source": "Mono3DRefer", - "scene_name": "006130", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[873.27, 289.39], [945.24, 291.33], [975.98, 307.71], [892.86, 305.15], [873.27, 146.21], [945.24, 145.57], [975.98, 140.21], [892.86, 141.05]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[587.61, 264.86], [633.04, 268.68], [600.49, 274.57], [553.65, 270.23], [587.61, 163.01], [633.04, 162.17], [600.49, 160.86], [553.65, 161.82]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[210.35, 294.87], [221.92, 287.68], [285.57, 285.32], [278.49, 292.19], [210.35, 164.0], [221.92, 165.11], [285.57, 165.48], [278.49, 164.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[475.13, 278.32], [424.47, 277.98], [439.48, 271.13], [486.58, 271.42], [475.13, 161.52], [424.47, 161.59], [439.48, 163.01], [486.58, 162.95]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[828.23, 278.58], [772.53, 278.81], [754.32, 269.69], [804.86, 269.5], [828.23, 143.26], [772.53, 143.16], [754.32, 146.76], [804.86, 146.83]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[415.6, 262.2], [425.37, 259.14], [462.2, 259.56], [453.83, 262.65], [415.6, 191.1], [425.37, 190.73], [462.2, 190.78], [453.83, 191.15]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.31, 220.45], [576.1, 220.45], [575.0, 222.19], [551.19, 222.19], [553.31, 167.67], [576.1, 167.67], [575.0, 167.06], [551.19, 167.06]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[527.95, 217.29], [548.41, 217.18], [549.31, 218.37], [528.16, 218.49], [527.95, 166.64], [548.41, 166.69], [549.31, 166.19], [528.16, 166.14]]}]\n```" - } - ] - }, - { - "id": 3479, - "data_source": "Mono3DRefer", - "scene_name": "006130", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, person_sitting, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[873.27, 289.39], [945.24, 291.33], [975.98, 307.71], [892.86, 305.15], [873.27, 146.21], [945.24, 145.57], [975.98, 140.21], [892.86, 141.05]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[587.61, 264.86], [633.04, 268.68], [600.49, 274.57], [553.65, 270.23], [587.61, 163.01], [633.04, 162.17], [600.49, 160.86], [553.65, 161.82]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[210.35, 294.87], [221.92, 287.68], [285.57, 285.32], [278.49, 292.19], [210.35, 164.0], [221.92, 165.11], [285.57, 165.48], [278.49, 164.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[475.13, 278.32], [424.47, 277.98], [439.48, 271.13], [486.58, 271.42], [475.13, 161.52], [424.47, 161.59], [439.48, 163.01], [486.58, 162.95]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[828.23, 278.58], [772.53, 278.81], [754.32, 269.69], [804.86, 269.5], [828.23, 143.26], [772.53, 143.16], [754.32, 146.76], [804.86, 146.83]]}, {\"category\": \"cyclist\", \"corners_3d\": [[528.82, 259.37], [493.28, 259.68], [498.92, 250.68], [530.36, 250.43], [528.82, 169.16], [493.28, 169.11], [498.92, 170.53], [530.36, 170.57]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[415.6, 262.2], [425.37, 259.14], [462.2, 259.56], [453.83, 262.65], [415.6, 191.1], [425.37, 190.73], [462.2, 190.78], [453.83, 191.15]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.31, 220.45], [576.1, 220.45], [575.0, 222.19], [551.19, 222.19], [553.31, 167.67], [576.1, 167.67], [575.0, 167.06], [551.19, 167.06]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[527.95, 217.29], [548.41, 217.18], [549.31, 218.37], [528.16, 218.49], [527.95, 166.64], [548.41, 166.69], [549.31, 166.19], [528.16, 166.14]]}]\n```" - } - ] - }, - { - "id": 3480, - "data_source": "Mono3DRefer", - "scene_name": "006131", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[389.57, 210.92], [399.54, 209.13], [476.41, 209.08], [470.25, 210.87], [389.57, 178.5], [399.54, 178.24], [476.41, 178.23], [470.25, 178.5]]}]\n```" - } - ] - }, - { - "id": 3481, - "data_source": "Mono3DRefer", - "scene_name": "006140", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[544.2, 233.77], [494.72, 233.86], [508.92, 225.18], [551.35, 225.12], [544.2, 181.81], [494.72, 181.82], [508.92, 180.55], [551.35, 180.54]]}, {\"category\": \"car\", \"corners_3d\": [[568.87, 210.72], [547.61, 210.7], [552.81, 208.11], [572.62, 208.12], [568.87, 187.05], [547.61, 187.04], [552.81, 186.07], [572.62, 186.08]]}]\n```" - } - ] - }, - { - "id": 3482, - "data_source": "Mono3DRefer", - "scene_name": "006146", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[904.04, 420.13], [1121.93, 417.92], [3820.37, 1657.59], [2591.6, 1742.49], [904.04, 187.35], [1121.93, 187.22], [3820.37, 259.9], [2591.6, 264.87]]}, {\"category\": \"car\", \"corners_3d\": [[733.19, 287.31], [836.32, 286.67], [982.8, 352.03], [821.23, 353.63], [733.19, 183.48], [836.32, 183.42], [982.8, 189.49], [821.23, 189.64]]}, {\"category\": \"car\", \"corners_3d\": [[372.23, 217.1], [384.96, 214.93], [459.83, 215.07], [450.92, 217.25], [372.23, 179.0], [384.96, 178.7], [459.83, 178.72], [450.92, 179.02]]}, {\"category\": \"car\", \"corners_3d\": [[722.36, 197.92], [693.44, 197.97], [682.26, 195.64], [708.53, 195.6], [722.36, 167.6], [693.44, 167.59], [682.26, 168.07], [708.53, 168.08]]}, {\"category\": \"car\", \"corners_3d\": [[589.61, 203.27], [565.07, 203.31], [566.96, 200.72], [589.42, 200.69], [589.61, 179.43], [565.07, 179.43], [566.96, 178.87], [589.42, 178.87]]}]\n```" - } - ] - }, - { - "id": 3483, - "data_source": "Mono3DRefer", - "scene_name": "006147", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[729.49, 277.86], [835.19, 277.87], [949.42, 331.17], [790.07, 331.14], [729.49, 187.01], [835.19, 187.01], [949.42, 194.19], [790.07, 194.19]]}, {\"category\": \"car\", \"corners_3d\": [[526.04, 215.62], [487.03, 215.62], [500.29, 210.97], [535.06, 210.97], [526.04, 183.93], [487.03, 183.93], [500.29, 182.72], [535.06, 182.72]]}, {\"category\": \"car\", \"corners_3d\": [[355.92, 292.5], [258.06, 292.18], [349.57, 262.3], [423.05, 262.47], [355.92, 201.06], [258.06, 200.99], [349.57, 193.94], [423.05, 193.98]]}, {\"category\": \"car\", \"corners_3d\": [[432.36, 255.91], [362.2, 255.5], [415.33, 239.98], [472.47, 240.25], [432.36, 192.21], [362.2, 192.11], [415.33, 188.5], [472.47, 188.56]]}, {\"category\": \"car\", \"corners_3d\": [[-289.0, 545.7], [-582.47, 541.1], [61.11, 345.74], [201.73, 346.75], [-289.0, 289.93], [-582.47, 288.49], [61.11, 227.14], [201.73, 227.46]]}]\n```" - } - ] - }, - { - "id": 3484, - "data_source": "Mono3DRefer", - "scene_name": "006152", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[142.0, 321.0], [57.86, 334.94], [2.32, 323.47], [85.4, 311.45], [142.0, 150.13], [57.86, 147.12], [2.32, 149.6], [85.4, 152.2]]}]\n```" - } - ] - }, - { - "id": 3485, - "data_source": "Mono3DRefer", - "scene_name": "006155", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[681.11, 252.31], [754.2, 252.24], [800.17, 276.36], [704.9, 276.49], [681.11, 189.36], [754.2, 189.34], [800.17, 194.35], [704.9, 194.38]]}, {\"category\": \"car\", \"corners_3d\": [[521.39, 218.3], [487.53, 218.24], [501.27, 213.88], [531.89, 213.93], [521.39, 190.64], [487.53, 190.62], [501.27, 188.91], [531.89, 188.93]]}, {\"category\": \"car\", \"corners_3d\": [[415.67, 265.78], [345.66, 265.7], [397.68, 247.85], [454.26, 247.9], [415.67, 200.77], [345.66, 200.74], [397.68, 195.38], [454.26, 195.39]]}, {\"category\": \"car\", \"corners_3d\": [[459.36, 244.16], [404.7, 243.94], [440.17, 233.13], [486.59, 233.28], [459.36, 194.84], [404.7, 194.77], [440.17, 191.44], [486.59, 191.48]]}, {\"category\": \"car\", \"corners_3d\": [[197.2, 342.94], [62.08, 342.33], [250.91, 285.32], [340.92, 285.59], [197.2, 229.24], [62.08, 229.04], [250.91, 210.14], [340.92, 210.23]]}]\n```" - } - ] - }, - { - "id": 3486, - "data_source": "Mono3DRefer", - "scene_name": "006158", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[810.59, 326.61], [732.8, 323.1], [736.96, 306.73], [805.71, 309.47], [810.59, 144.51], [732.8, 145.37], [736.96, 149.41], [805.71, 148.73]]}]\n```" - } - ] - }, - { - "id": 3487, - "data_source": "Mono3DRefer", - "scene_name": "006159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[328.87, 226.5], [350.65, 225.76], [357.22, 228.17], [334.53, 228.98], [328.87, 168.19], [350.65, 168.26], [357.22, 168.05], [334.53, 167.97]]}]\n```" - } - ] - }, - { - "id": 3488, - "data_source": "Mono3DRefer", - "scene_name": "006159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-698.51, 435.03], [-355.16, 366.65], [123.38, 368.16], [-52.96, 437.81], [-698.51, 210.91], [-355.16, 200.99], [123.38, 201.21], [-52.96, 211.32]]}, {\"category\": \"car\", \"corners_3d\": [[-45.46, 301.39], [33.31, 282.55], [327.67, 276.1], [301.39, 292.62], [-45.46, 197.64], [33.31, 194.0], [327.67, 192.76], [301.39, 195.95]]}, {\"category\": \"car\", \"corners_3d\": [[1030.98, 247.01], [1115.2, 246.93], [1248.93, 266.27], [1142.87, 266.4], [1030.98, 156.35], [1115.2, 156.37], [1248.93, 152.06], [1142.87, 152.03]]}, {\"category\": \"car\", \"corners_3d\": [[1154.0, 242.75], [1254.42, 242.46], [1500.22, 268.15], [1363.93, 268.69], [1154.0, 151.63], [1254.42, 151.72], [1500.22, 143.92], [1363.93, 143.75]]}, {\"category\": \"car\", \"corners_3d\": [[388.29, 239.61], [365.02, 245.17], [190.12, 243.73], [226.17, 238.37], [388.29, 183.51], [365.02, 184.39], [190.12, 184.16], [226.17, 183.31]]}]\n```" - } - ] - }, - { - "id": 3489, - "data_source": "Mono3DRefer", - "scene_name": "006159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-698.51, 435.03], [-355.16, 366.65], [123.38, 368.16], [-52.96, 437.81], [-698.51, 210.91], [-355.16, 200.99], [123.38, 201.21], [-52.96, 211.32]]}, {\"category\": \"car\", \"corners_3d\": [[-45.46, 301.39], [33.31, 282.55], [327.67, 276.1], [301.39, 292.62], [-45.46, 197.64], [33.31, 194.0], [327.67, 192.76], [301.39, 195.95]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[328.87, 226.5], [350.65, 225.76], [357.22, 228.17], [334.53, 228.98], [328.87, 168.19], [350.65, 168.26], [357.22, 168.05], [334.53, 167.97]]}, {\"category\": \"car\", \"corners_3d\": [[1030.98, 247.01], [1115.2, 246.93], [1248.93, 266.27], [1142.87, 266.4], [1030.98, 156.35], [1115.2, 156.37], [1248.93, 152.06], [1142.87, 152.03]]}, {\"category\": \"car\", \"corners_3d\": [[1154.0, 242.75], [1254.42, 242.46], [1500.22, 268.15], [1363.93, 268.69], [1154.0, 151.63], [1254.42, 151.72], [1500.22, 143.92], [1363.93, 143.75]]}, {\"category\": \"car\", \"corners_3d\": [[388.29, 239.61], [365.02, 245.17], [190.12, 243.73], [226.17, 238.37], [388.29, 183.51], [365.02, 184.39], [190.12, 184.16], [226.17, 183.31]]}]\n```" - } - ] - }, - { - "id": 3490, - "data_source": "Mono3DRefer", - "scene_name": "006181", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-222.29, 864.75], [-899.41, 877.94], [137.55, 388.54], [341.39, 387.29], [-222.29, 309.56], [-899.41, 312.16], [137.55, 215.47], [341.39, 215.22]]}, {\"category\": \"car\", \"corners_3d\": [[568.74, 198.05], [539.93, 198.08], [544.1, 195.86], [570.38, 195.84], [568.74, 167.78], [539.93, 167.77], [544.1, 168.22], [570.38, 168.22]]}, {\"category\": \"car\", \"corners_3d\": [[577.61, 187.1], [562.18, 187.09], [565.29, 186.37], [579.94, 186.37], [577.61, 169.92], [562.18, 169.92], [565.29, 170.07], [579.94, 170.07]]}]\n```" - } - ] - }, - { - "id": 3491, - "data_source": "Mono3DRefer", - "scene_name": "006187", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[643.13, 184.63], [661.78, 184.64], [664.4, 185.67], [644.12, 185.66], [643.13, 162.5], [661.78, 162.49], [664.4, 161.58], [644.12, 161.59]]}]\n```" - } - ] - }, - { - "id": 3492, - "data_source": "Mono3DRefer", - "scene_name": "006187", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-508.28, 405.4], [-259.27, 354.31], [177.67, 356.75], [49.51, 409.42], [-508.28, 214.98], [-259.27, 205.73], [177.67, 206.17], [49.51, 215.71]]}, {\"category\": \"car\", \"corners_3d\": [[-93.99, 327.23], [28.41, 301.72], [322.11, 304.55], [256.14, 331.3], [-93.99, 197.39], [28.41, 193.33], [322.11, 193.78], [256.14, 198.03]]}, {\"category\": \"car\", \"corners_3d\": [[349.68, 277.27], [303.07, 292.94], [25.32, 289.6], [106.16, 274.73], [349.68, 195.25], [303.07, 198.62], [25.32, 197.9], [106.16, 194.71]]}, {\"category\": \"car\", \"corners_3d\": [[385.8, 259.55], [355.24, 269.21], [187.0, 267.59], [233.71, 258.24], [385.8, 185.08], [355.24, 186.45], [187.0, 186.22], [233.71, 184.9]]}, {\"category\": \"car\", \"corners_3d\": [[711.8, 255.12], [793.14, 254.94], [876.02, 289.39], [760.68, 289.76], [711.8, 181.22], [793.14, 181.2], [876.02, 184.71], [760.68, 184.74]]}, {\"category\": \"car\", \"corners_3d\": [[496.0, 224.51], [481.42, 228.27], [369.33, 227.21], [391.11, 223.59], [496.0, 178.85], [481.42, 179.29], [369.33, 179.16], [391.11, 178.74]]}, {\"category\": \"car\", \"corners_3d\": [[671.22, 210.87], [711.43, 210.94], [722.14, 216.06], [676.51, 215.98], [671.22, 174.62], [711.43, 174.62], [722.14, 174.86], [676.51, 174.86]]}, {\"category\": \"car\", \"corners_3d\": [[404.66, 214.66], [413.99, 212.6], [509.82, 212.49], [505.47, 214.54], [404.66, 180.18], [413.99, 179.82], [509.82, 179.8], [505.47, 180.16]]}, {\"category\": \"car\", \"corners_3d\": [[445.3, 207.01], [451.98, 205.55], [522.99, 205.51], [519.49, 206.97], [445.3, 176.96], [451.98, 176.79], [522.99, 176.78], [519.49, 176.96]]}]\n```" - } - ] - }, - { - "id": 3493, - "data_source": "Mono3DRefer", - "scene_name": "006187", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-508.28, 405.4], [-259.27, 354.31], [177.67, 356.75], [49.51, 409.42], [-508.28, 214.98], [-259.27, 205.73], [177.67, 206.17], [49.51, 215.71]]}, {\"category\": \"car\", \"corners_3d\": [[-93.99, 327.23], [28.41, 301.72], [322.11, 304.55], [256.14, 331.3], [-93.99, 197.39], [28.41, 193.33], [322.11, 193.78], [256.14, 198.03]]}, {\"category\": \"car\", \"corners_3d\": [[349.68, 277.27], [303.07, 292.94], [25.32, 289.6], [106.16, 274.73], [349.68, 195.25], [303.07, 198.62], [25.32, 197.9], [106.16, 194.71]]}, {\"category\": \"car\", \"corners_3d\": [[385.8, 259.55], [355.24, 269.21], [187.0, 267.59], [233.71, 258.24], [385.8, 185.08], [355.24, 186.45], [187.0, 186.22], [233.71, 184.9]]}, {\"category\": \"car\", \"corners_3d\": [[711.8, 255.12], [793.14, 254.94], [876.02, 289.39], [760.68, 289.76], [711.8, 181.22], [793.14, 181.2], [876.02, 184.71], [760.68, 184.74]]}, {\"category\": \"car\", \"corners_3d\": [[496.0, 224.51], [481.42, 228.27], [369.33, 227.21], [391.11, 223.59], [496.0, 178.85], [481.42, 179.29], [369.33, 179.16], [391.11, 178.74]]}, {\"category\": \"car\", \"corners_3d\": [[671.22, 210.87], [711.43, 210.94], [722.14, 216.06], [676.51, 215.98], [671.22, 174.62], [711.43, 174.62], [722.14, 174.86], [676.51, 174.86]]}, {\"category\": \"car\", \"corners_3d\": [[404.66, 214.66], [413.99, 212.6], [509.82, 212.49], [505.47, 214.54], [404.66, 180.18], [413.99, 179.82], [509.82, 179.8], [505.47, 180.16]]}, {\"category\": \"car\", \"corners_3d\": [[445.3, 207.01], [451.98, 205.55], [522.99, 205.51], [519.49, 206.97], [445.3, 176.96], [451.98, 176.79], [522.99, 176.78], [519.49, 176.96]]}, {\"category\": \"van\", \"corners_3d\": [[643.13, 184.63], [661.78, 184.64], [664.4, 185.67], [644.12, 185.66], [643.13, 162.5], [661.78, 162.49], [664.4, 161.58], [644.12, 161.59]]}]\n```" - } - ] - }, - { - "id": 3494, - "data_source": "Mono3DRefer", - "scene_name": "006201", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[605.65, 295.8], [603.24, 306.19], [509.87, 305.53], [519.95, 295.24], [605.65, 159.81], [603.24, 157.95], [509.87, 158.07], [519.95, 159.91]]}]\n```" - } - ] - }, - { - "id": 3495, - "data_source": "Mono3DRefer", - "scene_name": "006207", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[517.78, 237.28], [458.7, 237.33], [482.23, 226.73], [531.58, 226.69], [517.78, 187.92], [458.7, 187.94], [482.23, 185.46], [531.58, 185.45]]}]\n```" - } - ] - }, - { - "id": 3496, - "data_source": "Mono3DRefer", - "scene_name": "006214", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[543.35, 206.0], [505.79, 205.74], [531.51, 201.75], [564.67, 201.95], [543.35, 176.12], [505.79, 176.09], [531.51, 175.7], [564.67, 175.72]]}]\n```" - } - ] - }, - { - "id": 3497, - "data_source": "Mono3DRefer", - "scene_name": "006214", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[505.08, 219.03], [444.23, 218.99], [482.21, 208.8], [529.63, 208.82], [505.08, 144.17], [444.23, 144.19], [482.21, 150.52], [529.63, 150.51]]}]\n```" - } - ] - }, - { - "id": 3498, - "data_source": "Mono3DRefer", - "scene_name": "006214", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[715.17, 192.38], [714.5, 192.14], [743.19, 192.19], [744.21, 192.44], [715.17, 163.16], [714.5, 163.28], [743.19, 163.26], [744.21, 163.13]]}]\n```" - } - ] - }, - { - "id": 3499, - "data_source": "Mono3DRefer", - "scene_name": "006214", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[505.08, 219.03], [444.23, 218.99], [482.21, 208.8], [529.63, 208.82], [505.08, 144.17], [444.23, 144.19], [482.21, 150.52], [529.63, 150.51]]}, {\"category\": \"car\", \"corners_3d\": [[543.35, 206.0], [505.79, 205.74], [531.51, 201.75], [564.67, 201.95], [543.35, 176.12], [505.79, 176.09], [531.51, 175.7], [564.67, 175.72]]}]\n```" - } - ] - }, - { - "id": 3500, - "data_source": "Mono3DRefer", - "scene_name": "006214", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[715.17, 192.38], [714.5, 192.14], [743.19, 192.19], [744.21, 192.44], [715.17, 163.16], [714.5, 163.28], [743.19, 163.26], [744.21, 163.13]]}, {\"category\": \"car\", \"corners_3d\": [[543.35, 206.0], [505.79, 205.74], [531.51, 201.75], [564.67, 201.95], [543.35, 176.12], [505.79, 176.09], [531.51, 175.7], [564.67, 175.72]]}]\n```" - } - ] - }, - { - "id": 3501, - "data_source": "Mono3DRefer", - "scene_name": "006214", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[715.17, 192.38], [714.5, 192.14], [743.19, 192.19], [744.21, 192.44], [715.17, 163.16], [714.5, 163.28], [743.19, 163.26], [744.21, 163.13]]}, {\"category\": \"van\", \"corners_3d\": [[505.08, 219.03], [444.23, 218.99], [482.21, 208.8], [529.63, 208.82], [505.08, 144.17], [444.23, 144.19], [482.21, 150.52], [529.63, 150.51]]}]\n```" - } - ] - }, - { - "id": 3502, - "data_source": "Mono3DRefer", - "scene_name": "006214", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[715.17, 192.38], [714.5, 192.14], [743.19, 192.19], [744.21, 192.44], [715.17, 163.16], [714.5, 163.28], [743.19, 163.26], [744.21, 163.13]]}, {\"category\": \"van\", \"corners_3d\": [[505.08, 219.03], [444.23, 218.99], [482.21, 208.8], [529.63, 208.82], [505.08, 144.17], [444.23, 144.19], [482.21, 150.52], [529.63, 150.51]]}, {\"category\": \"car\", \"corners_3d\": [[543.35, 206.0], [505.79, 205.74], [531.51, 201.75], [564.67, 201.95], [543.35, 176.12], [505.79, 176.09], [531.51, 175.7], [564.67, 175.72]]}]\n```" - } - ] - }, - { - "id": 3503, - "data_source": "Mono3DRefer", - "scene_name": "006216", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[551.14, 216.15], [507.2, 216.23], [517.48, 210.66], [555.78, 210.6], [551.14, 170.04], [507.2, 170.03], [517.48, 170.39], [555.78, 170.4]]}, {\"category\": \"car\", \"corners_3d\": [[569.84, 198.31], [549.9, 198.3], [554.65, 196.65], [573.3, 196.67], [569.84, 176.11], [549.9, 176.11], [554.65, 175.9], [573.3, 175.9]]}]\n```" - } - ] - }, - { - "id": 3504, - "data_source": "Mono3DRefer", - "scene_name": "006222", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[610.67, 299.98], [611.12, 308.11], [514.87, 308.11], [520.55, 299.98], [610.67, 159.06], [611.12, 157.6], [514.87, 157.6], [520.55, 159.06]]}]\n```" - } - ] - }, - { - "id": 3505, - "data_source": "Mono3DRefer", - "scene_name": "006223", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[444.96, 299.55], [327.67, 299.36], [402.44, 266.58], [489.38, 266.68], [444.96, 199.83], [327.67, 199.79], [402.44, 192.81], [489.38, 192.84]]}]\n```" - } - ] - }, - { - "id": 3506, - "data_source": "Mono3DRefer", - "scene_name": "006239", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[462.76, 288.55], [342.51, 288.57], [417.81, 255.87], [504.08, 255.86], [462.76, 193.4], [342.51, 193.41], [417.81, 187.6], [504.08, 187.6]]}]\n```" - } - ] - }, - { - "id": 3507, - "data_source": "Mono3DRefer", - "scene_name": "006240", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[694.44, 312.63], [705.91, 322.03], [578.26, 324.63], [575.22, 314.89], [694.44, 158.14], [705.91, 156.55], [578.26, 156.11], [575.22, 157.76]]}]\n```" - } - ] - }, - { - "id": 3508, - "data_source": "Mono3DRefer", - "scene_name": "006262", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[665.19, 213.97], [706.11, 214.09], [715.36, 220.47], [668.11, 220.31], [665.19, 167.6], [706.11, 167.58], [715.36, 166.77], [668.11, 166.79]]}, {\"category\": \"van\", \"corners_3d\": [[439.43, 208.9], [414.46, 208.88], [432.03, 205.85], [454.91, 205.87], [439.43, 181.38], [414.46, 181.37], [432.03, 180.66], [454.91, 180.66]]}]\n```" - } - ] - }, - { - "id": 3509, - "data_source": "Mono3DRefer", - "scene_name": "006262", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[460.83, 219.45], [487.03, 217.7], [543.18, 220.76], [517.46, 222.76], [460.83, 184.98], [487.03, 184.53], [543.18, 185.33], [517.46, 185.85]]}, {\"category\": \"car\", \"corners_3d\": [[484.94, 211.29], [504.95, 210.05], [561.42, 212.27], [542.17, 213.67], [484.94, 186.09], [504.95, 185.66], [561.42, 186.42], [542.17, 186.9]]}, {\"category\": \"car\", \"corners_3d\": [[590.76, 219.15], [629.81, 219.3], [626.66, 225.46], [582.47, 225.27], [590.76, 183.6], [629.81, 183.64], [626.66, 185.07], [582.47, 185.02]]}, {\"category\": \"car\", \"corners_3d\": [[599.39, 241.09], [664.55, 241.09], [676.55, 256.13], [597.02, 256.13], [599.39, 188.65], [664.55, 188.65], [676.55, 192.13], [597.02, 192.13]]}, {\"category\": \"car\", \"corners_3d\": [[352.04, 226.89], [310.51, 226.83], [346.52, 220.62], [383.31, 220.67], [352.04, 188.18], [310.51, 188.17], [346.52, 186.4], [383.31, 186.42]]}]\n```" - } - ] - }, - { - "id": 3510, - "data_source": "Mono3DRefer", - "scene_name": "006262", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[525.37, 230.06], [525.64, 231.57], [505.86, 231.78], [506.11, 230.25], [525.37, 175.19], [525.64, 175.25], [505.86, 175.26], [506.11, 175.2]]}]\n```" - } - ] - }, - { - "id": 3511, - "data_source": "Mono3DRefer", - "scene_name": "006262", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[795.13, 232.32], [790.1, 230.09], [855.05, 230.45], [862.67, 232.72], [795.13, 177.11], [790.1, 176.95], [855.05, 176.98], [862.67, 177.14]]}]\n```" - } - ] - }, - { - "id": 3512, - "data_source": "Mono3DRefer", - "scene_name": "006262", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[460.83, 219.45], [487.03, 217.7], [543.18, 220.76], [517.46, 222.76], [460.83, 184.98], [487.03, 184.53], [543.18, 185.33], [517.46, 185.85]]}, {\"category\": \"car\", \"corners_3d\": [[484.94, 211.29], [504.95, 210.05], [561.42, 212.27], [542.17, 213.67], [484.94, 186.09], [504.95, 185.66], [561.42, 186.42], [542.17, 186.9]]}, {\"category\": \"car\", \"corners_3d\": [[590.76, 219.15], [629.81, 219.3], [626.66, 225.46], [582.47, 225.27], [590.76, 183.6], [629.81, 183.64], [626.66, 185.07], [582.47, 185.02]]}, {\"category\": \"car\", \"corners_3d\": [[599.39, 241.09], [664.55, 241.09], [676.55, 256.13], [597.02, 256.13], [599.39, 188.65], [664.55, 188.65], [676.55, 192.13], [597.02, 192.13]]}, {\"category\": \"car\", \"corners_3d\": [[352.04, 226.89], [310.51, 226.83], [346.52, 220.62], [383.31, 220.67], [352.04, 188.18], [310.51, 188.17], [346.52, 186.4], [383.31, 186.42]]}, {\"category\": \"van\", \"corners_3d\": [[665.19, 213.97], [706.11, 214.09], [715.36, 220.47], [668.11, 220.31], [665.19, 167.6], [706.11, 167.58], [715.36, 166.77], [668.11, 166.79]]}, {\"category\": \"van\", \"corners_3d\": [[439.43, 208.9], [414.46, 208.88], [432.03, 205.85], [454.91, 205.87], [439.43, 181.38], [414.46, 181.37], [432.03, 180.66], [454.91, 180.66]]}]\n```" - } - ] - }, - { - "id": 3513, - "data_source": "Mono3DRefer", - "scene_name": "006262", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[665.19, 213.97], [706.11, 214.09], [715.36, 220.47], [668.11, 220.31], [665.19, 167.6], [706.11, 167.58], [715.36, 166.77], [668.11, 166.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[525.37, 230.06], [525.64, 231.57], [505.86, 231.78], [506.11, 230.25], [525.37, 175.19], [525.64, 175.25], [505.86, 175.26], [506.11, 175.2]]}, {\"category\": \"van\", \"corners_3d\": [[439.43, 208.9], [414.46, 208.88], [432.03, 205.85], [454.91, 205.87], [439.43, 181.38], [414.46, 181.37], [432.03, 180.66], [454.91, 180.66]]}]\n```" - } - ] - }, - { - "id": 3514, - "data_source": "Mono3DRefer", - "scene_name": "006262", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[795.13, 232.32], [790.1, 230.09], [855.05, 230.45], [862.67, 232.72], [795.13, 177.11], [790.1, 176.95], [855.05, 176.98], [862.67, 177.14]]}, {\"category\": \"van\", \"corners_3d\": [[665.19, 213.97], [706.11, 214.09], [715.36, 220.47], [668.11, 220.31], [665.19, 167.6], [706.11, 167.58], [715.36, 166.77], [668.11, 166.79]]}, {\"category\": \"van\", \"corners_3d\": [[439.43, 208.9], [414.46, 208.88], [432.03, 205.85], [454.91, 205.87], [439.43, 181.38], [414.46, 181.37], [432.03, 180.66], [454.91, 180.66]]}]\n```" - } - ] - }, - { - "id": 3515, - "data_source": "Mono3DRefer", - "scene_name": "006262", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[460.83, 219.45], [487.03, 217.7], [543.18, 220.76], [517.46, 222.76], [460.83, 184.98], [487.03, 184.53], [543.18, 185.33], [517.46, 185.85]]}, {\"category\": \"car\", \"corners_3d\": [[484.94, 211.29], [504.95, 210.05], [561.42, 212.27], [542.17, 213.67], [484.94, 186.09], [504.95, 185.66], [561.42, 186.42], [542.17, 186.9]]}, {\"category\": \"car\", \"corners_3d\": [[590.76, 219.15], [629.81, 219.3], [626.66, 225.46], [582.47, 225.27], [590.76, 183.6], [629.81, 183.64], [626.66, 185.07], [582.47, 185.02]]}, {\"category\": \"car\", \"corners_3d\": [[599.39, 241.09], [664.55, 241.09], [676.55, 256.13], [597.02, 256.13], [599.39, 188.65], [664.55, 188.65], [676.55, 192.13], [597.02, 192.13]]}, {\"category\": \"car\", \"corners_3d\": [[352.04, 226.89], [310.51, 226.83], [346.52, 220.62], [383.31, 220.67], [352.04, 188.18], [310.51, 188.17], [346.52, 186.4], [383.31, 186.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[525.37, 230.06], [525.64, 231.57], [505.86, 231.78], [506.11, 230.25], [525.37, 175.19], [525.64, 175.25], [505.86, 175.26], [506.11, 175.2]]}]\n```" - } - ] - }, - { - "id": 3516, - "data_source": "Mono3DRefer", - "scene_name": "006262", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[460.83, 219.45], [487.03, 217.7], [543.18, 220.76], [517.46, 222.76], [460.83, 184.98], [487.03, 184.53], [543.18, 185.33], [517.46, 185.85]]}, {\"category\": \"car\", \"corners_3d\": [[484.94, 211.29], [504.95, 210.05], [561.42, 212.27], [542.17, 213.67], [484.94, 186.09], [504.95, 185.66], [561.42, 186.42], [542.17, 186.9]]}, {\"category\": \"car\", \"corners_3d\": [[590.76, 219.15], [629.81, 219.3], [626.66, 225.46], [582.47, 225.27], [590.76, 183.6], [629.81, 183.64], [626.66, 185.07], [582.47, 185.02]]}, {\"category\": \"car\", \"corners_3d\": [[599.39, 241.09], [664.55, 241.09], [676.55, 256.13], [597.02, 256.13], [599.39, 188.65], [664.55, 188.65], [676.55, 192.13], [597.02, 192.13]]}, {\"category\": \"car\", \"corners_3d\": [[352.04, 226.89], [310.51, 226.83], [346.52, 220.62], [383.31, 220.67], [352.04, 188.18], [310.51, 188.17], [346.52, 186.4], [383.31, 186.42]]}, {\"category\": \"cyclist\", \"corners_3d\": [[795.13, 232.32], [790.1, 230.09], [855.05, 230.45], [862.67, 232.72], [795.13, 177.11], [790.1, 176.95], [855.05, 176.98], [862.67, 177.14]]}]\n```" - } - ] - }, - { - "id": 3517, - "data_source": "Mono3DRefer", - "scene_name": "006262", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[795.13, 232.32], [790.1, 230.09], [855.05, 230.45], [862.67, 232.72], [795.13, 177.11], [790.1, 176.95], [855.05, 176.98], [862.67, 177.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[525.37, 230.06], [525.64, 231.57], [505.86, 231.78], [506.11, 230.25], [525.37, 175.19], [525.64, 175.25], [505.86, 175.26], [506.11, 175.2]]}]\n```" - } - ] - }, - { - "id": 3518, - "data_source": "Mono3DRefer", - "scene_name": "006262", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[460.83, 219.45], [487.03, 217.7], [543.18, 220.76], [517.46, 222.76], [460.83, 184.98], [487.03, 184.53], [543.18, 185.33], [517.46, 185.85]]}, {\"category\": \"car\", \"corners_3d\": [[484.94, 211.29], [504.95, 210.05], [561.42, 212.27], [542.17, 213.67], [484.94, 186.09], [504.95, 185.66], [561.42, 186.42], [542.17, 186.9]]}, {\"category\": \"car\", \"corners_3d\": [[590.76, 219.15], [629.81, 219.3], [626.66, 225.46], [582.47, 225.27], [590.76, 183.6], [629.81, 183.64], [626.66, 185.07], [582.47, 185.02]]}, {\"category\": \"car\", \"corners_3d\": [[599.39, 241.09], [664.55, 241.09], [676.55, 256.13], [597.02, 256.13], [599.39, 188.65], [664.55, 188.65], [676.55, 192.13], [597.02, 192.13]]}, {\"category\": \"car\", \"corners_3d\": [[352.04, 226.89], [310.51, 226.83], [346.52, 220.62], [383.31, 220.67], [352.04, 188.18], [310.51, 188.17], [346.52, 186.4], [383.31, 186.42]]}, {\"category\": \"van\", \"corners_3d\": [[665.19, 213.97], [706.11, 214.09], [715.36, 220.47], [668.11, 220.31], [665.19, 167.6], [706.11, 167.58], [715.36, 166.77], [668.11, 166.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[525.37, 230.06], [525.64, 231.57], [505.86, 231.78], [506.11, 230.25], [525.37, 175.19], [525.64, 175.25], [505.86, 175.26], [506.11, 175.2]]}, {\"category\": \"van\", \"corners_3d\": [[439.43, 208.9], [414.46, 208.88], [432.03, 205.85], [454.91, 205.87], [439.43, 181.38], [414.46, 181.37], [432.03, 180.66], [454.91, 180.66]]}]\n```" - } - ] - }, - { - "id": 3519, - "data_source": "Mono3DRefer", - "scene_name": "006262", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[460.83, 219.45], [487.03, 217.7], [543.18, 220.76], [517.46, 222.76], [460.83, 184.98], [487.03, 184.53], [543.18, 185.33], [517.46, 185.85]]}, {\"category\": \"car\", \"corners_3d\": [[484.94, 211.29], [504.95, 210.05], [561.42, 212.27], [542.17, 213.67], [484.94, 186.09], [504.95, 185.66], [561.42, 186.42], [542.17, 186.9]]}, {\"category\": \"car\", \"corners_3d\": [[590.76, 219.15], [629.81, 219.3], [626.66, 225.46], [582.47, 225.27], [590.76, 183.6], [629.81, 183.64], [626.66, 185.07], [582.47, 185.02]]}, {\"category\": \"car\", \"corners_3d\": [[599.39, 241.09], [664.55, 241.09], [676.55, 256.13], [597.02, 256.13], [599.39, 188.65], [664.55, 188.65], [676.55, 192.13], [597.02, 192.13]]}, {\"category\": \"car\", \"corners_3d\": [[352.04, 226.89], [310.51, 226.83], [346.52, 220.62], [383.31, 220.67], [352.04, 188.18], [310.51, 188.17], [346.52, 186.4], [383.31, 186.42]]}, {\"category\": \"cyclist\", \"corners_3d\": [[795.13, 232.32], [790.1, 230.09], [855.05, 230.45], [862.67, 232.72], [795.13, 177.11], [790.1, 176.95], [855.05, 176.98], [862.67, 177.14]]}, {\"category\": \"van\", \"corners_3d\": [[665.19, 213.97], [706.11, 214.09], [715.36, 220.47], [668.11, 220.31], [665.19, 167.6], [706.11, 167.58], [715.36, 166.77], [668.11, 166.79]]}, {\"category\": \"van\", \"corners_3d\": [[439.43, 208.9], [414.46, 208.88], [432.03, 205.85], [454.91, 205.87], [439.43, 181.38], [414.46, 181.37], [432.03, 180.66], [454.91, 180.66]]}]\n```" - } - ] - }, - { - "id": 3520, - "data_source": "Mono3DRefer", - "scene_name": "006262", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[795.13, 232.32], [790.1, 230.09], [855.05, 230.45], [862.67, 232.72], [795.13, 177.11], [790.1, 176.95], [855.05, 176.98], [862.67, 177.14]]}, {\"category\": \"van\", \"corners_3d\": [[665.19, 213.97], [706.11, 214.09], [715.36, 220.47], [668.11, 220.31], [665.19, 167.6], [706.11, 167.58], [715.36, 166.77], [668.11, 166.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[525.37, 230.06], [525.64, 231.57], [505.86, 231.78], [506.11, 230.25], [525.37, 175.19], [525.64, 175.25], [505.86, 175.26], [506.11, 175.2]]}, {\"category\": \"van\", \"corners_3d\": [[439.43, 208.9], [414.46, 208.88], [432.03, 205.85], [454.91, 205.87], [439.43, 181.38], [414.46, 181.37], [432.03, 180.66], [454.91, 180.66]]}]\n```" - } - ] - }, - { - "id": 3521, - "data_source": "Mono3DRefer", - "scene_name": "006262", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[460.83, 219.45], [487.03, 217.7], [543.18, 220.76], [517.46, 222.76], [460.83, 184.98], [487.03, 184.53], [543.18, 185.33], [517.46, 185.85]]}, {\"category\": \"car\", \"corners_3d\": [[484.94, 211.29], [504.95, 210.05], [561.42, 212.27], [542.17, 213.67], [484.94, 186.09], [504.95, 185.66], [561.42, 186.42], [542.17, 186.9]]}, {\"category\": \"car\", \"corners_3d\": [[590.76, 219.15], [629.81, 219.3], [626.66, 225.46], [582.47, 225.27], [590.76, 183.6], [629.81, 183.64], [626.66, 185.07], [582.47, 185.02]]}, {\"category\": \"car\", \"corners_3d\": [[599.39, 241.09], [664.55, 241.09], [676.55, 256.13], [597.02, 256.13], [599.39, 188.65], [664.55, 188.65], [676.55, 192.13], [597.02, 192.13]]}, {\"category\": \"car\", \"corners_3d\": [[352.04, 226.89], [310.51, 226.83], [346.52, 220.62], [383.31, 220.67], [352.04, 188.18], [310.51, 188.17], [346.52, 186.4], [383.31, 186.42]]}, {\"category\": \"cyclist\", \"corners_3d\": [[795.13, 232.32], [790.1, 230.09], [855.05, 230.45], [862.67, 232.72], [795.13, 177.11], [790.1, 176.95], [855.05, 176.98], [862.67, 177.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[525.37, 230.06], [525.64, 231.57], [505.86, 231.78], [506.11, 230.25], [525.37, 175.19], [525.64, 175.25], [505.86, 175.26], [506.11, 175.2]]}]\n```" - } - ] - }, - { - "id": 3522, - "data_source": "Mono3DRefer", - "scene_name": "006262", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[460.83, 219.45], [487.03, 217.7], [543.18, 220.76], [517.46, 222.76], [460.83, 184.98], [487.03, 184.53], [543.18, 185.33], [517.46, 185.85]]}, {\"category\": \"car\", \"corners_3d\": [[484.94, 211.29], [504.95, 210.05], [561.42, 212.27], [542.17, 213.67], [484.94, 186.09], [504.95, 185.66], [561.42, 186.42], [542.17, 186.9]]}, {\"category\": \"car\", \"corners_3d\": [[590.76, 219.15], [629.81, 219.3], [626.66, 225.46], [582.47, 225.27], [590.76, 183.6], [629.81, 183.64], [626.66, 185.07], [582.47, 185.02]]}, {\"category\": \"car\", \"corners_3d\": [[599.39, 241.09], [664.55, 241.09], [676.55, 256.13], [597.02, 256.13], [599.39, 188.65], [664.55, 188.65], [676.55, 192.13], [597.02, 192.13]]}, {\"category\": \"car\", \"corners_3d\": [[352.04, 226.89], [310.51, 226.83], [346.52, 220.62], [383.31, 220.67], [352.04, 188.18], [310.51, 188.17], [346.52, 186.4], [383.31, 186.42]]}, {\"category\": \"cyclist\", \"corners_3d\": [[795.13, 232.32], [790.1, 230.09], [855.05, 230.45], [862.67, 232.72], [795.13, 177.11], [790.1, 176.95], [855.05, 176.98], [862.67, 177.14]]}, {\"category\": \"van\", \"corners_3d\": [[665.19, 213.97], [706.11, 214.09], [715.36, 220.47], [668.11, 220.31], [665.19, 167.6], [706.11, 167.58], [715.36, 166.77], [668.11, 166.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[525.37, 230.06], [525.64, 231.57], [505.86, 231.78], [506.11, 230.25], [525.37, 175.19], [525.64, 175.25], [505.86, 175.26], [506.11, 175.2]]}, {\"category\": \"van\", \"corners_3d\": [[439.43, 208.9], [414.46, 208.88], [432.03, 205.85], [454.91, 205.87], [439.43, 181.38], [414.46, 181.37], [432.03, 180.66], [454.91, 180.66]]}]\n```" - } - ] - }, - { - "id": 3523, - "data_source": "Mono3DRefer", - "scene_name": "006263", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[231.05, 311.49], [119.02, 311.72], [244.41, 275.65], [327.22, 275.52], [231.05, 208.94], [119.02, 209.0], [244.41, 199.61], [327.22, 199.58]]}, {\"category\": \"car\", \"corners_3d\": [[700.21, 239.26], [773.79, 239.27], [820.02, 258.03], [725.65, 258.02], [700.21, 175.78], [773.79, 175.78], [820.02, 176.61], [725.65, 176.61]]}, {\"category\": \"car\", \"corners_3d\": [[366.31, 261.53], [295.73, 261.45], [362.1, 243.1], [418.1, 243.15], [366.31, 197.71], [295.73, 197.69], [362.1, 192.54], [418.1, 192.56]]}, {\"category\": \"car\", \"corners_3d\": [[455.53, 233.13], [405.02, 233.13], [435.3, 224.18], [478.3, 224.18], [455.53, 179.88], [405.02, 179.88], [435.3, 178.84], [478.3, 178.84]]}, {\"category\": \"car\", \"corners_3d\": [[484.44, 221.48], [445.52, 221.43], [466.26, 215.76], [500.66, 215.8], [484.44, 183.1], [445.52, 183.09], [466.26, 181.89], [500.66, 181.9]]}, {\"category\": \"car\", \"corners_3d\": [[637.4, 192.88], [655.38, 192.88], [657.91, 194.21], [638.74, 194.21], [637.4, 176.1], [655.38, 176.1], [657.91, 176.31], [638.74, 176.31]]}]\n```" - } - ] - }, - { - "id": 3524, - "data_source": "Mono3DRefer", - "scene_name": "006278", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[480.03, 296.92], [356.16, 297.99], [425.1, 258.19], [509.39, 257.68], [480.03, 187.78], [356.16, 187.91], [425.1, 183.12], [509.39, 183.06]]}, {\"category\": \"car\", \"corners_3d\": [[549.53, 230.23], [487.5, 229.89], [517.53, 219.93], [568.84, 220.17], [549.53, 171.46], [487.5, 171.47], [517.53, 171.71], [568.84, 171.7]]}, {\"category\": \"car\", \"corners_3d\": [[664.14, 200.41], [689.96, 200.96], [669.01, 202.68], [642.09, 202.05], [664.14, 172.13], [689.96, 172.11], [669.01, 172.07], [642.09, 172.08]]}]\n```" - } - ] - }, - { - "id": 3525, - "data_source": "Mono3DRefer", - "scene_name": "006278", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-417.22, 1036.21], [-1507.88, 1086.55], [77.25, 392.9], [329.43, 389.86], [-417.22, 56.11], [-1507.88, 49.31], [77.25, 143.1], [329.43, 143.51]]}, {\"category\": \"van\", \"corners_3d\": [[884.45, 423.01], [1125.97, 419.67], [3101.13, 1314.66], [2056.1, 1389.69], [884.45, 152.97], [1125.97, 153.23], [3101.13, 82.08], [2056.1, 76.11]]}, {\"category\": \"van\", \"corners_3d\": [[521.58, 252.77], [443.12, 252.95], [476.02, 235.55], [537.4, 235.44], [521.58, 167.78], [443.12, 167.77], [476.02, 168.88], [537.4, 168.88]]}]\n```" - } - ] - }, - { - "id": 3526, - "data_source": "Mono3DRefer", - "scene_name": "006278", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-417.22, 1036.21], [-1507.88, 1086.55], [77.25, 392.9], [329.43, 389.86], [-417.22, 56.11], [-1507.88, 49.31], [77.25, 143.1], [329.43, 143.51]]}, {\"category\": \"van\", \"corners_3d\": [[884.45, 423.01], [1125.97, 419.67], [3101.13, 1314.66], [2056.1, 1389.69], [884.45, 152.97], [1125.97, 153.23], [3101.13, 82.08], [2056.1, 76.11]]}, {\"category\": \"car\", \"corners_3d\": [[480.03, 296.92], [356.16, 297.99], [425.1, 258.19], [509.39, 257.68], [480.03, 187.78], [356.16, 187.91], [425.1, 183.12], [509.39, 183.06]]}, {\"category\": \"van\", \"corners_3d\": [[521.58, 252.77], [443.12, 252.95], [476.02, 235.55], [537.4, 235.44], [521.58, 167.78], [443.12, 167.77], [476.02, 168.88], [537.4, 168.88]]}, {\"category\": \"car\", \"corners_3d\": [[549.53, 230.23], [487.5, 229.89], [517.53, 219.93], [568.84, 220.17], [549.53, 171.46], [487.5, 171.47], [517.53, 171.71], [568.84, 171.7]]}, {\"category\": \"car\", \"corners_3d\": [[664.14, 200.41], [689.96, 200.96], [669.01, 202.68], [642.09, 202.05], [664.14, 172.13], [689.96, 172.11], [669.01, 172.07], [642.09, 172.08]]}]\n```" - } - ] - }, - { - "id": 3527, - "data_source": "Mono3DRefer", - "scene_name": "006283", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[235.93, 177.66], [265.96, 177.65], [232.66, 178.17], [199.3, 178.18], [235.93, 151.54], [265.96, 151.58], [232.66, 149.29], [199.3, 149.23]]}, {\"category\": \"car\", \"corners_3d\": [[150.9, 194.41], [106.28, 194.46], [162.33, 191.91], [201.55, 191.87], [150.9, 154.64], [106.28, 154.59], [162.33, 156.75], [201.55, 156.79]]}]\n```" - } - ] - }, - { - "id": 3528, - "data_source": "Mono3DRefer", - "scene_name": "006296", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[11.53, 356.42], [-145.83, 356.45], [101.95, 296.18], [207.63, 296.17], [11.53, 211.88], [-145.83, 211.89], [101.95, 199.08], [207.63, 199.07]]}, {\"category\": \"car\", \"corners_3d\": [[614.55, 194.15], [636.26, 194.16], [637.64, 195.71], [614.35, 195.7], [614.55, 173.65], [636.26, 173.65], [637.64, 173.71], [614.35, 173.71]]}, {\"category\": \"car\", \"corners_3d\": [[500.85, 207.86], [474.02, 207.64], [495.08, 204.81], [519.84, 205.0], [500.85, 180.28], [474.02, 180.23], [495.08, 179.63], [519.84, 179.67]]}, {\"category\": \"car\", \"corners_3d\": [[522.86, 204.02], [498.43, 203.86], [516.0, 201.38], [538.54, 201.51], [522.86, 173.33], [498.43, 173.32], [516.0, 173.29], [538.54, 173.29]]}, {\"category\": \"car\", \"corners_3d\": [[656.86, 193.76], [679.64, 193.81], [681.11, 195.44], [656.56, 195.39], [656.86, 172.31], [679.64, 172.3], [681.11, 172.26], [656.56, 172.26]]}]\n```" - } - ] - }, - { - "id": 3529, - "data_source": "Mono3DRefer", - "scene_name": "006299", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[344.14, 299.95], [357.75, 293.49], [455.51, 293.31], [447.5, 299.75], [344.14, 135.0], [357.75, 137.46], [455.51, 137.53], [447.5, 135.08]]}]\n```" - } - ] - }, - { - "id": 3530, - "data_source": "Mono3DRefer", - "scene_name": "006309", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[526.2, 245.62], [456.54, 245.2], [496.2, 230.54], [551.87, 230.81], [526.2, 185.04], [456.54, 184.97], [496.2, 182.51], [551.87, 182.56]]}]\n```" - } - ] - }, - { - "id": 3531, - "data_source": "Mono3DRefer", - "scene_name": "006310", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[611.87, 208.05], [650.57, 208.42], [634.15, 214.35], [589.26, 213.86], [611.87, 168.61], [650.57, 168.57], [634.15, 167.85], [589.26, 167.91]]}, {\"category\": \"car\", \"corners_3d\": [[240.99, 255.53], [278.63, 251.97], [328.48, 257.63], [290.43, 261.72], [240.99, 206.49], [278.63, 205.04], [328.48, 207.34], [290.43, 209.01]]}, {\"category\": \"car\", \"corners_3d\": [[120.41, 269.4], [185.18, 262.98], [242.31, 270.91], [175.85, 278.56], [120.41, 209.92], [185.18, 207.45], [242.31, 210.5], [175.85, 213.43]]}, {\"category\": \"car\", \"corners_3d\": [[1159.29, 229.26], [1215.71, 235.93], [996.3, 234.15], [963.91, 227.84], [1159.29, 159.67], [1215.71, 158.11], [996.3, 158.53], [963.91, 160.0]]}, {\"category\": \"car\", \"corners_3d\": [[1033.96, 217.58], [1065.21, 221.55], [871.71, 220.32], [856.6, 216.54], [1033.96, 160.39], [1065.21, 159.29], [871.71, 159.63], [856.6, 160.68]]}, {\"category\": \"car\", \"corners_3d\": [[668.34, 194.22], [637.13, 194.06], [646.51, 192.13], [674.92, 192.27], [668.34, 164.22], [637.13, 164.29], [646.51, 165.07], [674.92, 165.01]]}, {\"category\": \"car\", \"corners_3d\": [[699.95, 176.74], [681.99, 176.72], [685.3, 176.5], [702.25, 176.52], [699.95, 161.4], [681.99, 161.45], [685.3, 162.1], [702.25, 162.06]]}]\n```" - } - ] - }, - { - "id": 3532, - "data_source": "Mono3DRefer", - "scene_name": "006311", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[517.52, 220.81], [476.12, 220.87], [490.37, 215.09], [526.78, 215.05], [517.52, 180.04], [476.12, 180.05], [490.37, 179.18], [526.78, 179.18]]}]\n```" - } - ] - }, - { - "id": 3533, - "data_source": "Mono3DRefer", - "scene_name": "006317", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[260.45, 230.38], [284.97, 226.44], [390.2, 226.53], [373.4, 230.49], [260.45, 175.89], [284.97, 175.68], [390.2, 175.68], [373.4, 175.89]]}, {\"category\": \"car\", \"corners_3d\": [[742.34, 203.73], [705.9, 203.84], [688.99, 200.29], [721.33, 200.2], [742.34, 165.34], [705.9, 165.31], [688.99, 166.18], [721.33, 166.2]]}, {\"category\": \"car\", \"corners_3d\": [[573.8, 206.07], [542.96, 206.14], [546.08, 202.67], [573.71, 202.61], [573.8, 176.15], [542.96, 176.16], [546.08, 175.81], [573.71, 175.81]]}]\n```" - } - ] - }, - { - "id": 3534, - "data_source": "Mono3DRefer", - "scene_name": "006329", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[995.84, 297.08], [925.0, 276.31], [1181.44, 274.36], [1301.04, 294.29], [995.84, 148.49], [925.0, 152.57], [1181.44, 152.95], [1301.04, 149.04]]}, {\"category\": \"car\", \"corners_3d\": [[-390.86, 714.11], [-1017.73, 762.97], [18.4, 365.86], [205.8, 360.32], [-390.86, 215.32], [-1017.73, 219.16], [18.4, 188.0], [205.8, 187.56]]}, {\"category\": \"car\", \"corners_3d\": [[267.85, 333.83], [111.62, 336.22], [269.49, 278.49], [369.53, 277.49], [267.85, 189.91], [111.62, 190.16], [269.49, 184.05], [369.53, 183.94]]}]\n```" - } - ] - }, - { - "id": 3535, - "data_source": "Mono3DRefer", - "scene_name": "006329", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[893.27, 271.82], [847.64, 257.42], [1132.61, 256.02], [1224.9, 269.89], [893.27, 148.25], [847.64, 151.83], [1132.61, 152.18], [1224.9, 148.73]]}]\n```" - } - ] - }, - { - "id": 3536, - "data_source": "Mono3DRefer", - "scene_name": "006329", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[659.81, 207.58], [718.13, 209.02], [604.17, 223.9], [529.51, 221.07], [659.81, 133.03], [718.13, 131.37], [604.17, 114.31], [529.51, 117.56]]}]\n```" - } - ] - }, - { - "id": 3537, - "data_source": "Mono3DRefer", - "scene_name": "006329", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[995.84, 297.08], [925.0, 276.31], [1181.44, 274.36], [1301.04, 294.29], [995.84, 148.49], [925.0, 152.57], [1181.44, 152.95], [1301.04, 149.04]]}, {\"category\": \"van\", \"corners_3d\": [[893.27, 271.82], [847.64, 257.42], [1132.61, 256.02], [1224.9, 269.89], [893.27, 148.25], [847.64, 151.83], [1132.61, 152.18], [1224.9, 148.73]]}, {\"category\": \"car\", \"corners_3d\": [[-390.86, 714.11], [-1017.73, 762.97], [18.4, 365.86], [205.8, 360.32], [-390.86, 215.32], [-1017.73, 219.16], [18.4, 188.0], [205.8, 187.56]]}, {\"category\": \"car\", \"corners_3d\": [[267.85, 333.83], [111.62, 336.22], [269.49, 278.49], [369.53, 277.49], [267.85, 189.91], [111.62, 190.16], [269.49, 184.05], [369.53, 183.94]]}]\n```" - } - ] - }, - { - "id": 3538, - "data_source": "Mono3DRefer", - "scene_name": "006329", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[995.84, 297.08], [925.0, 276.31], [1181.44, 274.36], [1301.04, 294.29], [995.84, 148.49], [925.0, 152.57], [1181.44, 152.95], [1301.04, 149.04]]}, {\"category\": \"car\", \"corners_3d\": [[-390.86, 714.11], [-1017.73, 762.97], [18.4, 365.86], [205.8, 360.32], [-390.86, 215.32], [-1017.73, 219.16], [18.4, 188.0], [205.8, 187.56]]}, {\"category\": \"car\", \"corners_3d\": [[267.85, 333.83], [111.62, 336.22], [269.49, 278.49], [369.53, 277.49], [267.85, 189.91], [111.62, 190.16], [269.49, 184.05], [369.53, 183.94]]}, {\"category\": \"truck\", \"corners_3d\": [[659.81, 207.58], [718.13, 209.02], [604.17, 223.9], [529.51, 221.07], [659.81, 133.03], [718.13, 131.37], [604.17, 114.31], [529.51, 117.56]]}]\n```" - } - ] - }, - { - "id": 3539, - "data_source": "Mono3DRefer", - "scene_name": "006329", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[893.27, 271.82], [847.64, 257.42], [1132.61, 256.02], [1224.9, 269.89], [893.27, 148.25], [847.64, 151.83], [1132.61, 152.18], [1224.9, 148.73]]}, {\"category\": \"truck\", \"corners_3d\": [[659.81, 207.58], [718.13, 209.02], [604.17, 223.9], [529.51, 221.07], [659.81, 133.03], [718.13, 131.37], [604.17, 114.31], [529.51, 117.56]]}]\n```" - } - ] - }, - { - "id": 3540, - "data_source": "Mono3DRefer", - "scene_name": "006329", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[995.84, 297.08], [925.0, 276.31], [1181.44, 274.36], [1301.04, 294.29], [995.84, 148.49], [925.0, 152.57], [1181.44, 152.95], [1301.04, 149.04]]}, {\"category\": \"van\", \"corners_3d\": [[893.27, 271.82], [847.64, 257.42], [1132.61, 256.02], [1224.9, 269.89], [893.27, 148.25], [847.64, 151.83], [1132.61, 152.18], [1224.9, 148.73]]}, {\"category\": \"car\", \"corners_3d\": [[-390.86, 714.11], [-1017.73, 762.97], [18.4, 365.86], [205.8, 360.32], [-390.86, 215.32], [-1017.73, 219.16], [18.4, 188.0], [205.8, 187.56]]}, {\"category\": \"car\", \"corners_3d\": [[267.85, 333.83], [111.62, 336.22], [269.49, 278.49], [369.53, 277.49], [267.85, 189.91], [111.62, 190.16], [269.49, 184.05], [369.53, 183.94]]}, {\"category\": \"truck\", \"corners_3d\": [[659.81, 207.58], [718.13, 209.02], [604.17, 223.9], [529.51, 221.07], [659.81, 133.03], [718.13, 131.37], [604.17, 114.31], [529.51, 117.56]]}]\n```" - } - ] - }, - { - "id": 3541, - "data_source": "Mono3DRefer", - "scene_name": "006347", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1056.88, 376.31], [831.61, 363.88], [807.28, 288.06], [938.79, 292.47], [1056.88, 180.3], [831.61, 179.85], [807.28, 177.07], [938.79, 177.23]]}]\n```" - } - ] - }, - { - "id": 3542, - "data_source": "Mono3DRefer", - "scene_name": "006348", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[449.52, 260.87], [360.31, 261.1], [416.87, 239.79], [484.47, 239.66], [449.52, 178.85], [360.31, 178.86], [416.87, 177.41], [484.47, 177.4]]}]\n```" - } - ] - }, - { - "id": 3543, - "data_source": "Mono3DRefer", - "scene_name": "006350", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[583.08, 198.76], [608.83, 198.77], [607.98, 201.34], [579.69, 201.32], [583.08, 175.01], [608.83, 175.01], [607.98, 175.22], [579.69, 175.22]]}, {\"category\": \"car\", \"corners_3d\": [[1254.16, 396.06], [1033.9, 394.77], [852.93, 296.93], [975.19, 297.34], [1254.16, 190.4], [1033.9, 190.29], [852.93, 182.61], [975.19, 182.64]]}, {\"category\": \"car\", \"corners_3d\": [[906.46, 237.31], [844.86, 237.31], [804.04, 226.16], [854.98, 226.16], [906.46, 178.14], [844.86, 178.14], [804.04, 177.23], [854.98, 177.23]]}, {\"category\": \"car\", \"corners_3d\": [[669.12, 206.2], [663.93, 204.53], [754.62, 204.29], [764.49, 205.92], [669.12, 171.47], [663.93, 171.53], [754.62, 171.55], [764.49, 171.48]]}]\n```" - } - ] - }, - { - "id": 3544, - "data_source": "Mono3DRefer", - "scene_name": "006374", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[608.78, 227.88], [657.02, 227.47], [685.09, 239.59], [626.31, 240.19], [608.78, 183.2], [657.02, 183.13], [685.09, 185.4], [626.31, 185.52]]}, {\"category\": \"car\", \"corners_3d\": [[707.91, 281.87], [820.73, 280.15], [965.85, 335.17], [797.31, 339.14], [707.91, 169.31], [820.73, 169.37], [965.85, 167.58], [797.31, 167.45]]}]\n```" - } - ] - }, - { - "id": 3545, - "data_source": "Mono3DRefer", - "scene_name": "006375", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[591.11, 203.9], [623.74, 203.9], [625.19, 207.23], [589.07, 207.23], [591.11, 172.45], [623.74, 172.45], [625.19, 172.41], [589.07, 172.41]]}]\n```" - } - ] - }, - { - "id": 3546, - "data_source": "Mono3DRefer", - "scene_name": "006388", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[618.52, 223.05], [660.48, 222.85], [675.5, 230.29], [627.34, 230.55], [618.52, 176.43], [660.48, 176.42], [675.5, 176.95], [627.34, 176.97]]}, {\"category\": \"car\", \"corners_3d\": [[821.56, 356.97], [990.14, 352.56], [1784.04, 667.92], [1354.81, 702.9], [821.56, 172.82], [990.14, 172.82], [1784.04, 172.75], [1354.81, 172.74]]}, {\"category\": \"car\", \"corners_3d\": [[406.87, 350.49], [249.82, 353.55], [353.4, 291.43], [455.91, 290.1], [406.87, 210.68], [249.82, 211.33], [353.4, 198.11], [455.91, 197.82]]}, {\"category\": \"car\", \"corners_3d\": [[482.4, 222.74], [522.34, 222.41], [522.21, 230.97], [475.31, 231.42], [482.4, 184.36], [522.34, 184.28], [522.21, 186.26], [475.31, 186.36]]}]\n```" - } - ] - }, - { - "id": 3547, - "data_source": "Mono3DRefer", - "scene_name": "006392", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.96, 210.41], [574.3, 210.45], [576.46, 206.44], [603.85, 206.41], [604.96, 183.25], [574.3, 183.26], [576.46, 182.15], [603.85, 182.14]]}]\n```" - } - ] - }, - { - "id": 3548, - "data_source": "Mono3DRefer", - "scene_name": "006398", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[498.82, 265.36], [430.93, 265.63], [457.74, 249.72], [513.96, 249.54], [498.82, 199.27], [430.93, 199.35], [457.74, 194.81], [513.96, 194.75]]}, {\"category\": \"car\", \"corners_3d\": [[550.67, 224.14], [515.69, 223.55], [542.54, 218.36], [574.18, 218.84], [550.67, 187.09], [515.69, 186.93], [542.54, 185.49], [574.18, 185.62]]}, {\"category\": \"car\", \"corners_3d\": [[268.27, 420.59], [69.24, 421.32], [285.77, 320.29], [403.64, 320.03], [268.27, 201.2], [69.24, 201.28], [285.77, 189.72], [403.64, 189.69]]}]\n```" - } - ] - }, - { - "id": 3549, - "data_source": "Mono3DRefer", - "scene_name": "006404", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-580.67, 1431.01], [-1723.06, 1412.39], [123.51, 433.93], [367.07, 434.74], [-580.67, 231.46], [-1723.06, 230.6], [123.51, 185.02], [367.07, 185.05]]}, {\"category\": \"car\", \"corners_3d\": [[755.77, 334.95], [901.17, 334.32], [1146.14, 463.82], [885.05, 465.87], [755.77, 193.89], [901.17, 193.81], [1146.14, 210.61], [885.05, 210.88]]}, {\"category\": \"car\", \"corners_3d\": [[362.09, 307.02], [474.04, 306.22], [414.99, 379.67], [240.25, 381.61], [362.09, 192.85], [474.04, 192.73], [414.99, 203.67], [240.25, 203.96]]}, {\"category\": \"car\", \"corners_3d\": [[492.02, 288.01], [388.54, 288.19], [436.68, 262.15], [516.77, 262.04], [492.02, 196.88], [388.54, 196.92], [436.68, 191.48], [516.77, 191.46]]}, {\"category\": \"car\", \"corners_3d\": [[697.71, 261.88], [773.8, 261.03], [840.19, 286.62], [742.58, 288.04], [697.71, 183.9], [773.8, 183.8], [840.19, 186.97], [742.58, 187.15]]}, {\"category\": \"car\", \"corners_3d\": [[663.81, 238.66], [722.13, 238.56], [752.12, 254.15], [680.0, 254.3], [663.81, 184.54], [722.13, 184.52], [752.12, 187.29], [680.0, 187.32]]}, {\"category\": \"car\", \"corners_3d\": [[553.43, 228.23], [511.6, 228.11], [526.77, 221.47], [563.6, 221.57], [553.43, 185.88], [511.6, 185.85], [526.77, 184.29], [563.6, 184.31]]}, {\"category\": \"car\", \"corners_3d\": [[579.34, 218.89], [541.81, 218.82], [552.19, 213.45], [585.35, 213.5], [579.34, 175.57], [541.81, 175.56], [552.19, 175.25], [585.35, 175.25]]}, {\"category\": \"car\", \"corners_3d\": [[595.13, 212.4], [563.06, 212.28], [572.83, 208.29], [601.68, 208.38], [595.13, 177.43], [563.06, 177.42], [572.83, 176.96], [601.68, 176.97]]}]\n```" - } - ] - }, - { - "id": 3550, - "data_source": "Mono3DRefer", - "scene_name": "006406", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[453.77, 264.3], [359.7, 264.07], [413.8, 245.36], [488.64, 245.51], [453.77, 187.45], [359.7, 187.41], [413.8, 184.42], [488.64, 184.45]]}]\n```" - } - ] - }, - { - "id": 3551, - "data_source": "Mono3DRefer", - "scene_name": "006406", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[491.21, 205.97], [479.05, 205.9], [487.64, 204.63], [499.35, 204.7], [491.21, 176.93], [479.05, 176.92], [487.64, 176.76], [499.35, 176.77]]}]\n```" - } - ] - }, - { - "id": 3552, - "data_source": "Mono3DRefer", - "scene_name": "006406", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[491.21, 205.97], [479.05, 205.9], [487.64, 204.63], [499.35, 204.7], [491.21, 176.93], [479.05, 176.92], [487.64, 176.76], [499.35, 176.77]]}, {\"category\": \"car\", \"corners_3d\": [[453.77, 264.3], [359.7, 264.07], [413.8, 245.36], [488.64, 245.51], [453.77, 187.45], [359.7, 187.41], [413.8, 184.42], [488.64, 184.45]]}]\n```" - } - ] - }, - { - "id": 3553, - "data_source": "Mono3DRefer", - "scene_name": "006455", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[454.31, 231.85], [390.32, 231.63], [428.29, 223.88], [482.48, 224.04], [454.31, 180.0], [390.32, 180.0], [428.29, 180.24], [482.48, 180.23]]}]\n```" - } - ] - }, - { - "id": 3554, - "data_source": "Mono3DRefer", - "scene_name": "006456", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[509.51, 247.47], [438.48, 247.47], [469.76, 233.78], [527.76, 233.78], [509.51, 189.29], [438.48, 189.29], [469.76, 186.28], [527.76, 186.27]]}]\n```" - } - ] - }, - { - "id": 3555, - "data_source": "Mono3DRefer", - "scene_name": "006458", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[524.17, 299.13], [518.79, 310.0], [422.61, 310.52], [436.12, 299.57], [524.17, 161.11], [518.79, 159.33], [422.61, 159.24], [436.12, 161.03]]}]\n```" - } - ] - }, - { - "id": 3556, - "data_source": "Mono3DRefer", - "scene_name": "006484", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[585.58, 199.44], [609.85, 199.42], [610.96, 201.52], [584.78, 201.54], [585.58, 175.46], [609.85, 175.46], [610.96, 175.67], [584.78, 175.67]]}]\n```" - } - ] - }, - { - "id": 3557, - "data_source": "Mono3DRefer", - "scene_name": "006484", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[170.11, 230.07], [99.07, 230.45], [353.0, 198.61], [384.13, 198.53], [170.11, 139.57], [99.07, 139.35], [353.0, 157.87], [384.13, 157.92]]}]\n```" - } - ] - }, - { - "id": 3558, - "data_source": "Mono3DRefer", - "scene_name": "006484", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[689.23, 211.88], [680.0, 211.92], [673.34, 209.78], [682.08, 209.75], [689.23, 165.29], [680.0, 165.28], [673.34, 165.7], [682.08, 165.7]]}]\n```" - } - ] - }, - { - "id": 3559, - "data_source": "Mono3DRefer", - "scene_name": "006484", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[585.58, 199.44], [609.85, 199.42], [610.96, 201.52], [584.78, 201.54], [585.58, 175.46], [609.85, 175.46], [610.96, 175.67], [584.78, 175.67]]}, {\"category\": \"tram\", \"corners_3d\": [[170.11, 230.07], [99.07, 230.45], [353.0, 198.61], [384.13, 198.53], [170.11, 139.57], [99.07, 139.35], [353.0, 157.87], [384.13, 157.92]]}]\n```" - } - ] - }, - { - "id": 3560, - "data_source": "Mono3DRefer", - "scene_name": "006484", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[585.58, 199.44], [609.85, 199.42], [610.96, 201.52], [584.78, 201.54], [585.58, 175.46], [609.85, 175.46], [610.96, 175.67], [584.78, 175.67]]}, {\"category\": \"cyclist\", \"corners_3d\": [[689.23, 211.88], [680.0, 211.92], [673.34, 209.78], [682.08, 209.75], [689.23, 165.29], [680.0, 165.28], [673.34, 165.7], [682.08, 165.7]]}]\n```" - } - ] - }, - { - "id": 3561, - "data_source": "Mono3DRefer", - "scene_name": "006484", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[170.11, 230.07], [99.07, 230.45], [353.0, 198.61], [384.13, 198.53], [170.11, 139.57], [99.07, 139.35], [353.0, 157.87], [384.13, 157.92]]}, {\"category\": \"cyclist\", \"corners_3d\": [[689.23, 211.88], [680.0, 211.92], [673.34, 209.78], [682.08, 209.75], [689.23, 165.29], [680.0, 165.28], [673.34, 165.7], [682.08, 165.7]]}]\n```" - } - ] - }, - { - "id": 3562, - "data_source": "Mono3DRefer", - "scene_name": "006484", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, tram, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[585.58, 199.44], [609.85, 199.42], [610.96, 201.52], [584.78, 201.54], [585.58, 175.46], [609.85, 175.46], [610.96, 175.67], [584.78, 175.67]]}, {\"category\": \"tram\", \"corners_3d\": [[170.11, 230.07], [99.07, 230.45], [353.0, 198.61], [384.13, 198.53], [170.11, 139.57], [99.07, 139.35], [353.0, 157.87], [384.13, 157.92]]}, {\"category\": \"cyclist\", \"corners_3d\": [[689.23, 211.88], [680.0, 211.92], [673.34, 209.78], [682.08, 209.75], [689.23, 165.29], [680.0, 165.28], [673.34, 165.7], [682.08, 165.7]]}]\n```" - } - ] - }, - { - "id": 3563, - "data_source": "Mono3DRefer", - "scene_name": "006485", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[500.7, 254.96], [428.49, 254.89], [462.71, 239.93], [521.77, 239.98], [500.7, 187.69], [428.49, 187.67], [462.71, 184.97], [521.77, 184.98]]}]\n```" - } - ] - }, - { - "id": 3564, - "data_source": "Mono3DRefer", - "scene_name": "006487", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[481.95, 251.57], [532.86, 257.78], [359.46, 277.41], [313.16, 268.16], [481.95, 184.29], [532.86, 185.19], [359.46, 188.04], [313.16, 186.7]]}, {\"category\": \"car\", \"corners_3d\": [[406.07, 244.58], [457.82, 250.13], [259.52, 270.61], [212.94, 261.9], [406.07, 171.18], [457.82, 171.05], [259.52, 170.57], [212.94, 170.78]]}, {\"category\": \"car\", \"corners_3d\": [[72.78, 224.22], [20.64, 226.49], [-1.08, 221.13], [46.89, 219.29], [72.78, 189.2], [20.64, 189.92], [-1.08, 188.22], [46.89, 187.63]]}]\n```" - } - ] - }, - { - "id": 3565, - "data_source": "Mono3DRefer", - "scene_name": "006488", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[703.38, 359.85], [763.15, 348.07], [858.99, 371.73], [797.86, 387.25], [703.38, 157.16], [763.15, 158.76], [858.99, 155.53], [797.86, 153.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[682.01, 323.27], [729.27, 317.72], [788.06, 332.83], [738.0, 339.72], [682.01, 165.4], [729.27, 166.03], [788.06, 164.31], [738.0, 163.53]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[522.28, 313.47], [574.15, 308.0], [616.33, 320.91], [560.93, 327.58], [522.28, 163.89], [574.15, 164.62], [616.33, 162.89], [560.93, 162.0]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[403.61, 290.71], [457.57, 285.74], [487.76, 294.02], [430.85, 299.84], [403.61, 166.1], [457.57, 166.8], [487.76, 165.63], [430.85, 164.81]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[621.29, 285.21], [659.69, 279.8], [707.77, 286.18], [669.56, 292.33], [621.29, 160.65], [659.69, 161.74], [707.77, 160.45], [669.56, 159.22]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[488.62, 270.55], [515.83, 266.48], [561.98, 270.46], [535.76, 274.94], [488.62, 159.44], [515.83, 160.45], [561.98, 159.47], [535.76, 158.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[541.5, 295.93], [583.74, 290.53], [627.56, 298.5], [584.4, 304.73], [541.5, 170.64], [583.74, 171.15], [627.56, 170.39], [584.4, 169.8]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[537.05, 255.76], [563.32, 252.32], [603.56, 255.21], [578.2, 258.94], [537.05, 155.74], [563.32, 156.94], [603.56, 155.94], [578.2, 154.64]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[577.56, 278.25], [614.45, 274.49], [653.42, 280.84], [615.59, 285.14], [577.56, 160.12], [614.45, 160.95], [653.42, 159.55], [615.59, 158.59]]}]\n```" - } - ] - }, - { - "id": 3566, - "data_source": "Mono3DRefer", - "scene_name": "006488", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[508.16, 238.76], [435.81, 238.7], [470.04, 227.27], [527.94, 227.3], [508.16, 170.59], [435.81, 170.6], [470.04, 172.79], [527.94, 172.78]]}]\n```" - } - ] - }, - { - "id": 3567, - "data_source": "Mono3DRefer", - "scene_name": "006488", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[703.38, 359.85], [763.15, 348.07], [858.99, 371.73], [797.86, 387.25], [703.38, 157.16], [763.15, 158.76], [858.99, 155.53], [797.86, 153.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[682.01, 323.27], [729.27, 317.72], [788.06, 332.83], [738.0, 339.72], [682.01, 165.4], [729.27, 166.03], [788.06, 164.31], [738.0, 163.53]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[522.28, 313.47], [574.15, 308.0], [616.33, 320.91], [560.93, 327.58], [522.28, 163.89], [574.15, 164.62], [616.33, 162.89], [560.93, 162.0]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[403.61, 290.71], [457.57, 285.74], [487.76, 294.02], [430.85, 299.84], [403.61, 166.1], [457.57, 166.8], [487.76, 165.63], [430.85, 164.81]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[621.29, 285.21], [659.69, 279.8], [707.77, 286.18], [669.56, 292.33], [621.29, 160.65], [659.69, 161.74], [707.77, 160.45], [669.56, 159.22]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[488.62, 270.55], [515.83, 266.48], [561.98, 270.46], [535.76, 274.94], [488.62, 159.44], [515.83, 160.45], [561.98, 159.47], [535.76, 158.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[541.5, 295.93], [583.74, 290.53], [627.56, 298.5], [584.4, 304.73], [541.5, 170.64], [583.74, 171.15], [627.56, 170.39], [584.4, 169.8]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[537.05, 255.76], [563.32, 252.32], [603.56, 255.21], [578.2, 258.94], [537.05, 155.74], [563.32, 156.94], [603.56, 155.94], [578.2, 154.64]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[577.56, 278.25], [614.45, 274.49], [653.42, 280.84], [615.59, 285.14], [577.56, 160.12], [614.45, 160.95], [653.42, 159.55], [615.59, 158.59]]}, {\"category\": \"car\", \"corners_3d\": [[508.16, 238.76], [435.81, 238.7], [470.04, 227.27], [527.94, 227.3], [508.16, 170.59], [435.81, 170.6], [470.04, 172.79], [527.94, 172.78]]}]\n```" - } - ] - }, - { - "id": 3568, - "data_source": "Mono3DRefer", - "scene_name": "006510", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[563.6, 195.16], [538.64, 195.11], [547.96, 194.21], [571.3, 194.25], [563.6, 174.94], [538.64, 174.96], [547.96, 175.4], [571.3, 175.38]]}]\n```" - } - ] - }, - { - "id": 3569, - "data_source": "Mono3DRefer", - "scene_name": "006514", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[971.75, 237.56], [973.49, 232.9], [1154.36, 239.24], [1168.09, 244.99], [971.75, 179.27], [973.49, 178.81], [1154.36, 179.44], [1168.09, 180.01]]}, {\"category\": \"car\", \"corners_3d\": [[767.52, 224.27], [790.98, 226.13], [729.1, 229.9], [706.33, 227.78], [767.52, 197.59], [790.98, 198.48], [729.1, 200.29], [706.33, 199.27]]}]\n```" - } - ] - }, - { - "id": 3570, - "data_source": "Mono3DRefer", - "scene_name": "006514", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[625.69, 220.57], [661.71, 220.45], [673.8, 227.01], [632.82, 227.16], [625.69, 179.76], [661.71, 179.74], [673.8, 180.69], [632.82, 180.71]]}, {\"category\": \"van\", \"corners_3d\": [[546.03, 200.66], [566.32, 200.56], [569.79, 202.46], [548.11, 202.57], [546.03, 175.6], [566.32, 175.59], [569.79, 175.78], [548.11, 175.79]]}]\n```" - } - ] - }, - { - "id": 3571, - "data_source": "Mono3DRefer", - "scene_name": "006514", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[971.75, 237.56], [973.49, 232.9], [1154.36, 239.24], [1168.09, 244.99], [971.75, 179.27], [973.49, 178.81], [1154.36, 179.44], [1168.09, 180.01]]}, {\"category\": \"van\", \"corners_3d\": [[625.69, 220.57], [661.71, 220.45], [673.8, 227.01], [632.82, 227.16], [625.69, 179.76], [661.71, 179.74], [673.8, 180.69], [632.82, 180.71]]}, {\"category\": \"car\", \"corners_3d\": [[767.52, 224.27], [790.98, 226.13], [729.1, 229.9], [706.33, 227.78], [767.52, 197.59], [790.98, 198.48], [729.1, 200.29], [706.33, 199.27]]}, {\"category\": \"van\", \"corners_3d\": [[546.03, 200.66], [566.32, 200.56], [569.79, 202.46], [548.11, 202.57], [546.03, 175.6], [566.32, 175.59], [569.79, 175.78], [548.11, 175.79]]}]\n```" - } - ] - }, - { - "id": 3572, - "data_source": "Mono3DRefer", - "scene_name": "006518", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[560.78, 203.72], [526.71, 203.77], [532.92, 200.6], [563.5, 200.56], [560.78, 167.94], [526.71, 167.94], [532.92, 168.44], [563.5, 168.45]]}, {\"category\": \"car\", \"corners_3d\": [[574.12, 190.61], [556.95, 190.61], [560.68, 189.61], [576.89, 189.62], [574.12, 171.5], [556.95, 171.5], [560.68, 171.58], [576.89, 171.58]]}]\n```" - } - ] - }, - { - "id": 3573, - "data_source": "Mono3DRefer", - "scene_name": "006520", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[821.26, 239.55], [845.01, 245.28], [708.35, 246.55], [695.19, 240.63], [821.26, 183.05], [845.01, 183.92], [708.35, 184.12], [695.19, 183.21]]}, {\"category\": \"car\", \"corners_3d\": [[1121.02, 309.92], [1254.99, 342.18], [888.1, 349.61], [821.76, 314.75], [1121.02, 193.43], [1254.99, 198.27], [888.1, 199.39], [821.76, 194.16]]}, {\"category\": \"car\", \"corners_3d\": [[200.98, 281.11], [139.48, 301.81], [-216.88, 308.93], [-92.73, 286.08], [200.98, 163.75], [139.48, 162.01], [-216.88, 161.41], [-92.73, 163.33]]}, {\"category\": \"car\", \"corners_3d\": [[-35.98, 280.56], [41.14, 266.74], [281.44, 264.9], [240.64, 278.15], [-35.98, 185.47], [41.14, 183.85], [281.44, 183.63], [240.64, 185.18]]}, {\"category\": \"car\", \"corners_3d\": [[339.04, 247.16], [306.25, 256.41], [87.93, 256.66], [145.04, 247.36], [339.04, 176.09], [306.25, 176.49], [87.93, 176.5], [145.04, 176.1]]}, {\"category\": \"car\", \"corners_3d\": [[227.17, 237.57], [252.11, 232.58], [397.64, 231.63], [385.08, 236.46], [227.17, 179.0], [252.11, 178.53], [397.64, 178.44], [385.08, 178.9]]}, {\"category\": \"car\", \"corners_3d\": [[781.33, 235.01], [802.14, 239.87], [685.02, 241.35], [672.46, 236.28], [781.33, 187.09], [802.14, 188.2], [685.02, 188.54], [672.46, 187.38]]}, {\"category\": \"car\", \"corners_3d\": [[459.08, 199.77], [478.95, 199.43], [493.89, 201.12], [472.94, 201.51], [459.08, 179.37], [478.95, 179.28], [493.89, 179.69], [472.94, 179.78]]}]\n```" - } - ] - }, - { - "id": 3574, - "data_source": "Mono3DRefer", - "scene_name": "006521", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[679.12, 246.84], [740.62, 246.65], [787.59, 269.44], [707.19, 269.75], [679.12, 191.63], [740.62, 191.59], [787.59, 197.37], [707.19, 197.45]]}]\n```" - } - ] - }, - { - "id": 3575, - "data_source": "Mono3DRefer", - "scene_name": "006525", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[430.38, 268.47], [517.83, 268.48], [480.14, 307.53], [357.0, 307.51], [430.38, 200.45], [517.83, 200.46], [480.14, 211.72], [357.0, 211.72]]}, {\"category\": \"car\", \"corners_3d\": [[660.68, 220.93], [700.48, 220.88], [719.5, 229.6], [672.49, 229.68], [660.68, 182.82], [700.48, 182.81], [719.5, 184.62], [672.49, 184.63]]}, {\"category\": \"car\", \"corners_3d\": [[500.99, 234.11], [548.78, 234.16], [533.86, 247.49], [475.71, 247.43], [500.99, 189.72], [548.78, 189.74], [533.86, 193.41], [475.71, 193.39]]}, {\"category\": \"car\", \"corners_3d\": [[704.01, 252.13], [777.58, 252.06], [835.73, 278.42], [737.71, 278.55], [704.01, 178.4], [777.58, 178.4], [835.73, 180.24], [737.71, 180.25]]}, {\"category\": \"car\", \"corners_3d\": [[641.92, 205.95], [668.49, 206.05], [668.53, 208.87], [639.7, 208.76], [641.92, 183.67], [668.49, 183.7], [668.53, 184.62], [639.7, 184.58]]}]\n```" - } - ] - }, - { - "id": 3576, - "data_source": "Mono3DRefer", - "scene_name": "006525", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[737.07, 274.95], [856.79, 274.8], [1196.58, 411.24], [917.26, 412.09], [737.07, 109.25], [856.79, 109.34], [1196.58, 24.34], [917.26, 23.81]]}]\n```" - } - ] - }, - { - "id": 3577, - "data_source": "Mono3DRefer", - "scene_name": "006525", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1201.49, 438.29], [1124.8, 405.56], [1188.16, 404.69], [1273.4, 437.15], [1201.49, 122.9], [1124.8, 129.06], [1188.16, 129.22], [1273.4, 123.11]]}]\n```" - } - ] - }, - { - "id": 3578, - "data_source": "Mono3DRefer", - "scene_name": "006525", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[430.38, 268.47], [517.83, 268.48], [480.14, 307.53], [357.0, 307.51], [430.38, 200.45], [517.83, 200.46], [480.14, 211.72], [357.0, 211.72]]}, {\"category\": \"van\", \"corners_3d\": [[737.07, 274.95], [856.79, 274.8], [1196.58, 411.24], [917.26, 412.09], [737.07, 109.25], [856.79, 109.34], [1196.58, 24.34], [917.26, 23.81]]}, {\"category\": \"car\", \"corners_3d\": [[660.68, 220.93], [700.48, 220.88], [719.5, 229.6], [672.49, 229.68], [660.68, 182.82], [700.48, 182.81], [719.5, 184.62], [672.49, 184.63]]}, {\"category\": \"car\", \"corners_3d\": [[500.99, 234.11], [548.78, 234.16], [533.86, 247.49], [475.71, 247.43], [500.99, 189.72], [548.78, 189.74], [533.86, 193.41], [475.71, 193.39]]}, {\"category\": \"car\", \"corners_3d\": [[704.01, 252.13], [777.58, 252.06], [835.73, 278.42], [737.71, 278.55], [704.01, 178.4], [777.58, 178.4], [835.73, 180.24], [737.71, 180.25]]}, {\"category\": \"car\", \"corners_3d\": [[641.92, 205.95], [668.49, 206.05], [668.53, 208.87], [639.7, 208.76], [641.92, 183.67], [668.49, 183.7], [668.53, 184.62], [639.7, 184.58]]}]\n```" - } - ] - }, - { - "id": 3579, - "data_source": "Mono3DRefer", - "scene_name": "006525", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[430.38, 268.47], [517.83, 268.48], [480.14, 307.53], [357.0, 307.51], [430.38, 200.45], [517.83, 200.46], [480.14, 211.72], [357.0, 211.72]]}, {\"category\": \"car\", \"corners_3d\": [[660.68, 220.93], [700.48, 220.88], [719.5, 229.6], [672.49, 229.68], [660.68, 182.82], [700.48, 182.81], [719.5, 184.62], [672.49, 184.63]]}, {\"category\": \"car\", \"corners_3d\": [[500.99, 234.11], [548.78, 234.16], [533.86, 247.49], [475.71, 247.43], [500.99, 189.72], [548.78, 189.74], [533.86, 193.41], [475.71, 193.39]]}, {\"category\": \"car\", \"corners_3d\": [[704.01, 252.13], [777.58, 252.06], [835.73, 278.42], [737.71, 278.55], [704.01, 178.4], [777.58, 178.4], [835.73, 180.24], [737.71, 180.25]]}, {\"category\": \"car\", \"corners_3d\": [[641.92, 205.95], [668.49, 206.05], [668.53, 208.87], [639.7, 208.76], [641.92, 183.67], [668.49, 183.7], [668.53, 184.62], [639.7, 184.58]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1201.49, 438.29], [1124.8, 405.56], [1188.16, 404.69], [1273.4, 437.15], [1201.49, 122.9], [1124.8, 129.06], [1188.16, 129.22], [1273.4, 123.11]]}]\n```" - } - ] - }, - { - "id": 3580, - "data_source": "Mono3DRefer", - "scene_name": "006525", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[737.07, 274.95], [856.79, 274.8], [1196.58, 411.24], [917.26, 412.09], [737.07, 109.25], [856.79, 109.34], [1196.58, 24.34], [917.26, 23.81]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1201.49, 438.29], [1124.8, 405.56], [1188.16, 404.69], [1273.4, 437.15], [1201.49, 122.9], [1124.8, 129.06], [1188.16, 129.22], [1273.4, 123.11]]}]\n```" - } - ] - }, - { - "id": 3581, - "data_source": "Mono3DRefer", - "scene_name": "006525", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[430.38, 268.47], [517.83, 268.48], [480.14, 307.53], [357.0, 307.51], [430.38, 200.45], [517.83, 200.46], [480.14, 211.72], [357.0, 211.72]]}, {\"category\": \"van\", \"corners_3d\": [[737.07, 274.95], [856.79, 274.8], [1196.58, 411.24], [917.26, 412.09], [737.07, 109.25], [856.79, 109.34], [1196.58, 24.34], [917.26, 23.81]]}, {\"category\": \"car\", \"corners_3d\": [[660.68, 220.93], [700.48, 220.88], [719.5, 229.6], [672.49, 229.68], [660.68, 182.82], [700.48, 182.81], [719.5, 184.62], [672.49, 184.63]]}, {\"category\": \"car\", \"corners_3d\": [[500.99, 234.11], [548.78, 234.16], [533.86, 247.49], [475.71, 247.43], [500.99, 189.72], [548.78, 189.74], [533.86, 193.41], [475.71, 193.39]]}, {\"category\": \"car\", \"corners_3d\": [[704.01, 252.13], [777.58, 252.06], [835.73, 278.42], [737.71, 278.55], [704.01, 178.4], [777.58, 178.4], [835.73, 180.24], [737.71, 180.25]]}, {\"category\": \"car\", \"corners_3d\": [[641.92, 205.95], [668.49, 206.05], [668.53, 208.87], [639.7, 208.76], [641.92, 183.67], [668.49, 183.7], [668.53, 184.62], [639.7, 184.58]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1201.49, 438.29], [1124.8, 405.56], [1188.16, 404.69], [1273.4, 437.15], [1201.49, 122.9], [1124.8, 129.06], [1188.16, 129.22], [1273.4, 123.11]]}]\n```" - } - ] - }, - { - "id": 3582, - "data_source": "Mono3DRefer", - "scene_name": "006532", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[70.72, 274.38], [136.98, 261.92], [373.86, 261.97], [340.72, 274.44], [70.72, 182.86], [136.98, 181.63], [373.86, 181.64], [340.72, 182.87]]}, {\"category\": \"car\", \"corners_3d\": [[252.17, 244.82], [279.2, 237.79], [459.01, 236.07], [451.68, 242.71], [252.17, 164.84], [279.2, 165.62], [459.01, 165.82], [451.68, 165.07]]}, {\"category\": \"car\", \"corners_3d\": [[413.78, 210.48], [423.53, 208.74], [494.62, 208.82], [488.3, 210.56], [413.78, 174.66], [423.53, 174.58], [494.62, 174.58], [488.3, 174.66]]}, {\"category\": \"car\", \"corners_3d\": [[522.23, 207.45], [517.52, 209.17], [432.29, 209.12], [441.02, 207.41], [522.23, 173.85], [517.52, 173.9], [432.29, 173.9], [441.02, 173.85]]}, {\"category\": \"car\", \"corners_3d\": [[1190.85, 394.18], [1047.54, 341.04], [1481.69, 338.83], [1757.35, 390.36], [1190.85, 159.47], [1047.54, 162.68], [1481.69, 162.82], [1757.35, 159.7]]}]\n```" - } - ] - }, - { - "id": 3583, - "data_source": "Mono3DRefer", - "scene_name": "006532", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[683.18, 230.4], [755.98, 230.41], [822.9, 256.82], [716.69, 256.81], [683.18, 143.87], [755.98, 143.87], [822.9, 130.57], [716.69, 130.58]]}, {\"category\": \"van\", \"corners_3d\": [[644.99, 203.58], [683.61, 203.55], [702.75, 210.23], [655.73, 210.28], [644.99, 153.59], [683.61, 153.61], [702.75, 149.42], [655.73, 149.39]]}, {\"category\": \"van\", \"corners_3d\": [[562.19, 195.08], [560.3, 195.85], [483.18, 195.83], [487.65, 195.06], [562.19, 158.33], [560.3, 157.82], [483.18, 157.84], [487.65, 158.34]]}]\n```" - } - ] - }, - { - "id": 3584, - "data_source": "Mono3DRefer", - "scene_name": "006532", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[683.18, 230.4], [755.98, 230.41], [822.9, 256.82], [716.69, 256.81], [683.18, 143.87], [755.98, 143.87], [822.9, 130.57], [716.69, 130.58]]}, {\"category\": \"car\", \"corners_3d\": [[70.72, 274.38], [136.98, 261.92], [373.86, 261.97], [340.72, 274.44], [70.72, 182.86], [136.98, 181.63], [373.86, 181.64], [340.72, 182.87]]}, {\"category\": \"car\", \"corners_3d\": [[252.17, 244.82], [279.2, 237.79], [459.01, 236.07], [451.68, 242.71], [252.17, 164.84], [279.2, 165.62], [459.01, 165.82], [451.68, 165.07]]}, {\"category\": \"van\", \"corners_3d\": [[644.99, 203.58], [683.61, 203.55], [702.75, 210.23], [655.73, 210.28], [644.99, 153.59], [683.61, 153.61], [702.75, 149.42], [655.73, 149.39]]}, {\"category\": \"car\", \"corners_3d\": [[413.78, 210.48], [423.53, 208.74], [494.62, 208.82], [488.3, 210.56], [413.78, 174.66], [423.53, 174.58], [494.62, 174.58], [488.3, 174.66]]}, {\"category\": \"car\", \"corners_3d\": [[522.23, 207.45], [517.52, 209.17], [432.29, 209.12], [441.02, 207.41], [522.23, 173.85], [517.52, 173.9], [432.29, 173.9], [441.02, 173.85]]}, {\"category\": \"van\", \"corners_3d\": [[562.19, 195.08], [560.3, 195.85], [483.18, 195.83], [487.65, 195.06], [562.19, 158.33], [560.3, 157.82], [483.18, 157.84], [487.65, 158.34]]}, {\"category\": \"car\", \"corners_3d\": [[1190.85, 394.18], [1047.54, 341.04], [1481.69, 338.83], [1757.35, 390.36], [1190.85, 159.47], [1047.54, 162.68], [1481.69, 162.82], [1757.35, 159.7]]}]\n```" - } - ] - }, - { - "id": 3585, - "data_source": "Mono3DRefer", - "scene_name": "006534", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-1961.59, 1303.03], [-3038.5, 1304.37], [-68.39, 382.99], [131.14, 382.94], [-1961.59, 365.31], [-3038.5, 365.53], [-68.39, 208.64], [131.14, 208.63]]}, {\"category\": \"car\", \"corners_3d\": [[485.82, 238.01], [424.29, 237.39], [472.24, 225.68], [522.9, 226.1], [485.82, 185.73], [424.29, 185.6], [472.24, 183.29], [522.9, 183.37]]}, {\"category\": \"car\", \"corners_3d\": [[702.35, 219.17], [754.61, 219.6], [764.06, 228.42], [701.96, 227.81], [702.35, 177.03], [754.61, 177.07], [764.06, 177.86], [701.96, 177.81]]}, {\"category\": \"car\", \"corners_3d\": [[669.85, 191.77], [649.47, 191.57], [664.23, 190.38], [683.44, 190.55], [669.85, 170.56], [649.47, 170.58], [664.23, 170.73], [683.44, 170.71]]}]\n```" - } - ] - }, - { - "id": 3586, - "data_source": "Mono3DRefer", - "scene_name": "006534", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[851.75, 252.12], [889.51, 252.12], [920.87, 261.02], [878.86, 261.01], [851.75, 148.56], [889.51, 148.56], [920.87, 145.84], [878.86, 145.84]]}]\n```" - } - ] - }, - { - "id": 3587, - "data_source": "Mono3DRefer", - "scene_name": "006534", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-1961.59, 1303.03], [-3038.5, 1304.37], [-68.39, 382.99], [131.14, 382.94], [-1961.59, 365.31], [-3038.5, 365.53], [-68.39, 208.64], [131.14, 208.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[851.75, 252.12], [889.51, 252.12], [920.87, 261.02], [878.86, 261.01], [851.75, 148.56], [889.51, 148.56], [920.87, 145.84], [878.86, 145.84]]}, {\"category\": \"car\", \"corners_3d\": [[485.82, 238.01], [424.29, 237.39], [472.24, 225.68], [522.9, 226.1], [485.82, 185.73], [424.29, 185.6], [472.24, 183.29], [522.9, 183.37]]}, {\"category\": \"car\", \"corners_3d\": [[702.35, 219.17], [754.61, 219.6], [764.06, 228.42], [701.96, 227.81], [702.35, 177.03], [754.61, 177.07], [764.06, 177.86], [701.96, 177.81]]}, {\"category\": \"car\", \"corners_3d\": [[669.85, 191.77], [649.47, 191.57], [664.23, 190.38], [683.44, 190.55], [669.85, 170.56], [649.47, 170.58], [664.23, 170.73], [683.44, 170.71]]}]\n```" - } - ] - }, - { - "id": 3588, - "data_source": "Mono3DRefer", - "scene_name": "006535", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[35.99, 289.07], [109.87, 274.5], [301.96, 275.09], [255.32, 289.85], [35.99, 182.91], [109.87, 181.65], [301.96, 181.7], [255.32, 182.98]]}, {\"category\": \"car\", \"corners_3d\": [[398.37, 256.61], [370.24, 267.39], [150.48, 267.1], [203.54, 256.38], [398.37, 175.73], [370.24, 176.1], [150.48, 176.09], [203.54, 175.72]]}, {\"category\": \"car\", \"corners_3d\": [[213.06, 252.48], [248.54, 244.61], [431.93, 243.7], [416.76, 251.35], [213.06, 181.22], [248.54, 180.39], [431.93, 180.3], [416.76, 181.1]]}, {\"category\": \"car\", \"corners_3d\": [[485.24, 231.84], [474.2, 236.79], [331.9, 236.67], [353.91, 231.73], [485.24, 183.51], [474.2, 184.41], [331.9, 184.38], [353.91, 183.49]]}, {\"category\": \"car\", \"corners_3d\": [[523.96, 216.6], [521.29, 218.97], [416.5, 219.3], [424.6, 216.9], [523.96, 183.11], [521.29, 183.67], [416.5, 183.75], [424.6, 183.19]]}, {\"category\": \"car\", \"corners_3d\": [[650.66, 214.38], [686.99, 214.36], [699.58, 220.58], [657.81, 220.6], [650.66, 182.72], [686.99, 182.72], [699.58, 184.2], [657.81, 184.2]]}]\n```" - } - ] - }, - { - "id": 3589, - "data_source": "Mono3DRefer", - "scene_name": "006535", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[696.97, 247.28], [786.76, 247.28], [912.17, 300.13], [758.61, 300.11], [696.97, 131.33], [786.76, 131.33], [912.17, 101.85], [758.61, 101.86]]}, {\"category\": \"van\", \"corners_3d\": [[534.92, 214.34], [530.26, 216.7], [403.88, 216.62], [415.31, 214.27], [534.92, 155.3], [530.26, 154.3], [403.88, 154.34], [415.31, 155.33]]}]\n```" - } - ] - }, - { - "id": 3590, - "data_source": "Mono3DRefer", - "scene_name": "006535", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[696.97, 247.28], [786.76, 247.28], [912.17, 300.13], [758.61, 300.11], [696.97, 131.33], [786.76, 131.33], [912.17, 101.85], [758.61, 101.86]]}, {\"category\": \"car\", \"corners_3d\": [[35.99, 289.07], [109.87, 274.5], [301.96, 275.09], [255.32, 289.85], [35.99, 182.91], [109.87, 181.65], [301.96, 181.7], [255.32, 182.98]]}, {\"category\": \"car\", \"corners_3d\": [[398.37, 256.61], [370.24, 267.39], [150.48, 267.1], [203.54, 256.38], [398.37, 175.73], [370.24, 176.1], [150.48, 176.09], [203.54, 175.72]]}, {\"category\": \"car\", \"corners_3d\": [[213.06, 252.48], [248.54, 244.61], [431.93, 243.7], [416.76, 251.35], [213.06, 181.22], [248.54, 180.39], [431.93, 180.3], [416.76, 181.1]]}, {\"category\": \"car\", \"corners_3d\": [[485.24, 231.84], [474.2, 236.79], [331.9, 236.67], [353.91, 231.73], [485.24, 183.51], [474.2, 184.41], [331.9, 184.38], [353.91, 183.49]]}, {\"category\": \"car\", \"corners_3d\": [[523.96, 216.6], [521.29, 218.97], [416.5, 219.3], [424.6, 216.9], [523.96, 183.11], [521.29, 183.67], [416.5, 183.75], [424.6, 183.19]]}, {\"category\": \"van\", \"corners_3d\": [[534.92, 214.34], [530.26, 216.7], [403.88, 216.62], [415.31, 214.27], [534.92, 155.3], [530.26, 154.3], [403.88, 154.34], [415.31, 155.33]]}, {\"category\": \"car\", \"corners_3d\": [[650.66, 214.38], [686.99, 214.36], [699.58, 220.58], [657.81, 220.6], [650.66, 182.72], [686.99, 182.72], [699.58, 184.2], [657.81, 184.2]]}]\n```" - } - ] - }, - { - "id": 3591, - "data_source": "Mono3DRefer", - "scene_name": "006550", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[301.7, 443.64], [62.65, 442.81], [305.23, 324.51], [439.76, 324.77], [301.7, 214.36], [62.65, 214.23], [305.23, 196.1], [439.76, 196.14]]}, {\"category\": \"car\", \"corners_3d\": [[566.17, 221.38], [530.83, 221.38], [539.28, 216.13], [570.8, 216.13], [566.17, 191.15], [530.83, 191.15], [539.28, 189.17], [570.8, 189.17]]}]\n```" - } - ] - }, - { - "id": 3592, - "data_source": "Mono3DRefer", - "scene_name": "006553", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1050.18, 237.76], [1120.43, 241.27], [1104.38, 252.49], [1024.25, 247.77], [1050.18, 176.53], [1120.43, 176.73], [1104.38, 177.36], [1024.25, 177.1]]}, {\"category\": \"car\", \"corners_3d\": [[980.01, 208.39], [943.66, 207.35], [965.36, 204.48], [999.18, 205.35], [980.01, 178.25], [943.66, 178.09], [965.36, 177.66], [999.18, 177.79]]}]\n```" - } - ] - }, - { - "id": 3593, - "data_source": "Mono3DRefer", - "scene_name": "006556", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[-212.24, 822.71], [-1258.4, 915.75], [186.8, 315.18], [371.22, 311.38], [-212.24, -412.65], [-1258.4, -496.48], [186.8, 44.62], [371.22, 48.04]]}]\n```" - } - ] - }, - { - "id": 3594, - "data_source": "Mono3DRefer", - "scene_name": "006556", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[611.14, 273.7], [636.89, 271.08], [653.99, 274.12], [627.9, 276.91], [611.14, 163.66], [636.89, 163.9], [653.99, 163.62], [627.9, 163.37]]}]\n```" - } - ] - }, - { - "id": 3595, - "data_source": "Mono3DRefer", - "scene_name": "006556", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[252.67, 235.66], [281.31, 231.18], [391.06, 231.74], [370.65, 236.32], [252.67, 174.76], [281.31, 174.62], [391.06, 174.64], [370.65, 174.78]]}, {\"category\": \"car\", \"corners_3d\": [[635.33, 259.5], [723.74, 259.51], [762.4, 288.99], [643.91, 288.97], [635.33, 175.14], [723.74, 175.14], [762.4, 175.92], [643.91, 175.92]]}]\n```" - } - ] - }, - { - "id": 3596, - "data_source": "Mono3DRefer", - "scene_name": "006556", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[611.14, 273.7], [636.89, 271.08], [653.99, 274.12], [627.9, 276.91], [611.14, 163.66], [636.89, 163.9], [653.99, 163.62], [627.9, 163.37]]}, {\"category\": \"truck\", \"corners_3d\": [[-212.24, 822.71], [-1258.4, 915.75], [186.8, 315.18], [371.22, 311.38], [-212.24, -412.65], [-1258.4, -496.48], [186.8, 44.62], [371.22, 48.04]]}]\n```" - } - ] - }, - { - "id": 3597, - "data_source": "Mono3DRefer", - "scene_name": "006556", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[252.67, 235.66], [281.31, 231.18], [391.06, 231.74], [370.65, 236.32], [252.67, 174.76], [281.31, 174.62], [391.06, 174.64], [370.65, 174.78]]}, {\"category\": \"truck\", \"corners_3d\": [[-212.24, 822.71], [-1258.4, 915.75], [186.8, 315.18], [371.22, 311.38], [-212.24, -412.65], [-1258.4, -496.48], [186.8, 44.62], [371.22, 48.04]]}, {\"category\": \"car\", \"corners_3d\": [[635.33, 259.5], [723.74, 259.51], [762.4, 288.99], [643.91, 288.97], [635.33, 175.14], [723.74, 175.14], [762.4, 175.92], [643.91, 175.92]]}]\n```" - } - ] - }, - { - "id": 3598, - "data_source": "Mono3DRefer", - "scene_name": "006556", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[611.14, 273.7], [636.89, 271.08], [653.99, 274.12], [627.9, 276.91], [611.14, 163.66], [636.89, 163.9], [653.99, 163.62], [627.9, 163.37]]}, {\"category\": \"car\", \"corners_3d\": [[252.67, 235.66], [281.31, 231.18], [391.06, 231.74], [370.65, 236.32], [252.67, 174.76], [281.31, 174.62], [391.06, 174.64], [370.65, 174.78]]}, {\"category\": \"car\", \"corners_3d\": [[635.33, 259.5], [723.74, 259.51], [762.4, 288.99], [643.91, 288.97], [635.33, 175.14], [723.74, 175.14], [762.4, 175.92], [643.91, 175.92]]}]\n```" - } - ] - }, - { - "id": 3599, - "data_source": "Mono3DRefer", - "scene_name": "006556", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[611.14, 273.7], [636.89, 271.08], [653.99, 274.12], [627.9, 276.91], [611.14, 163.66], [636.89, 163.9], [653.99, 163.62], [627.9, 163.37]]}, {\"category\": \"car\", \"corners_3d\": [[252.67, 235.66], [281.31, 231.18], [391.06, 231.74], [370.65, 236.32], [252.67, 174.76], [281.31, 174.62], [391.06, 174.64], [370.65, 174.78]]}, {\"category\": \"truck\", \"corners_3d\": [[-212.24, 822.71], [-1258.4, 915.75], [186.8, 315.18], [371.22, 311.38], [-212.24, -412.65], [-1258.4, -496.48], [186.8, 44.62], [371.22, 48.04]]}, {\"category\": \"car\", \"corners_3d\": [[635.33, 259.5], [723.74, 259.51], [762.4, 288.99], [643.91, 288.97], [635.33, 175.14], [723.74, 175.14], [762.4, 175.92], [643.91, 175.92]]}]\n```" - } - ] - }, - { - "id": 3600, - "data_source": "Mono3DRefer", - "scene_name": "006566", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[614.46, 213.08], [581.81, 213.1], [584.06, 208.56], [613.03, 208.55], [614.46, 184.14], [581.81, 184.14], [584.06, 182.87], [613.03, 182.86]]}]\n```" - } - ] - }, - { - "id": 3601, - "data_source": "Mono3DRefer", - "scene_name": "006568", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[668.4, 220.47], [729.03, 220.48], [771.22, 237.38], [689.07, 237.37], [668.4, 148.41], [729.03, 148.41], [771.22, 139.73], [689.07, 139.73]]}, {\"category\": \"van\", \"corners_3d\": [[640.38, 199.95], [675.34, 199.93], [690.72, 205.15], [649.02, 205.19], [640.38, 154.72], [675.34, 154.74], [690.72, 151.24], [649.02, 151.21]]}, {\"category\": \"van\", \"corners_3d\": [[563.83, 191.83], [562.12, 192.45], [490.18, 192.43], [494.14, 191.81], [563.83, 157.47], [562.12, 156.97], [490.18, 156.99], [494.14, 157.49]]}]\n```" - } - ] - }, - { - "id": 3602, - "data_source": "Mono3DRefer", - "scene_name": "006568", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[940.57, 300.65], [887.1, 280.93], [1168.76, 280.01], [1272.58, 299.37], [940.57, 165.13], [887.1, 166.32], [1168.76, 166.37], [1272.58, 165.2]]}, {\"category\": \"car\", \"corners_3d\": [[158.02, 332.35], [26.84, 378.69], [-482.23, 378.69], [-236.45, 332.35], [158.02, 117.71], [26.84, 101.69], [-482.23, 101.69], [-236.45, 117.71]]}, {\"category\": \"car\", \"corners_3d\": [[-124.63, 308.33], [-13.84, 287.85], [227.61, 287.79], [159.86, 308.24], [-124.63, 185.14], [-13.84, 183.28], [227.61, 183.28], [159.86, 185.13]]}, {\"category\": \"car\", \"corners_3d\": [[203.42, 249.34], [241.81, 242.09], [424.99, 242.06], [405.79, 249.3], [203.42, 180.82], [241.81, 180.06], [424.99, 180.06], [405.79, 180.81]]}, {\"category\": \"car\", \"corners_3d\": [[317.97, 231.36], [334.33, 226.66], [481.98, 225.37], [478.62, 229.85], [317.97, 166.74], [334.33, 167.23], [481.98, 167.36], [478.62, 166.9]]}, {\"category\": \"car\", \"corners_3d\": [[431.87, 205.61], [439.86, 204.25], [503.62, 204.31], [498.37, 205.67], [431.87, 173.66], [439.86, 173.62], [503.62, 173.62], [498.37, 173.66]]}, {\"category\": \"car\", \"corners_3d\": [[529.73, 203.15], [526.17, 204.51], [449.47, 204.51], [456.31, 203.15], [529.73, 172.85], [526.17, 172.85], [449.47, 172.85], [456.31, 172.85]]}]\n```" - } - ] - }, - { - "id": 3603, - "data_source": "Mono3DRefer", - "scene_name": "006568", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[940.57, 300.65], [887.1, 280.93], [1168.76, 280.01], [1272.58, 299.37], [940.57, 165.13], [887.1, 166.32], [1168.76, 166.37], [1272.58, 165.2]]}, {\"category\": \"car\", \"corners_3d\": [[158.02, 332.35], [26.84, 378.69], [-482.23, 378.69], [-236.45, 332.35], [158.02, 117.71], [26.84, 101.69], [-482.23, 101.69], [-236.45, 117.71]]}, {\"category\": \"car\", \"corners_3d\": [[-124.63, 308.33], [-13.84, 287.85], [227.61, 287.79], [159.86, 308.24], [-124.63, 185.14], [-13.84, 183.28], [227.61, 183.28], [159.86, 185.13]]}, {\"category\": \"van\", \"corners_3d\": [[668.4, 220.47], [729.03, 220.48], [771.22, 237.38], [689.07, 237.37], [668.4, 148.41], [729.03, 148.41], [771.22, 139.73], [689.07, 139.73]]}, {\"category\": \"car\", \"corners_3d\": [[203.42, 249.34], [241.81, 242.09], [424.99, 242.06], [405.79, 249.3], [203.42, 180.82], [241.81, 180.06], [424.99, 180.06], [405.79, 180.81]]}, {\"category\": \"car\", \"corners_3d\": [[317.97, 231.36], [334.33, 226.66], [481.98, 225.37], [478.62, 229.85], [317.97, 166.74], [334.33, 167.23], [481.98, 167.36], [478.62, 166.9]]}, {\"category\": \"van\", \"corners_3d\": [[640.38, 199.95], [675.34, 199.93], [690.72, 205.15], [649.02, 205.19], [640.38, 154.72], [675.34, 154.74], [690.72, 151.24], [649.02, 151.21]]}, {\"category\": \"car\", \"corners_3d\": [[431.87, 205.61], [439.86, 204.25], [503.62, 204.31], [498.37, 205.67], [431.87, 173.66], [439.86, 173.62], [503.62, 173.62], [498.37, 173.66]]}, {\"category\": \"car\", \"corners_3d\": [[529.73, 203.15], [526.17, 204.51], [449.47, 204.51], [456.31, 203.15], [529.73, 172.85], [526.17, 172.85], [449.47, 172.85], [456.31, 172.85]]}, {\"category\": \"van\", \"corners_3d\": [[563.83, 191.83], [562.12, 192.45], [490.18, 192.43], [494.14, 191.81], [563.83, 157.47], [562.12, 156.97], [490.18, 156.99], [494.14, 157.49]]}]\n```" - } - ] - }, - { - "id": 3604, - "data_source": "Mono3DRefer", - "scene_name": "006594", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[498.25, 281.88], [398.55, 281.59], [453.46, 255.04], [528.87, 255.21], [498.25, 179.73], [398.55, 179.71], [453.46, 178.03], [528.87, 178.04]]}, {\"category\": \"car\", \"corners_3d\": [[822.27, 355.79], [984.65, 359.83], [1257.94, 528.97], [943.56, 514.58], [822.27, 205.9], [984.65, 206.63], [1257.94, 237.18], [943.56, 234.58]]}, {\"category\": \"car\", \"corners_3d\": [[713.36, 268.8], [794.65, 268.59], [862.11, 301.05], [753.37, 301.43], [713.36, 181.61], [794.65, 181.59], [862.11, 184.55], [753.37, 184.59]]}, {\"category\": \"car\", \"corners_3d\": [[666.48, 238.72], [728.16, 238.56], [760.61, 253.82], [684.66, 254.07], [666.48, 179.04], [728.16, 179.03], [760.61, 180.46], [684.66, 180.48]]}, {\"category\": \"car\", \"corners_3d\": [[491.07, 243.19], [543.82, 243.19], [530.27, 257.57], [466.74, 257.56], [491.07, 189.02], [543.82, 189.02], [530.27, 192.33], [466.74, 192.32]]}, {\"category\": \"car\", \"corners_3d\": [[544.88, 231.66], [492.0, 231.62], [508.14, 223.98], [554.16, 224.01], [544.88, 184.87], [492.0, 184.86], [508.14, 183.3], [554.16, 183.3]]}, {\"category\": \"car\", \"corners_3d\": [[653.75, 221.84], [697.5, 221.57], [720.01, 228.76], [669.9, 229.12], [653.75, 177.43], [697.5, 177.4], [720.01, 178.07], [669.9, 178.11]]}, {\"category\": \"car\", \"corners_3d\": [[638.45, 212.28], [676.29, 212.2], [689.79, 217.79], [646.6, 217.89], [638.45, 177.12], [676.29, 177.11], [689.79, 177.72], [646.6, 177.73]]}, {\"category\": \"car\", \"corners_3d\": [[564.72, 207.61], [534.72, 207.57], [543.27, 204.48], [570.6, 204.51], [564.72, 177.31], [534.72, 177.31], [543.27, 176.91], [570.6, 176.92]]}, {\"category\": \"car\", \"corners_3d\": [[580.17, 203.08], [552.02, 203.05], [558.47, 200.33], [584.07, 200.35], [580.17, 170.64], [552.02, 170.64], [558.47, 170.84], [584.07, 170.84]]}, {\"category\": \"car\", \"corners_3d\": [[590.42, 200.0], [565.34, 199.94], [572.39, 197.75], [595.44, 197.8], [590.42, 172.7], [565.34, 172.7], [572.39, 172.71], [595.44, 172.71]]}]\n```" - } - ] - }, - { - "id": 3605, - "data_source": "Mono3DRefer", - "scene_name": "006597", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[490.9, 241.77], [467.47, 241.75], [481.83, 235.1], [503.0, 235.12], [490.9, 172.84], [467.47, 172.84], [481.83, 172.84], [503.0, 172.84]]}, {\"category\": \"cyclist\", \"corners_3d\": [[523.95, 238.08], [499.99, 238.21], [505.49, 232.73], [527.44, 232.62], [523.95, 174.64], [499.99, 174.65], [505.49, 174.5], [527.44, 174.49]]}, {\"category\": \"cyclist\", \"corners_3d\": [[542.67, 212.35], [528.63, 212.35], [532.51, 210.26], [545.8, 210.25], [542.67, 174.55], [528.63, 174.55], [532.51, 174.46], [545.8, 174.46]]}]\n```" - } - ] - }, - { - "id": 3606, - "data_source": "Mono3DRefer", - "scene_name": "006597", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[238.02, 339.85], [84.25, 340.58], [250.67, 285.88], [353.95, 285.55], [238.02, 191.99], [84.25, 192.07], [250.67, 185.8], [353.95, 185.77]]}, {\"category\": \"car\", \"corners_3d\": [[471.97, 228.57], [420.82, 228.58], [453.33, 218.95], [495.64, 218.95], [471.97, 183.87], [420.82, 183.87], [453.33, 181.96], [495.64, 181.96]]}, {\"category\": \"car\", \"corners_3d\": [[535.75, 206.42], [507.05, 206.3], [522.56, 203.19], [548.64, 203.29], [535.75, 182.46], [507.05, 182.43], [522.56, 181.54], [548.64, 181.57]]}, {\"category\": \"car\", \"corners_3d\": [[642.59, 199.13], [669.85, 199.2], [670.68, 201.61], [640.93, 201.53], [642.59, 176.83], [669.85, 176.85], [670.68, 177.21], [640.93, 177.2]]}, {\"category\": \"car\", \"corners_3d\": [[772.6, 318.28], [920.13, 318.01], [1173.58, 433.99], [908.64, 434.88], [772.6, 175.48], [920.13, 175.48], [1173.58, 177.57], [908.64, 177.59]]}]\n```" - } - ] - }, - { - "id": 3607, - "data_source": "Mono3DRefer", - "scene_name": "006597", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[238.02, 339.85], [84.25, 340.58], [250.67, 285.88], [353.95, 285.55], [238.02, 191.99], [84.25, 192.07], [250.67, 185.8], [353.95, 185.77]]}, {\"category\": \"cyclist\", \"corners_3d\": [[490.9, 241.77], [467.47, 241.75], [481.83, 235.1], [503.0, 235.12], [490.9, 172.84], [467.47, 172.84], [481.83, 172.84], [503.0, 172.84]]}, {\"category\": \"cyclist\", \"corners_3d\": [[523.95, 238.08], [499.99, 238.21], [505.49, 232.73], [527.44, 232.62], [523.95, 174.64], [499.99, 174.65], [505.49, 174.5], [527.44, 174.49]]}, {\"category\": \"car\", \"corners_3d\": [[471.97, 228.57], [420.82, 228.58], [453.33, 218.95], [495.64, 218.95], [471.97, 183.87], [420.82, 183.87], [453.33, 181.96], [495.64, 181.96]]}, {\"category\": \"cyclist\", \"corners_3d\": [[542.67, 212.35], [528.63, 212.35], [532.51, 210.26], [545.8, 210.25], [542.67, 174.55], [528.63, 174.55], [532.51, 174.46], [545.8, 174.46]]}, {\"category\": \"car\", \"corners_3d\": [[535.75, 206.42], [507.05, 206.3], [522.56, 203.19], [548.64, 203.29], [535.75, 182.46], [507.05, 182.43], [522.56, 181.54], [548.64, 181.57]]}, {\"category\": \"car\", \"corners_3d\": [[642.59, 199.13], [669.85, 199.2], [670.68, 201.61], [640.93, 201.53], [642.59, 176.83], [669.85, 176.85], [670.68, 177.21], [640.93, 177.2]]}, {\"category\": \"car\", \"corners_3d\": [[772.6, 318.28], [920.13, 318.01], [1173.58, 433.99], [908.64, 434.88], [772.6, 175.48], [920.13, 175.48], [1173.58, 177.57], [908.64, 177.59]]}]\n```" - } - ] - }, - { - "id": 3608, - "data_source": "Mono3DRefer", - "scene_name": "006606", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[558.16, 196.46], [531.35, 196.41], [541.32, 195.35], [566.25, 195.4], [558.16, 174.77], [531.35, 174.79], [541.32, 175.27], [566.25, 175.25]]}]\n```" - } - ] - }, - { - "id": 3609, - "data_source": "Mono3DRefer", - "scene_name": "006610", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[42.55, 237.34], [-12.9, 237.09], [103.25, 225.75], [149.26, 225.92], [42.55, 172.84], [-12.9, 172.84], [103.25, 172.84], [149.26, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[176.34, 226.47], [137.43, 226.2], [203.47, 219.63], [237.86, 219.84], [176.34, 184.13], [137.43, 184.08], [203.47, 182.7], [237.86, 182.74]]}]\n```" - } - ] - }, - { - "id": 3610, - "data_source": "Mono3DRefer", - "scene_name": "006638", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[134.56, 273.96], [186.18, 262.77], [382.93, 262.49], [355.87, 273.6], [134.56, 188.74], [186.18, 186.98], [382.93, 186.94], [355.87, 188.69]]}, {\"category\": \"car\", \"corners_3d\": [[984.45, 255.75], [1034.72, 266.07], [822.09, 267.2], [795.08, 256.64], [984.45, 181.72], [1034.72, 182.82], [822.09, 182.94], [795.08, 181.82]]}, {\"category\": \"car\", \"corners_3d\": [[4.01, 296.46], [82.11, 280.09], [326.98, 279.32], [286.58, 295.43], [4.01, 194.14], [82.11, 191.32], [326.98, 191.19], [286.58, 193.97]]}, {\"category\": \"car\", \"corners_3d\": [[-206.48, 339.38], [-67.47, 309.09], [265.37, 304.87], [202.8, 333.12], [-206.48, 169.95], [-67.47, 170.48], [265.37, 170.55], [202.8, 170.06]]}, {\"category\": \"car\", \"corners_3d\": [[170.53, 361.95], [31.04, 421.05], [-566.18, 418.98], [-286.26, 360.74], [170.53, 198.75], [31.04, 206.85], [-566.18, 206.57], [-286.26, 198.59]]}, {\"category\": \"car\", \"corners_3d\": [[524.38, 211.16], [521.61, 213.04], [430.78, 213.24], [437.83, 211.34], [524.38, 178.96], [521.61, 179.26], [430.78, 179.29], [437.83, 178.99]]}, {\"category\": \"car\", \"corners_3d\": [[641.06, 213.96], [679.67, 213.9], [693.88, 220.29], [649.26, 220.38], [641.06, 182.59], [679.67, 182.57], [693.88, 184.09], [649.26, 184.11]]}, {\"category\": \"car\", \"corners_3d\": [[472.78, 229.79], [462.62, 234.02], [316.54, 234.02], [336.81, 229.79], [472.78, 179.24], [462.62, 179.72], [316.54, 179.72], [336.81, 179.24]]}]\n```" - } - ] - }, - { - "id": 3611, - "data_source": "Mono3DRefer", - "scene_name": "006641", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[234.68, 225.64], [200.2, 226.68], [202.89, 222.28], [234.56, 221.41], [234.68, 198.78], [200.2, 199.29], [202.89, 197.13], [234.56, 196.7]]}]\n```" - } - ] - }, - { - "id": 3612, - "data_source": "Mono3DRefer", - "scene_name": "006646", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[442.89, 318.4], [318.48, 316.63], [416.7, 275.23], [505.83, 276.12], [442.89, 190.1], [318.48, 189.89], [416.7, 184.98], [505.83, 185.09]]}, {\"category\": \"car\", \"corners_3d\": [[501.12, 194.96], [481.37, 195.1], [478.46, 193.78], [497.06, 193.65], [501.12, 174.88], [481.37, 174.89], [478.46, 174.77], [497.06, 174.76]]}, {\"category\": \"car\", \"corners_3d\": [[-175.31, 677.54], [-610.11, 674.71], [106.75, 381.34], [288.46, 381.83], [-175.31, 241.59], [-610.11, 241.2], [106.75, 201.25], [288.46, 201.31]]}, {\"category\": \"car\", \"corners_3d\": [[-76.39, 309.55], [21.38, 291.61], [243.34, 294.81], [177.09, 313.81], [-76.39, 206.84], [21.38, 202.38], [243.34, 203.18], [177.09, 207.9]]}]\n```" - } - ] - }, - { - "id": 3613, - "data_source": "Mono3DRefer", - "scene_name": "006652", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1002.81, 317.41], [973.34, 307.5], [1081.09, 307.27], [1118.9, 317.15], [1002.81, 144.17], [973.34, 146.8], [1081.09, 146.86], [1118.9, 144.24]]}]\n```" - } - ] - }, - { - "id": 3614, - "data_source": "Mono3DRefer", - "scene_name": "006654", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1237.3, 271.96], [1345.64, 287.3], [1064.89, 292.27], [991.52, 275.67], [1237.3, 183.92], [1345.64, 185.63], [1064.89, 186.19], [991.52, 184.33]]}, {\"category\": \"car\", \"corners_3d\": [[1187.04, 254.38], [1275.04, 265.67], [1005.5, 268.67], [948.73, 256.68], [1187.04, 182.62], [1275.04, 183.97], [1005.5, 184.33], [948.73, 182.9]]}]\n```" - } - ] - }, - { - "id": 3615, - "data_source": "Mono3DRefer", - "scene_name": "006656", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[274.63, 300.82], [301.06, 291.2], [392.51, 291.22], [374.02, 300.85], [274.63, 143.98], [301.06, 146.9], [392.51, 146.9], [374.02, 143.98]]}]\n```" - } - ] - }, - { - "id": 3616, - "data_source": "Mono3DRefer", - "scene_name": "006660", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[839.9, 346.91], [993.77, 346.57], [1356.86, 507.95], [1060.89, 509.22], [839.9, 204.56], [993.77, 204.5], [1356.86, 233.9], [1060.89, 234.13]]}, {\"category\": \"car\", \"corners_3d\": [[99.57, 373.18], [-82.09, 444.53], [-604.28, 444.53], [-285.48, 373.18], [99.57, 167.98], [-82.09, 166.25], [-604.28, 166.25], [-285.48, 167.98]]}, {\"category\": \"car\", \"corners_3d\": [[-13.59, 310.82], [67.18, 290.81], [302.23, 287.51], [262.66, 306.33], [-13.59, 181.55], [67.18, 180.29], [302.23, 180.08], [262.66, 181.27]]}, {\"category\": \"car\", \"corners_3d\": [[869.23, 274.14], [845.14, 262.8], [1065.4, 264.46], [1118.33, 276.25], [869.23, 193.81], [845.14, 191.47], [1065.4, 191.81], [1118.33, 194.25]]}, {\"category\": \"car\", \"corners_3d\": [[984.14, 253.45], [1015.65, 262.06], [835.79, 259.75], [821.97, 251.56], [984.14, 178.79], [1015.65, 179.43], [835.79, 179.26], [821.97, 178.65]]}, {\"category\": \"car\", \"corners_3d\": [[745.46, 228.6], [734.08, 224.6], [865.7, 224.3], [887.12, 228.25], [745.46, 171.21], [734.08, 171.33], [865.7, 171.34], [887.12, 171.22]]}, {\"category\": \"car\", \"corners_3d\": [[438.51, 233.95], [423.54, 239.53], [262.73, 239.68], [291.21, 234.07], [438.51, 154.15], [423.54, 152.44], [262.73, 152.4], [291.21, 154.11]]}, {\"category\": \"car\", \"corners_3d\": [[379.42, 216.73], [391.5, 214.34], [496.88, 214.27], [490.88, 216.65], [379.42, 179.04], [391.5, 178.7], [496.88, 178.69], [490.88, 179.03]]}, {\"category\": \"car\", \"corners_3d\": [[424.05, 211.44], [429.01, 209.47], [521.99, 208.91], [522.02, 210.82], [424.05, 171.74], [429.01, 171.8], [521.99, 171.82], [522.02, 171.76]]}]\n```" - } - ] - }, - { - "id": 3617, - "data_source": "Mono3DRefer", - "scene_name": "006660", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[641.88, 207.54], [681.15, 207.52], [697.15, 214.61], [649.85, 214.64], [641.88, 160.81], [681.15, 160.82], [697.15, 158.36], [649.85, 158.35]]}, {\"category\": \"van\", \"corners_3d\": [[626.95, 195.55], [653.66, 195.54], [660.82, 198.74], [630.34, 198.75], [626.95, 161.03], [653.66, 161.03], [660.82, 159.37], [630.34, 159.36]]}, {\"category\": \"van\", \"corners_3d\": [[565.79, 193.62], [564.63, 194.17], [505.37, 194.17], [508.07, 193.62], [565.79, 165.2], [564.63, 165.0], [505.37, 165.0], [508.07, 165.2]]}]\n```" - } - ] - }, - { - "id": 3618, - "data_source": "Mono3DRefer", - "scene_name": "006660", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[839.9, 346.91], [993.77, 346.57], [1356.86, 507.95], [1060.89, 509.22], [839.9, 204.56], [993.77, 204.5], [1356.86, 233.9], [1060.89, 234.13]]}, {\"category\": \"car\", \"corners_3d\": [[99.57, 373.18], [-82.09, 444.53], [-604.28, 444.53], [-285.48, 373.18], [99.57, 167.98], [-82.09, 166.25], [-604.28, 166.25], [-285.48, 167.98]]}, {\"category\": \"car\", \"corners_3d\": [[-13.59, 310.82], [67.18, 290.81], [302.23, 287.51], [262.66, 306.33], [-13.59, 181.55], [67.18, 180.29], [302.23, 180.08], [262.66, 181.27]]}, {\"category\": \"car\", \"corners_3d\": [[869.23, 274.14], [845.14, 262.8], [1065.4, 264.46], [1118.33, 276.25], [869.23, 193.81], [845.14, 191.47], [1065.4, 191.81], [1118.33, 194.25]]}, {\"category\": \"car\", \"corners_3d\": [[984.14, 253.45], [1015.65, 262.06], [835.79, 259.75], [821.97, 251.56], [984.14, 178.79], [1015.65, 179.43], [835.79, 179.26], [821.97, 178.65]]}, {\"category\": \"car\", \"corners_3d\": [[745.46, 228.6], [734.08, 224.6], [865.7, 224.3], [887.12, 228.25], [745.46, 171.21], [734.08, 171.33], [865.7, 171.34], [887.12, 171.22]]}, {\"category\": \"car\", \"corners_3d\": [[438.51, 233.95], [423.54, 239.53], [262.73, 239.68], [291.21, 234.07], [438.51, 154.15], [423.54, 152.44], [262.73, 152.4], [291.21, 154.11]]}, {\"category\": \"van\", \"corners_3d\": [[641.88, 207.54], [681.15, 207.52], [697.15, 214.61], [649.85, 214.64], [641.88, 160.81], [681.15, 160.82], [697.15, 158.36], [649.85, 158.35]]}, {\"category\": \"car\", \"corners_3d\": [[379.42, 216.73], [391.5, 214.34], [496.88, 214.27], [490.88, 216.65], [379.42, 179.04], [391.5, 178.7], [496.88, 178.69], [490.88, 179.03]]}, {\"category\": \"car\", \"corners_3d\": [[424.05, 211.44], [429.01, 209.47], [521.99, 208.91], [522.02, 210.82], [424.05, 171.74], [429.01, 171.8], [521.99, 171.82], [522.02, 171.76]]}, {\"category\": \"van\", \"corners_3d\": [[626.95, 195.55], [653.66, 195.54], [660.82, 198.74], [630.34, 198.75], [626.95, 161.03], [653.66, 161.03], [660.82, 159.37], [630.34, 159.36]]}, {\"category\": \"van\", \"corners_3d\": [[565.79, 193.62], [564.63, 194.17], [505.37, 194.17], [508.07, 193.62], [565.79, 165.2], [564.63, 165.0], [505.37, 165.0], [508.07, 165.2]]}]\n```" - } - ] - }, - { - "id": 3619, - "data_source": "Mono3DRefer", - "scene_name": "006663", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[823.3, 177.66], [851.94, 177.74], [839.79, 178.48], [807.13, 178.38], [823.3, 144.44], [851.94, 144.0], [839.79, 139.65], [807.13, 140.23]]}]\n```" - } - ] - }, - { - "id": 3620, - "data_source": "Mono3DRefer", - "scene_name": "006663", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[238.69, 394.05], [42.03, 392.27], [265.01, 309.15], [387.92, 309.83], [238.69, 196.65], [42.03, 196.46], [265.01, 187.52], [387.92, 187.59]]}, {\"category\": \"car\", \"corners_3d\": [[816.31, 289.8], [933.32, 289.81], [1176.53, 377.82], [971.43, 377.78], [816.31, 181.34], [933.32, 181.34], [1176.53, 187.72], [971.43, 187.72]]}, {\"category\": \"car\", \"corners_3d\": [[449.44, 269.06], [355.59, 268.82], [422.19, 244.96], [492.78, 245.09], [449.44, 169.4], [355.59, 169.41], [422.19, 170.26], [492.78, 170.26]]}, {\"category\": \"car\", \"corners_3d\": [[543.1, 234.96], [493.36, 233.89], [535.65, 225.05], [578.72, 225.83], [543.1, 174.68], [493.36, 174.65], [535.65, 174.39], [578.72, 174.41]]}, {\"category\": \"car\", \"corners_3d\": [[752.95, 242.09], [824.05, 242.84], [864.57, 263.96], [771.74, 262.71], [752.95, 175.72], [824.05, 175.76], [864.57, 176.63], [771.74, 176.58]]}, {\"category\": \"car\", \"corners_3d\": [[746.24, 226.92], [797.43, 227.6], [810.21, 240.08], [747.32, 239.07], [746.24, 179.13], [797.43, 179.21], [810.21, 180.66], [747.32, 180.54]]}, {\"category\": \"car\", \"corners_3d\": [[676.86, 212.81], [640.66, 211.84], [676.3, 207.68], [709.35, 208.46], [676.86, 175.51], [640.66, 175.44], [676.3, 175.17], [709.35, 175.22]]}]\n```" - } - ] - }, - { - "id": 3621, - "data_source": "Mono3DRefer", - "scene_name": "006663", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[238.69, 394.05], [42.03, 392.27], [265.01, 309.15], [387.92, 309.83], [238.69, 196.65], [42.03, 196.46], [265.01, 187.52], [387.92, 187.59]]}, {\"category\": \"car\", \"corners_3d\": [[816.31, 289.8], [933.32, 289.81], [1176.53, 377.82], [971.43, 377.78], [816.31, 181.34], [933.32, 181.34], [1176.53, 187.72], [971.43, 187.72]]}, {\"category\": \"car\", \"corners_3d\": [[449.44, 269.06], [355.59, 268.82], [422.19, 244.96], [492.78, 245.09], [449.44, 169.4], [355.59, 169.41], [422.19, 170.26], [492.78, 170.26]]}, {\"category\": \"car\", \"corners_3d\": [[543.1, 234.96], [493.36, 233.89], [535.65, 225.05], [578.72, 225.83], [543.1, 174.68], [493.36, 174.65], [535.65, 174.39], [578.72, 174.41]]}, {\"category\": \"car\", \"corners_3d\": [[752.95, 242.09], [824.05, 242.84], [864.57, 263.96], [771.74, 262.71], [752.95, 175.72], [824.05, 175.76], [864.57, 176.63], [771.74, 176.58]]}, {\"category\": \"car\", \"corners_3d\": [[746.24, 226.92], [797.43, 227.6], [810.21, 240.08], [747.32, 239.07], [746.24, 179.13], [797.43, 179.21], [810.21, 180.66], [747.32, 180.54]]}, {\"category\": \"car\", \"corners_3d\": [[676.86, 212.81], [640.66, 211.84], [676.3, 207.68], [709.35, 208.46], [676.86, 175.51], [640.66, 175.44], [676.3, 175.17], [709.35, 175.22]]}, {\"category\": \"truck\", \"corners_3d\": [[823.3, 177.66], [851.94, 177.74], [839.79, 178.48], [807.13, 178.38], [823.3, 144.44], [851.94, 144.0], [839.79, 139.65], [807.13, 140.23]]}]\n```" - } - ] - }, - { - "id": 3622, - "data_source": "Mono3DRefer", - "scene_name": "006683", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[389.78, 225.72], [356.02, 226.62], [352.76, 222.61], [384.1, 221.84], [389.78, 195.64], [356.02, 196.03], [352.76, 194.3], [384.1, 193.96]]}, {\"category\": \"car\", \"corners_3d\": [[365.18, 375.18], [166.31, 379.58], [323.8, 294.98], [440.32, 293.43], [365.18, 210.83], [166.31, 211.65], [323.8, 195.77], [440.32, 195.48]]}]\n```" - } - ] - }, - { - "id": 3623, - "data_source": "Mono3DRefer", - "scene_name": "006697", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[551.86, 198.13], [522.92, 198.07], [533.71, 196.81], [560.47, 196.86], [551.86, 174.73], [522.92, 174.76], [533.71, 175.27], [560.47, 175.25]]}]\n```" - } - ] - }, - { - "id": 3624, - "data_source": "Mono3DRefer", - "scene_name": "006703", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[52.53, 362.13], [131.93, 358.79], [100.23, 375.77], [12.42, 379.78], [52.53, 157.76], [131.93, 158.2], [100.23, 155.97], [12.42, 155.44]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[105.38, 356.59], [45.83, 356.16], [87.0, 343.66], [142.38, 344.04], [105.38, 158.49], [45.83, 158.55], [87.0, 160.19], [142.38, 160.14]]}]\n```" - } - ] - }, - { - "id": 3625, - "data_source": "Mono3DRefer", - "scene_name": "006703", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[576.42, 195.44], [590.56, 196.21], [387.99, 197.81], [384.15, 196.87], [576.42, 141.97], [590.56, 139.78], [387.99, 135.23], [384.15, 137.92]]}]\n```" - } - ] - }, - { - "id": 3626, - "data_source": "Mono3DRefer", - "scene_name": "006703", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[328.81, 357.88], [408.01, 357.89], [382.15, 381.52], [292.35, 381.5], [328.81, 208.5], [408.01, 208.5], [382.15, 212.12], [292.35, 212.12]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[259.99, 356.06], [336.62, 356.07], [300.82, 379.71], [213.82, 379.69], [259.99, 209.52], [336.62, 209.53], [300.82, 213.32], [213.82, 213.32]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[354.18, 346.94], [291.97, 346.52], [320.52, 332.21], [377.39, 332.57], [354.18, 209.13], [291.97, 209.06], [320.52, 206.67], [377.39, 206.73]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[361.04, 310.87], [407.31, 311.04], [389.75, 321.97], [339.6, 321.77], [361.04, 201.7], [407.31, 201.72], [389.75, 203.43], [339.6, 203.4]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[415.99, 300.91], [369.79, 300.84], [384.63, 293.37], [427.95, 293.44], [415.99, 194.4], [369.79, 194.39], [384.63, 193.59], [427.95, 193.59]]}]\n```" - } - ] - }, - { - "id": 3627, - "data_source": "Mono3DRefer", - "scene_name": "006703", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[952.2, 412.38], [1206.25, 410.8], [3361.43, 1208.26], [2264.06, 1240.79], [952.2, 134.46], [1206.25, 134.79], [3361.43, -27.73], [2264.06, -34.35]]}]\n```" - } - ] - }, - { - "id": 3628, - "data_source": "Mono3DRefer", - "scene_name": "006703", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[576.42, 195.44], [590.56, 196.21], [387.99, 197.81], [384.15, 196.87], [576.42, 141.97], [590.56, 139.78], [387.99, 135.23], [384.15, 137.92]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[52.53, 362.13], [131.93, 358.79], [100.23, 375.77], [12.42, 379.78], [52.53, 157.76], [131.93, 158.2], [100.23, 155.97], [12.42, 155.44]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[105.38, 356.59], [45.83, 356.16], [87.0, 343.66], [142.38, 344.04], [105.38, 158.49], [45.83, 158.55], [87.0, 160.19], [142.38, 160.14]]}]\n```" - } - ] - }, - { - "id": 3629, - "data_source": "Mono3DRefer", - "scene_name": "006703", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[328.81, 357.88], [408.01, 357.89], [382.15, 381.52], [292.35, 381.5], [328.81, 208.5], [408.01, 208.5], [382.15, 212.12], [292.35, 212.12]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[259.99, 356.06], [336.62, 356.07], [300.82, 379.71], [213.82, 379.69], [259.99, 209.52], [336.62, 209.53], [300.82, 213.32], [213.82, 213.32]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[354.18, 346.94], [291.97, 346.52], [320.52, 332.21], [377.39, 332.57], [354.18, 209.13], [291.97, 209.06], [320.52, 206.67], [377.39, 206.73]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[361.04, 310.87], [407.31, 311.04], [389.75, 321.97], [339.6, 321.77], [361.04, 201.7], [407.31, 201.72], [389.75, 203.43], [339.6, 203.4]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[415.99, 300.91], [369.79, 300.84], [384.63, 293.37], [427.95, 293.44], [415.99, 194.4], [369.79, 194.39], [384.63, 193.59], [427.95, 193.59]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[52.53, 362.13], [131.93, 358.79], [100.23, 375.77], [12.42, 379.78], [52.53, 157.76], [131.93, 158.2], [100.23, 155.97], [12.42, 155.44]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[105.38, 356.59], [45.83, 356.16], [87.0, 343.66], [142.38, 344.04], [105.38, 158.49], [45.83, 158.55], [87.0, 160.19], [142.38, 160.14]]}]\n```" - } - ] - }, - { - "id": 3630, - "data_source": "Mono3DRefer", - "scene_name": "006703", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[952.2, 412.38], [1206.25, 410.8], [3361.43, 1208.26], [2264.06, 1240.79], [952.2, 134.46], [1206.25, 134.79], [3361.43, -27.73], [2264.06, -34.35]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[52.53, 362.13], [131.93, 358.79], [100.23, 375.77], [12.42, 379.78], [52.53, 157.76], [131.93, 158.2], [100.23, 155.97], [12.42, 155.44]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[105.38, 356.59], [45.83, 356.16], [87.0, 343.66], [142.38, 344.04], [105.38, 158.49], [45.83, 158.55], [87.0, 160.19], [142.38, 160.14]]}]\n```" - } - ] - }, - { - "id": 3631, - "data_source": "Mono3DRefer", - "scene_name": "006703", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[576.42, 195.44], [590.56, 196.21], [387.99, 197.81], [384.15, 196.87], [576.42, 141.97], [590.56, 139.78], [387.99, 135.23], [384.15, 137.92]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[328.81, 357.88], [408.01, 357.89], [382.15, 381.52], [292.35, 381.5], [328.81, 208.5], [408.01, 208.5], [382.15, 212.12], [292.35, 212.12]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[259.99, 356.06], [336.62, 356.07], [300.82, 379.71], [213.82, 379.69], [259.99, 209.52], [336.62, 209.53], [300.82, 213.32], [213.82, 213.32]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[354.18, 346.94], [291.97, 346.52], [320.52, 332.21], [377.39, 332.57], [354.18, 209.13], [291.97, 209.06], [320.52, 206.67], [377.39, 206.73]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[361.04, 310.87], [407.31, 311.04], [389.75, 321.97], [339.6, 321.77], [361.04, 201.7], [407.31, 201.72], [389.75, 203.43], [339.6, 203.4]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[415.99, 300.91], [369.79, 300.84], [384.63, 293.37], [427.95, 293.44], [415.99, 194.4], [369.79, 194.39], [384.63, 193.59], [427.95, 193.59]]}]\n```" - } - ] - }, - { - "id": 3632, - "data_source": "Mono3DRefer", - "scene_name": "006703", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[952.2, 412.38], [1206.25, 410.8], [3361.43, 1208.26], [2264.06, 1240.79], [952.2, 134.46], [1206.25, 134.79], [3361.43, -27.73], [2264.06, -34.35]]}, {\"category\": \"tram\", \"corners_3d\": [[576.42, 195.44], [590.56, 196.21], [387.99, 197.81], [384.15, 196.87], [576.42, 141.97], [590.56, 139.78], [387.99, 135.23], [384.15, 137.92]]}]\n```" - } - ] - }, - { - "id": 3633, - "data_source": "Mono3DRefer", - "scene_name": "006703", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[952.2, 412.38], [1206.25, 410.8], [3361.43, 1208.26], [2264.06, 1240.79], [952.2, 134.46], [1206.25, 134.79], [3361.43, -27.73], [2264.06, -34.35]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[328.81, 357.88], [408.01, 357.89], [382.15, 381.52], [292.35, 381.5], [328.81, 208.5], [408.01, 208.5], [382.15, 212.12], [292.35, 212.12]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[259.99, 356.06], [336.62, 356.07], [300.82, 379.71], [213.82, 379.69], [259.99, 209.52], [336.62, 209.53], [300.82, 213.32], [213.82, 213.32]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[354.18, 346.94], [291.97, 346.52], [320.52, 332.21], [377.39, 332.57], [354.18, 209.13], [291.97, 209.06], [320.52, 206.67], [377.39, 206.73]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[361.04, 310.87], [407.31, 311.04], [389.75, 321.97], [339.6, 321.77], [361.04, 201.7], [407.31, 201.72], [389.75, 203.43], [339.6, 203.4]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[415.99, 300.91], [369.79, 300.84], [384.63, 293.37], [427.95, 293.44], [415.99, 194.4], [369.79, 194.39], [384.63, 193.59], [427.95, 193.59]]}]\n```" - } - ] - }, - { - "id": 3634, - "data_source": "Mono3DRefer", - "scene_name": "006703", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, tram, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[576.42, 195.44], [590.56, 196.21], [387.99, 197.81], [384.15, 196.87], [576.42, 141.97], [590.56, 139.78], [387.99, 135.23], [384.15, 137.92]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[328.81, 357.88], [408.01, 357.89], [382.15, 381.52], [292.35, 381.5], [328.81, 208.5], [408.01, 208.5], [382.15, 212.12], [292.35, 212.12]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[259.99, 356.06], [336.62, 356.07], [300.82, 379.71], [213.82, 379.69], [259.99, 209.52], [336.62, 209.53], [300.82, 213.32], [213.82, 213.32]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[354.18, 346.94], [291.97, 346.52], [320.52, 332.21], [377.39, 332.57], [354.18, 209.13], [291.97, 209.06], [320.52, 206.67], [377.39, 206.73]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[361.04, 310.87], [407.31, 311.04], [389.75, 321.97], [339.6, 321.77], [361.04, 201.7], [407.31, 201.72], [389.75, 203.43], [339.6, 203.4]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[415.99, 300.91], [369.79, 300.84], [384.63, 293.37], [427.95, 293.44], [415.99, 194.4], [369.79, 194.39], [384.63, 193.59], [427.95, 193.59]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[52.53, 362.13], [131.93, 358.79], [100.23, 375.77], [12.42, 379.78], [52.53, 157.76], [131.93, 158.2], [100.23, 155.97], [12.42, 155.44]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[105.38, 356.59], [45.83, 356.16], [87.0, 343.66], [142.38, 344.04], [105.38, 158.49], [45.83, 158.55], [87.0, 160.19], [142.38, 160.14]]}]\n```" - } - ] - }, - { - "id": 3635, - "data_source": "Mono3DRefer", - "scene_name": "006703", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, tram, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[952.2, 412.38], [1206.25, 410.8], [3361.43, 1208.26], [2264.06, 1240.79], [952.2, 134.46], [1206.25, 134.79], [3361.43, -27.73], [2264.06, -34.35]]}, {\"category\": \"tram\", \"corners_3d\": [[576.42, 195.44], [590.56, 196.21], [387.99, 197.81], [384.15, 196.87], [576.42, 141.97], [590.56, 139.78], [387.99, 135.23], [384.15, 137.92]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[52.53, 362.13], [131.93, 358.79], [100.23, 375.77], [12.42, 379.78], [52.53, 157.76], [131.93, 158.2], [100.23, 155.97], [12.42, 155.44]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[105.38, 356.59], [45.83, 356.16], [87.0, 343.66], [142.38, 344.04], [105.38, 158.49], [45.83, 158.55], [87.0, 160.19], [142.38, 160.14]]}]\n```" - } - ] - }, - { - "id": 3636, - "data_source": "Mono3DRefer", - "scene_name": "006703", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, person_sitting, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[952.2, 412.38], [1206.25, 410.8], [3361.43, 1208.26], [2264.06, 1240.79], [952.2, 134.46], [1206.25, 134.79], [3361.43, -27.73], [2264.06, -34.35]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[328.81, 357.88], [408.01, 357.89], [382.15, 381.52], [292.35, 381.5], [328.81, 208.5], [408.01, 208.5], [382.15, 212.12], [292.35, 212.12]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[259.99, 356.06], [336.62, 356.07], [300.82, 379.71], [213.82, 379.69], [259.99, 209.52], [336.62, 209.53], [300.82, 213.32], [213.82, 213.32]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[354.18, 346.94], [291.97, 346.52], [320.52, 332.21], [377.39, 332.57], [354.18, 209.13], [291.97, 209.06], [320.52, 206.67], [377.39, 206.73]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[361.04, 310.87], [407.31, 311.04], [389.75, 321.97], [339.6, 321.77], [361.04, 201.7], [407.31, 201.72], [389.75, 203.43], [339.6, 203.4]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[415.99, 300.91], [369.79, 300.84], [384.63, 293.37], [427.95, 293.44], [415.99, 194.4], [369.79, 194.39], [384.63, 193.59], [427.95, 193.59]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[52.53, 362.13], [131.93, 358.79], [100.23, 375.77], [12.42, 379.78], [52.53, 157.76], [131.93, 158.2], [100.23, 155.97], [12.42, 155.44]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[105.38, 356.59], [45.83, 356.16], [87.0, 343.66], [142.38, 344.04], [105.38, 158.49], [45.83, 158.55], [87.0, 160.19], [142.38, 160.14]]}]\n```" - } - ] - }, - { - "id": 3637, - "data_source": "Mono3DRefer", - "scene_name": "006703", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, person_sitting, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[952.2, 412.38], [1206.25, 410.8], [3361.43, 1208.26], [2264.06, 1240.79], [952.2, 134.46], [1206.25, 134.79], [3361.43, -27.73], [2264.06, -34.35]]}, {\"category\": \"tram\", \"corners_3d\": [[576.42, 195.44], [590.56, 196.21], [387.99, 197.81], [384.15, 196.87], [576.42, 141.97], [590.56, 139.78], [387.99, 135.23], [384.15, 137.92]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[328.81, 357.88], [408.01, 357.89], [382.15, 381.52], [292.35, 381.5], [328.81, 208.5], [408.01, 208.5], [382.15, 212.12], [292.35, 212.12]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[259.99, 356.06], [336.62, 356.07], [300.82, 379.71], [213.82, 379.69], [259.99, 209.52], [336.62, 209.53], [300.82, 213.32], [213.82, 213.32]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[354.18, 346.94], [291.97, 346.52], [320.52, 332.21], [377.39, 332.57], [354.18, 209.13], [291.97, 209.06], [320.52, 206.67], [377.39, 206.73]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[361.04, 310.87], [407.31, 311.04], [389.75, 321.97], [339.6, 321.77], [361.04, 201.7], [407.31, 201.72], [389.75, 203.43], [339.6, 203.4]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[415.99, 300.91], [369.79, 300.84], [384.63, 293.37], [427.95, 293.44], [415.99, 194.4], [369.79, 194.39], [384.63, 193.59], [427.95, 193.59]]}]\n```" - } - ] - }, - { - "id": 3638, - "data_source": "Mono3DRefer", - "scene_name": "006703", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, tram, person_sitting, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[952.2, 412.38], [1206.25, 410.8], [3361.43, 1208.26], [2264.06, 1240.79], [952.2, 134.46], [1206.25, 134.79], [3361.43, -27.73], [2264.06, -34.35]]}, {\"category\": \"tram\", \"corners_3d\": [[576.42, 195.44], [590.56, 196.21], [387.99, 197.81], [384.15, 196.87], [576.42, 141.97], [590.56, 139.78], [387.99, 135.23], [384.15, 137.92]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[328.81, 357.88], [408.01, 357.89], [382.15, 381.52], [292.35, 381.5], [328.81, 208.5], [408.01, 208.5], [382.15, 212.12], [292.35, 212.12]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[259.99, 356.06], [336.62, 356.07], [300.82, 379.71], [213.82, 379.69], [259.99, 209.52], [336.62, 209.53], [300.82, 213.32], [213.82, 213.32]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[354.18, 346.94], [291.97, 346.52], [320.52, 332.21], [377.39, 332.57], [354.18, 209.13], [291.97, 209.06], [320.52, 206.67], [377.39, 206.73]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[361.04, 310.87], [407.31, 311.04], [389.75, 321.97], [339.6, 321.77], [361.04, 201.7], [407.31, 201.72], [389.75, 203.43], [339.6, 203.4]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[415.99, 300.91], [369.79, 300.84], [384.63, 293.37], [427.95, 293.44], [415.99, 194.4], [369.79, 194.39], [384.63, 193.59], [427.95, 193.59]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[52.53, 362.13], [131.93, 358.79], [100.23, 375.77], [12.42, 379.78], [52.53, 157.76], [131.93, 158.2], [100.23, 155.97], [12.42, 155.44]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[105.38, 356.59], [45.83, 356.16], [87.0, 343.66], [142.38, 344.04], [105.38, 158.49], [45.83, 158.55], [87.0, 160.19], [142.38, 160.14]]}]\n```" - } - ] - }, - { - "id": 3639, - "data_source": "Mono3DRefer", - "scene_name": "006705", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[518.46, 253.91], [444.98, 254.08], [475.04, 237.76], [533.72, 237.65], [518.46, 191.34], [444.98, 191.38], [475.04, 187.66], [533.72, 187.64]]}, {\"category\": \"car\", \"corners_3d\": [[586.33, 188.27], [564.47, 188.28], [566.52, 187.22], [586.89, 187.22], [586.33, 165.26], [564.47, 165.25], [566.52, 165.77], [586.89, 165.78]]}]\n```" - } - ] - }, - { - "id": 3640, - "data_source": "Mono3DRefer", - "scene_name": "006727", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[428.44, 338.3], [289.95, 337.69], [393.8, 286.35], [489.32, 286.64], [428.44, 205.25], [289.95, 205.13], [393.8, 195.08], [489.32, 195.14]]}, {\"category\": \"car\", \"corners_3d\": [[573.09, 222.03], [541.19, 222.03], [547.88, 217.18], [576.63, 217.18], [573.09, 194.75], [541.19, 194.75], [547.88, 192.59], [576.63, 192.59]]}]\n```" - } - ] - }, - { - "id": 3641, - "data_source": "Mono3DRefer", - "scene_name": "006734", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[-130.8, 615.03], [-284.43, 616.95], [72.89, 436.42], [163.31, 435.74], [-130.8, 175.15], [-284.43, 175.16], [72.89, 174.22], [163.31, 174.22]]}, {\"category\": \"cyclist\", \"corners_3d\": [[102.35, 543.03], [-37.87, 545.2], [178.02, 416.63], [269.18, 415.69], [102.35, 178.99], [-37.87, 179.02], [178.02, 176.89], [269.18, 176.88]]}, {\"category\": \"cyclist\", \"corners_3d\": [[475.56, 251.66], [447.49, 251.57], [466.03, 243.6], [491.28, 243.67], [475.56, 175.41], [447.49, 175.41], [466.03, 175.15], [491.28, 175.15]]}]\n```" - } - ] - }, - { - "id": 3642, - "data_source": "Mono3DRefer", - "scene_name": "006734", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[352.23, 276.9], [261.21, 276.91], [355.42, 248.72], [421.78, 248.71], [352.23, 197.35], [261.21, 197.35], [355.42, 190.72], [421.78, 190.72]]}, {\"category\": \"car\", \"corners_3d\": [[507.78, 216.35], [469.62, 216.14], [494.19, 210.94], [527.85, 211.1], [507.78, 184.34], [469.62, 184.29], [494.19, 182.91], [527.85, 182.96]]}, {\"category\": \"car\", \"corners_3d\": [[646.22, 203.93], [680.83, 204.05], [682.36, 207.76], [643.64, 207.61], [646.22, 175.66], [680.83, 175.67], [682.36, 176.0], [643.64, 175.99]]}]\n```" - } - ] - }, - { - "id": 3643, - "data_source": "Mono3DRefer", - "scene_name": "006734", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[-130.8, 615.03], [-284.43, 616.95], [72.89, 436.42], [163.31, 435.74], [-130.8, 175.15], [-284.43, 175.16], [72.89, 174.22], [163.31, 174.22]]}, {\"category\": \"cyclist\", \"corners_3d\": [[102.35, 543.03], [-37.87, 545.2], [178.02, 416.63], [269.18, 415.69], [102.35, 178.99], [-37.87, 179.02], [178.02, 176.89], [269.18, 176.88]]}, {\"category\": \"car\", \"corners_3d\": [[352.23, 276.9], [261.21, 276.91], [355.42, 248.72], [421.78, 248.71], [352.23, 197.35], [261.21, 197.35], [355.42, 190.72], [421.78, 190.72]]}, {\"category\": \"cyclist\", \"corners_3d\": [[475.56, 251.66], [447.49, 251.57], [466.03, 243.6], [491.28, 243.67], [475.56, 175.41], [447.49, 175.41], [466.03, 175.15], [491.28, 175.15]]}, {\"category\": \"car\", \"corners_3d\": [[507.78, 216.35], [469.62, 216.14], [494.19, 210.94], [527.85, 211.1], [507.78, 184.34], [469.62, 184.29], [494.19, 182.91], [527.85, 182.96]]}, {\"category\": \"car\", \"corners_3d\": [[646.22, 203.93], [680.83, 204.05], [682.36, 207.76], [643.64, 207.61], [646.22, 175.66], [680.83, 175.67], [682.36, 176.0], [643.64, 175.99]]}]\n```" - } - ] - }, - { - "id": 3644, - "data_source": "Mono3DRefer", - "scene_name": "006739", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[486.41, 239.86], [427.4, 240.67], [439.01, 229.8], [488.56, 229.23], [486.41, 180.6], [427.4, 180.69], [439.01, 179.44], [488.56, 179.37]]}, {\"category\": \"car\", \"corners_3d\": [[485.82, 216.32], [449.9, 216.95], [442.92, 212.16], [475.07, 211.65], [485.82, 185.89], [449.9, 186.08], [442.92, 184.64], [475.07, 184.49]]}]\n```" - } - ] - }, - { - "id": 3645, - "data_source": "Mono3DRefer", - "scene_name": "006740", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[591.4, 231.99], [644.36, 231.34], [667.87, 240.94], [606.41, 241.82], [591.4, 179.67], [644.36, 179.59], [667.87, 180.7], [606.41, 180.8]]}, {\"category\": \"car\", \"corners_3d\": [[330.01, 219.57], [295.62, 219.55], [324.46, 215.35], [355.77, 215.37], [330.01, 185.05], [295.62, 185.04], [324.46, 183.94], [355.77, 183.95]]}, {\"category\": \"car\", \"corners_3d\": [[416.67, 221.23], [376.18, 221.34], [400.48, 215.56], [436.11, 215.48], [416.67, 185.3], [376.18, 185.33], [400.48, 183.84], [436.11, 183.82]]}]\n```" - } - ] - }, - { - "id": 3646, - "data_source": "Mono3DRefer", - "scene_name": "006748", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[385.38, 219.88], [353.73, 220.69], [349.08, 217.37], [378.64, 216.66], [385.38, 191.66], [353.73, 191.98], [349.08, 190.65], [378.64, 190.37]]}, {\"category\": \"car\", \"corners_3d\": [[426.57, 309.31], [294.86, 311.77], [370.97, 268.03], [460.79, 266.87], [426.57, 199.68], [294.86, 200.16], [370.97, 191.56], [460.79, 191.33]]}]\n```" - } - ] - }, - { - "id": 3647, - "data_source": "Mono3DRefer", - "scene_name": "006750", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[523.16, 249.67], [513.03, 241.51], [712.83, 237.56], [747.03, 244.65], [523.16, 156.88], [513.03, 159.67], [712.83, 161.02], [747.03, 158.6]]}, {\"category\": \"car\", \"corners_3d\": [[198.14, 249.32], [226.87, 241.31], [468.66, 237.13], [471.56, 244.01], [198.14, 144.89], [226.87, 149.04], [468.66, 151.21], [471.56, 147.64]]}, {\"category\": \"car\", \"corners_3d\": [[-265.88, 264.04], [-162.8, 252.05], [86.5, 247.66], [29.56, 258.11], [-265.88, 148.31], [-162.8, 152.93], [86.5, 154.62], [29.56, 150.59]]}]\n```" - } - ] - }, - { - "id": 3648, - "data_source": "Mono3DRefer", - "scene_name": "006750", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[832.23, 273.6], [807.7, 269.21], [896.02, 264.02], [923.16, 267.9], [832.23, 159.26], [807.7, 160.26], [896.02, 161.44], [923.16, 160.56]]}]\n```" - } - ] - }, - { - "id": 3649, - "data_source": "Mono3DRefer", - "scene_name": "006750", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[523.16, 249.67], [513.03, 241.51], [712.83, 237.56], [747.03, 244.65], [523.16, 156.88], [513.03, 159.67], [712.83, 161.02], [747.03, 158.6]]}, {\"category\": \"car\", \"corners_3d\": [[198.14, 249.32], [226.87, 241.31], [468.66, 237.13], [471.56, 244.01], [198.14, 144.89], [226.87, 149.04], [468.66, 151.21], [471.56, 147.64]]}, {\"category\": \"car\", \"corners_3d\": [[-265.88, 264.04], [-162.8, 252.05], [86.5, 247.66], [29.56, 258.11], [-265.88, 148.31], [-162.8, 152.93], [86.5, 154.62], [29.56, 150.59]]}, {\"category\": \"cyclist\", \"corners_3d\": [[832.23, 273.6], [807.7, 269.21], [896.02, 264.02], [923.16, 267.9], [832.23, 159.26], [807.7, 160.26], [896.02, 161.44], [923.16, 160.56]]}]\n```" - } - ] - }, - { - "id": 3650, - "data_source": "Mono3DRefer", - "scene_name": "006757", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1019.48, 194.13], [1062.11, 194.15], [1144.56, 198.09], [1093.97, 198.07], [1019.48, 153.82], [1062.11, 153.81], [1144.56, 150.28], [1093.97, 150.3]]}, {\"category\": \"van\", \"corners_3d\": [[1237.47, 204.35], [1314.61, 204.42], [1554.74, 215.4], [1450.13, 215.28], [1237.47, 135.38], [1314.61, 135.3], [1554.74, 122.23], [1450.13, 122.38]]}]\n```" - } - ] - }, - { - "id": 3651, - "data_source": "Mono3DRefer", - "scene_name": "006757", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[571.95, 220.59], [618.02, 220.56], [620.14, 227.25], [567.61, 227.29], [571.95, 175.9], [618.02, 175.9], [620.14, 176.32], [567.61, 176.32]]}, {\"category\": \"car\", \"corners_3d\": [[506.45, 210.54], [476.55, 210.54], [487.38, 207.46], [514.83, 207.46], [506.45, 179.72], [476.55, 179.72], [487.38, 179.16], [514.83, 179.16]]}, {\"category\": \"car\", \"corners_3d\": [[550.64, 193.75], [532.92, 193.75], [537.52, 192.49], [554.17, 192.49], [550.64, 177.94], [532.92, 177.94], [537.52, 177.64], [554.17, 177.64]]}, {\"category\": \"car\", \"corners_3d\": [[957.91, 192.97], [986.64, 192.96], [1036.35, 195.52], [1004.0, 195.53], [957.91, 166.65], [986.64, 166.66], [1036.35, 165.87], [1004.0, 165.87]]}]\n```" - } - ] - }, - { - "id": 3652, - "data_source": "Mono3DRefer", - "scene_name": "006757", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[571.95, 220.59], [618.02, 220.56], [620.14, 227.25], [567.61, 227.29], [571.95, 175.9], [618.02, 175.9], [620.14, 176.32], [567.61, 176.32]]}, {\"category\": \"car\", \"corners_3d\": [[506.45, 210.54], [476.55, 210.54], [487.38, 207.46], [514.83, 207.46], [506.45, 179.72], [476.55, 179.72], [487.38, 179.16], [514.83, 179.16]]}, {\"category\": \"car\", \"corners_3d\": [[550.64, 193.75], [532.92, 193.75], [537.52, 192.49], [554.17, 192.49], [550.64, 177.94], [532.92, 177.94], [537.52, 177.64], [554.17, 177.64]]}, {\"category\": \"car\", \"corners_3d\": [[957.91, 192.97], [986.64, 192.96], [1036.35, 195.52], [1004.0, 195.53], [957.91, 166.65], [986.64, 166.66], [1036.35, 165.87], [1004.0, 165.87]]}, {\"category\": \"van\", \"corners_3d\": [[1019.48, 194.13], [1062.11, 194.15], [1144.56, 198.09], [1093.97, 198.07], [1019.48, 153.82], [1062.11, 153.81], [1144.56, 150.28], [1093.97, 150.3]]}, {\"category\": \"van\", \"corners_3d\": [[1237.47, 204.35], [1314.61, 204.42], [1554.74, 215.4], [1450.13, 215.28], [1237.47, 135.38], [1314.61, 135.3], [1554.74, 122.23], [1450.13, 122.38]]}]\n```" - } - ] - }, - { - "id": 3653, - "data_source": "Mono3DRefer", - "scene_name": "006763", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[444.99, 276.3], [344.66, 277.58], [393.63, 251.84], [469.07, 251.11], [444.99, 175.82], [344.66, 175.86], [393.63, 175.12], [469.07, 175.1]]}, {\"category\": \"car\", \"corners_3d\": [[482.31, 219.89], [434.46, 220.66], [433.09, 213.93], [474.31, 213.36], [482.31, 179.68], [434.46, 179.79], [433.09, 178.82], [474.31, 178.73]]}]\n```" - } - ] - }, - { - "id": 3654, - "data_source": "Mono3DRefer", - "scene_name": "006776", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[316.45, 234.75], [300.89, 234.67], [321.21, 231.1], [335.9, 231.17], [316.45, 187.36], [300.89, 187.34], [321.21, 186.5], [335.9, 186.52]]}]\n```" - } - ] - }, - { - "id": 3655, - "data_source": "Mono3DRefer", - "scene_name": "006776", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[510.56, 243.3], [452.86, 243.19], [482.29, 231.05], [530.06, 231.12], [510.56, 188.37], [452.86, 188.35], [482.29, 185.67], [530.06, 185.69]]}]\n```" - } - ] - }, - { - "id": 3656, - "data_source": "Mono3DRefer", - "scene_name": "006776", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[510.56, 243.3], [452.86, 243.19], [482.29, 231.05], [530.06, 231.12], [510.56, 188.37], [452.86, 188.35], [482.29, 185.67], [530.06, 185.69]]}, {\"category\": \"cyclist\", \"corners_3d\": [[316.45, 234.75], [300.89, 234.67], [321.21, 231.1], [335.9, 231.17], [316.45, 187.36], [300.89, 187.34], [321.21, 186.5], [335.9, 186.52]]}]\n```" - } - ] - }, - { - "id": 3657, - "data_source": "Mono3DRefer", - "scene_name": "006785", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[594.2, 194.92], [619.09, 194.9], [620.99, 196.7], [594.08, 196.71], [594.2, 170.31], [619.09, 170.32], [620.99, 170.11], [594.08, 170.11]]}, {\"category\": \"car\", \"corners_3d\": [[368.89, 292.7], [474.11, 292.71], [369.51, 384.88], [183.42, 384.83], [368.89, 180.5], [474.11, 180.5], [369.51, 186.37], [183.42, 186.37]]}]\n```" - } - ] - }, - { - "id": 3658, - "data_source": "Mono3DRefer", - "scene_name": "006785", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[705.09, 231.9], [714.27, 230.51], [775.69, 233.51], [767.58, 235.06], [705.09, 157.53], [714.27, 157.89], [775.69, 157.11], [767.58, 156.71]]}, {\"category\": \"cyclist\", \"corners_3d\": [[710.95, 244.26], [725.49, 242.65], [798.18, 248.88], [784.14, 250.8], [710.95, 156.51], [725.49, 156.88], [798.18, 155.45], [784.14, 155.01]]}]\n```" - } - ] - }, - { - "id": 3659, - "data_source": "Mono3DRefer", - "scene_name": "006785", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[720.32, 239.62], [718.29, 237.94], [737.01, 238.01], [739.52, 239.69], [720.32, 163.3], [718.29, 163.54], [737.01, 163.53], [739.52, 163.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[698.13, 234.43], [687.33, 233.22], [698.69, 232.57], [709.6, 233.75], [698.13, 170.27], [687.33, 170.32], [698.69, 170.35], [709.6, 170.3]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[351.06, 247.12], [344.74, 249.04], [329.88, 249.07], [336.57, 247.15], [351.06, 170.44], [344.74, 170.38], [329.88, 170.38], [336.57, 170.44]]}]\n```" - } - ] - }, - { - "id": 3660, - "data_source": "Mono3DRefer", - "scene_name": "006785", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[705.09, 231.9], [714.27, 230.51], [775.69, 233.51], [767.58, 235.06], [705.09, 157.53], [714.27, 157.89], [775.69, 157.11], [767.58, 156.71]]}, {\"category\": \"car\", \"corners_3d\": [[594.2, 194.92], [619.09, 194.9], [620.99, 196.7], [594.08, 196.71], [594.2, 170.31], [619.09, 170.32], [620.99, 170.11], [594.08, 170.11]]}, {\"category\": \"car\", \"corners_3d\": [[368.89, 292.7], [474.11, 292.71], [369.51, 384.88], [183.42, 384.83], [368.89, 180.5], [474.11, 180.5], [369.51, 186.37], [183.42, 186.37]]}, {\"category\": \"cyclist\", \"corners_3d\": [[710.95, 244.26], [725.49, 242.65], [798.18, 248.88], [784.14, 250.8], [710.95, 156.51], [725.49, 156.88], [798.18, 155.45], [784.14, 155.01]]}]\n```" - } - ] - }, - { - "id": 3661, - "data_source": "Mono3DRefer", - "scene_name": "006785", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[720.32, 239.62], [718.29, 237.94], [737.01, 238.01], [739.52, 239.69], [720.32, 163.3], [718.29, 163.54], [737.01, 163.53], [739.52, 163.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[698.13, 234.43], [687.33, 233.22], [698.69, 232.57], [709.6, 233.75], [698.13, 170.27], [687.33, 170.32], [698.69, 170.35], [709.6, 170.3]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[351.06, 247.12], [344.74, 249.04], [329.88, 249.07], [336.57, 247.15], [351.06, 170.44], [344.74, 170.38], [329.88, 170.38], [336.57, 170.44]]}, {\"category\": \"car\", \"corners_3d\": [[594.2, 194.92], [619.09, 194.9], [620.99, 196.7], [594.08, 196.71], [594.2, 170.31], [619.09, 170.32], [620.99, 170.11], [594.08, 170.11]]}, {\"category\": \"car\", \"corners_3d\": [[368.89, 292.7], [474.11, 292.71], [369.51, 384.88], [183.42, 384.83], [368.89, 180.5], [474.11, 180.5], [369.51, 186.37], [183.42, 186.37]]}]\n```" - } - ] - }, - { - "id": 3662, - "data_source": "Mono3DRefer", - "scene_name": "006785", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[720.32, 239.62], [718.29, 237.94], [737.01, 238.01], [739.52, 239.69], [720.32, 163.3], [718.29, 163.54], [737.01, 163.53], [739.52, 163.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[698.13, 234.43], [687.33, 233.22], [698.69, 232.57], [709.6, 233.75], [698.13, 170.27], [687.33, 170.32], [698.69, 170.35], [709.6, 170.3]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[351.06, 247.12], [344.74, 249.04], [329.88, 249.07], [336.57, 247.15], [351.06, 170.44], [344.74, 170.38], [329.88, 170.38], [336.57, 170.44]]}, {\"category\": \"cyclist\", \"corners_3d\": [[705.09, 231.9], [714.27, 230.51], [775.69, 233.51], [767.58, 235.06], [705.09, 157.53], [714.27, 157.89], [775.69, 157.11], [767.58, 156.71]]}, {\"category\": \"cyclist\", \"corners_3d\": [[710.95, 244.26], [725.49, 242.65], [798.18, 248.88], [784.14, 250.8], [710.95, 156.51], [725.49, 156.88], [798.18, 155.45], [784.14, 155.01]]}]\n```" - } - ] - }, - { - "id": 3663, - "data_source": "Mono3DRefer", - "scene_name": "006785", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[720.32, 239.62], [718.29, 237.94], [737.01, 238.01], [739.52, 239.69], [720.32, 163.3], [718.29, 163.54], [737.01, 163.53], [739.52, 163.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[698.13, 234.43], [687.33, 233.22], [698.69, 232.57], [709.6, 233.75], [698.13, 170.27], [687.33, 170.32], [698.69, 170.35], [709.6, 170.3]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[351.06, 247.12], [344.74, 249.04], [329.88, 249.07], [336.57, 247.15], [351.06, 170.44], [344.74, 170.38], [329.88, 170.38], [336.57, 170.44]]}, {\"category\": \"cyclist\", \"corners_3d\": [[705.09, 231.9], [714.27, 230.51], [775.69, 233.51], [767.58, 235.06], [705.09, 157.53], [714.27, 157.89], [775.69, 157.11], [767.58, 156.71]]}, {\"category\": \"car\", \"corners_3d\": [[594.2, 194.92], [619.09, 194.9], [620.99, 196.7], [594.08, 196.71], [594.2, 170.31], [619.09, 170.32], [620.99, 170.11], [594.08, 170.11]]}, {\"category\": \"car\", \"corners_3d\": [[368.89, 292.7], [474.11, 292.71], [369.51, 384.88], [183.42, 384.83], [368.89, 180.5], [474.11, 180.5], [369.51, 186.37], [183.42, 186.37]]}, {\"category\": \"cyclist\", \"corners_3d\": [[710.95, 244.26], [725.49, 242.65], [798.18, 248.88], [784.14, 250.8], [710.95, 156.51], [725.49, 156.88], [798.18, 155.45], [784.14, 155.01]]}]\n```" - } - ] - }, - { - "id": 3664, - "data_source": "Mono3DRefer", - "scene_name": "006790", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[537.14, 226.02], [487.42, 225.99], [504.94, 218.76], [547.9, 218.79], [537.14, 185.22], [487.42, 185.22], [504.94, 183.53], [547.9, 183.54]]}]\n```" - } - ] - }, - { - "id": 3665, - "data_source": "Mono3DRefer", - "scene_name": "006802", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[683.98, 200.63], [683.97, 200.26], [714.76, 200.38], [715.19, 200.75], [683.98, 169.2], [683.97, 169.25], [714.76, 169.23], [715.19, 169.19]]}]\n```" - } - ] - }, - { - "id": 3666, - "data_source": "Mono3DRefer", - "scene_name": "006802", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[487.57, 230.64], [416.56, 230.59], [466.19, 216.23], [519.57, 216.26], [487.57, 143.24], [416.56, 143.26], [466.19, 150.62], [519.57, 150.61]]}]\n```" - } - ] - }, - { - "id": 3667, - "data_source": "Mono3DRefer", - "scene_name": "006802", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[537.88, 212.27], [496.06, 211.92], [525.63, 206.71], [562.07, 206.98], [537.88, 178.93], [496.06, 178.88], [525.63, 178.07], [562.07, 178.11]]}]\n```" - } - ] - }, - { - "id": 3668, - "data_source": "Mono3DRefer", - "scene_name": "006802", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[683.98, 200.63], [683.97, 200.26], [714.76, 200.38], [715.19, 200.75], [683.98, 169.2], [683.97, 169.25], [714.76, 169.23], [715.19, 169.19]]}, {\"category\": \"van\", \"corners_3d\": [[487.57, 230.64], [416.56, 230.59], [466.19, 216.23], [519.57, 216.26], [487.57, 143.24], [416.56, 143.26], [466.19, 150.62], [519.57, 150.61]]}]\n```" - } - ] - }, - { - "id": 3669, - "data_source": "Mono3DRefer", - "scene_name": "006802", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[537.88, 212.27], [496.06, 211.92], [525.63, 206.71], [562.07, 206.98], [537.88, 178.93], [496.06, 178.88], [525.63, 178.07], [562.07, 178.11]]}, {\"category\": \"cyclist\", \"corners_3d\": [[683.98, 200.63], [683.97, 200.26], [714.76, 200.38], [715.19, 200.75], [683.98, 169.2], [683.97, 169.25], [714.76, 169.23], [715.19, 169.19]]}]\n```" - } - ] - }, - { - "id": 3670, - "data_source": "Mono3DRefer", - "scene_name": "006802", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[537.88, 212.27], [496.06, 211.92], [525.63, 206.71], [562.07, 206.98], [537.88, 178.93], [496.06, 178.88], [525.63, 178.07], [562.07, 178.11]]}, {\"category\": \"van\", \"corners_3d\": [[487.57, 230.64], [416.56, 230.59], [466.19, 216.23], [519.57, 216.26], [487.57, 143.24], [416.56, 143.26], [466.19, 150.62], [519.57, 150.61]]}]\n```" - } - ] - }, - { - "id": 3671, - "data_source": "Mono3DRefer", - "scene_name": "006802", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[537.88, 212.27], [496.06, 211.92], [525.63, 206.71], [562.07, 206.98], [537.88, 178.93], [496.06, 178.88], [525.63, 178.07], [562.07, 178.11]]}, {\"category\": \"cyclist\", \"corners_3d\": [[683.98, 200.63], [683.97, 200.26], [714.76, 200.38], [715.19, 200.75], [683.98, 169.2], [683.97, 169.25], [714.76, 169.23], [715.19, 169.19]]}, {\"category\": \"van\", \"corners_3d\": [[487.57, 230.64], [416.56, 230.59], [466.19, 216.23], [519.57, 216.26], [487.57, 143.24], [416.56, 143.26], [466.19, 150.62], [519.57, 150.61]]}]\n```" - } - ] - }, - { - "id": 3672, - "data_source": "Mono3DRefer", - "scene_name": "006808", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[784.24, 281.08], [877.58, 280.95], [1021.06, 337.42], [879.14, 337.72], [784.24, 184.61], [877.58, 184.6], [1021.06, 190.73], [879.14, 190.77]]}, {\"category\": \"car\", \"corners_3d\": [[563.72, 284.14], [482.5, 283.9], [513.1, 259.78], [576.71, 259.93], [563.72, 208.89], [482.5, 208.81], [513.1, 201.0], [576.71, 201.05]]}, {\"category\": \"car\", \"corners_3d\": [[574.83, 225.49], [606.96, 225.45], [608.5, 232.67], [571.96, 232.73], [574.83, 194.4], [606.96, 194.38], [608.5, 197.34], [571.96, 197.37]]}, {\"category\": \"car\", \"corners_3d\": [[685.58, 224.16], [720.34, 224.19], [732.69, 230.34], [693.76, 230.31], [685.58, 185.89], [720.34, 185.89], [732.69, 187.46], [693.76, 187.45]]}]\n```" - } - ] - }, - { - "id": 3673, - "data_source": "Mono3DRefer", - "scene_name": "006817", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[63.77, 513.23], [-180.93, 506.07], [210.9, 349.38], [343.66, 351.37], [63.77, 227.41], [-180.93, 226.26], [210.9, 201.14], [343.66, 201.46]]}, {\"category\": \"car\", \"corners_3d\": [[443.68, 314.46], [338.07, 311.69], [432.7, 275.72], [511.94, 277.23], [443.68, 205.46], [338.07, 204.82], [432.7, 196.54], [511.94, 196.88]]}, {\"category\": \"car\", \"corners_3d\": [[717.9, 201.61], [748.19, 202.03], [733.29, 205.49], [699.69, 204.97], [717.9, 171.13], [748.19, 171.1], [733.29, 170.89], [699.69, 170.92]]}, {\"category\": \"car\", \"corners_3d\": [[461.13, 227.6], [483.33, 226.15], [520.65, 229.3], [498.18, 230.93], [461.13, 193.97], [483.33, 193.41], [520.65, 194.62], [498.18, 195.25]]}, {\"category\": \"car\", \"corners_3d\": [[1078.26, 212.93], [1098.4, 215.85], [953.9, 213.91], [944.0, 211.24], [1078.26, 166.94], [1098.4, 166.51], [953.9, 166.79], [944.0, 167.19]]}, {\"category\": \"car\", \"corners_3d\": [[1015.01, 205.57], [1024.19, 207.47], [886.22, 205.58], [884.69, 203.88], [1015.01, 165.14], [1024.19, 164.7], [886.22, 165.14], [884.69, 165.54]]}, {\"category\": \"car\", \"corners_3d\": [[767.46, 193.28], [741.74, 193.04], [750.56, 191.59], [774.5, 191.79], [767.46, 168.88], [741.74, 168.93], [750.56, 169.21], [774.5, 169.17]]}]\n```" - } - ] - }, - { - "id": 3674, - "data_source": "Mono3DRefer", - "scene_name": "006819", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[874.57, 142.75], [846.73, 143.29], [859.4, 145.55], [885.28, 145.09], [874.57, 117.8], [846.73, 118.79], [859.4, 122.92], [885.28, 122.08]]}]\n```" - } - ] - }, - { - "id": 3675, - "data_source": "Mono3DRefer", - "scene_name": "006819", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (motorcyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"motorcyclist\", \"corners_3d\": [[632.19, 283.34], [577.96, 280.72], [615.45, 265.29], [662.54, 267.21], [632.19, 144.78], [577.96, 145.45], [615.45, 149.37], [662.54, 148.88]]}]\n```" - } - ] - }, - { - "id": 3676, - "data_source": "Mono3DRefer", - "scene_name": "006819", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, motorcyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"motorcyclist\", \"corners_3d\": [[632.19, 283.34], [577.96, 280.72], [615.45, 265.29], [662.54, 267.21], [632.19, 144.78], [577.96, 145.45], [615.45, 149.37], [662.54, 148.88]]}, {\"category\": \"car\", \"corners_3d\": [[874.57, 142.75], [846.73, 143.29], [859.4, 145.55], [885.28, 145.09], [874.57, 117.8], [846.73, 118.79], [859.4, 122.92], [885.28, 122.08]]}]\n```" - } - ] - }, - { - "id": 3677, - "data_source": "Mono3DRefer", - "scene_name": "006823", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[314.84, 193.17], [293.79, 193.2], [312.96, 191.83], [332.59, 191.81], [314.84, 174.0], [293.79, 174.0], [312.96, 173.92], [332.59, 173.92]]}, {\"category\": \"car\", \"corners_3d\": [[572.02, 219.13], [614.92, 219.05], [618.76, 225.77], [569.62, 225.88], [572.02, 177.76], [614.92, 177.75], [618.76, 178.46], [569.62, 178.47]]}, {\"category\": \"car\", \"corners_3d\": [[280.89, 195.02], [255.85, 195.03], [280.66, 193.44], [303.9, 193.44], [280.89, 174.43], [255.85, 174.43], [280.66, 174.32], [303.9, 174.32]]}]\n```" - } - ] - }, - { - "id": 3678, - "data_source": "Mono3DRefer", - "scene_name": "006833", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[164.97, 362.73], [6.27, 429.43], [-557.89, 427.41], [-254.32, 361.62], [164.97, 183.36], [6.27, 187.05], [-557.89, 186.94], [-254.32, 183.3]]}, {\"category\": \"car\", \"corners_3d\": [[241.69, 318.51], [149.9, 351.61], [-157.58, 347.67], [-11.54, 315.89], [241.69, 177.08], [149.9, 178.04], [-157.58, 177.93], [-11.54, 177.01]]}, {\"category\": \"car\", \"corners_3d\": [[348.85, 283.68], [294.84, 304.35], [-34.49, 301.9], [69.62, 281.94], [348.85, 173.49], [294.84, 173.61], [-34.49, 173.6], [69.62, 173.48]]}, {\"category\": \"car\", \"corners_3d\": [[100.45, 270.77], [162.63, 259.0], [336.33, 259.24], [297.75, 271.08], [100.45, 174.66], [162.63, 174.44], [336.33, 174.44], [297.75, 174.66]]}, {\"category\": \"car\", \"corners_3d\": [[197.94, 264.19], [236.65, 254.56], [410.47, 253.38], [392.55, 262.72], [197.94, 186.69], [236.65, 185.23], [410.47, 185.06], [392.55, 186.47]]}, {\"category\": \"car\", \"corners_3d\": [[226.87, 242.86], [261.52, 236.88], [431.77, 237.36], [412.88, 243.44], [226.87, 175.85], [261.52, 175.59], [431.77, 175.61], [412.88, 175.87]]}, {\"category\": \"car\", \"corners_3d\": [[319.35, 227.01], [336.93, 223.02], [461.63, 222.41], [454.08, 226.3], [319.35, 170.44], [336.93, 170.62], [461.63, 170.65], [454.08, 170.47]]}, {\"category\": \"car\", \"corners_3d\": [[480.85, 220.33], [472.18, 223.93], [338.31, 224.12], [356.48, 220.49], [480.85, 161.54], [472.18, 160.68], [338.31, 160.63], [356.48, 161.5]]}, {\"category\": \"car\", \"corners_3d\": [[770.63, 297.25], [884.13, 297.46], [1031.59, 366.33], [855.13, 365.82], [770.63, 192.55], [884.13, 192.59], [1031.59, 203.49], [855.13, 203.41]]}, {\"category\": \"car\", \"corners_3d\": [[1055.89, 264.41], [1110.08, 276.72], [866.04, 274.7], [841.21, 262.83], [1055.89, 174.98], [1110.08, 175.26], [866.04, 175.22], [841.21, 174.94]]}, {\"category\": \"car\", \"corners_3d\": [[964.44, 249.42], [1003.75, 258.86], [830.32, 257.85], [810.26, 248.62], [964.44, 179.64], [1003.75, 180.47], [830.32, 180.38], [810.26, 179.56]]}, {\"category\": \"car\", \"corners_3d\": [[906.1, 237.07], [930.97, 242.73], [790.92, 242.47], [777.45, 236.85], [906.1, 180.15], [930.97, 180.79], [790.92, 180.76], [777.45, 180.12]]}, {\"category\": \"car\", \"corners_3d\": [[743.69, 229.34], [731.66, 225.38], [849.3, 224.93], [870.02, 228.82], [743.69, 178.36], [731.66, 177.98], [849.3, 177.93], [870.02, 178.31]]}, {\"category\": \"car\", \"corners_3d\": [[823.56, 217.75], [839.82, 220.67], [719.62, 221.08], [710.63, 218.11], [823.56, 179.26], [839.82, 179.68], [719.62, 179.73], [710.63, 179.31]]}, {\"category\": \"car\", \"corners_3d\": [[807.2, 214.83], [820.22, 217.41], [720.41, 217.53], [713.15, 214.94], [807.2, 171.12], [820.22, 171.01], [720.41, 171.0], [713.15, 171.11]]}, {\"category\": \"car\", \"corners_3d\": [[426.07, 204.42], [433.69, 203.05], [521.45, 203.01], [517.83, 204.38], [426.07, 173.67], [433.69, 173.64], [521.45, 173.64], [517.83, 173.67]]}, {\"category\": \"car\", \"corners_3d\": [[450.74, 204.96], [457.96, 203.64], [532.91, 203.7], [528.92, 205.04], [450.74, 176.67], [457.96, 176.51], [532.91, 176.52], [528.92, 176.68]]}, {\"category\": \"car\", \"corners_3d\": [[635.54, 201.14], [661.23, 201.09], [670.2, 203.99], [641.88, 204.05], [635.54, 177.81], [661.23, 177.8], [670.2, 178.31], [641.88, 178.32]]}]\n```" - } - ] - }, - { - "id": 3679, - "data_source": "Mono3DRefer", - "scene_name": "006833", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[614.97, 182.75], [628.35, 182.75], [629.47, 183.36], [615.26, 183.36], [614.97, 166.84], [628.35, 166.84], [629.47, 166.48], [615.26, 166.48]]}]\n```" - } - ] - }, - { - "id": 3680, - "data_source": "Mono3DRefer", - "scene_name": "006833", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[559.78, 205.78], [548.46, 205.8], [550.16, 204.36], [560.99, 204.34], [559.78, 174.8], [548.46, 174.8], [550.16, 174.71], [560.99, 174.71]]}]\n```" - } - ] - }, - { - "id": 3681, - "data_source": "Mono3DRefer", - "scene_name": "006833", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[164.97, 362.73], [6.27, 429.43], [-557.89, 427.41], [-254.32, 361.62], [164.97, 183.36], [6.27, 187.05], [-557.89, 186.94], [-254.32, 183.3]]}, {\"category\": \"car\", \"corners_3d\": [[241.69, 318.51], [149.9, 351.61], [-157.58, 347.67], [-11.54, 315.89], [241.69, 177.08], [149.9, 178.04], [-157.58, 177.93], [-11.54, 177.01]]}, {\"category\": \"car\", \"corners_3d\": [[348.85, 283.68], [294.84, 304.35], [-34.49, 301.9], [69.62, 281.94], [348.85, 173.49], [294.84, 173.61], [-34.49, 173.6], [69.62, 173.48]]}, {\"category\": \"car\", \"corners_3d\": [[100.45, 270.77], [162.63, 259.0], [336.33, 259.24], [297.75, 271.08], [100.45, 174.66], [162.63, 174.44], [336.33, 174.44], [297.75, 174.66]]}, {\"category\": \"car\", \"corners_3d\": [[197.94, 264.19], [236.65, 254.56], [410.47, 253.38], [392.55, 262.72], [197.94, 186.69], [236.65, 185.23], [410.47, 185.06], [392.55, 186.47]]}, {\"category\": \"car\", \"corners_3d\": [[226.87, 242.86], [261.52, 236.88], [431.77, 237.36], [412.88, 243.44], [226.87, 175.85], [261.52, 175.59], [431.77, 175.61], [412.88, 175.87]]}, {\"category\": \"car\", \"corners_3d\": [[319.35, 227.01], [336.93, 223.02], [461.63, 222.41], [454.08, 226.3], [319.35, 170.44], [336.93, 170.62], [461.63, 170.65], [454.08, 170.47]]}, {\"category\": \"car\", \"corners_3d\": [[480.85, 220.33], [472.18, 223.93], [338.31, 224.12], [356.48, 220.49], [480.85, 161.54], [472.18, 160.68], [338.31, 160.63], [356.48, 161.5]]}, {\"category\": \"car\", \"corners_3d\": [[770.63, 297.25], [884.13, 297.46], [1031.59, 366.33], [855.13, 365.82], [770.63, 192.55], [884.13, 192.59], [1031.59, 203.49], [855.13, 203.41]]}, {\"category\": \"car\", \"corners_3d\": [[1055.89, 264.41], [1110.08, 276.72], [866.04, 274.7], [841.21, 262.83], [1055.89, 174.98], [1110.08, 175.26], [866.04, 175.22], [841.21, 174.94]]}, {\"category\": \"car\", \"corners_3d\": [[964.44, 249.42], [1003.75, 258.86], [830.32, 257.85], [810.26, 248.62], [964.44, 179.64], [1003.75, 180.47], [830.32, 180.38], [810.26, 179.56]]}, {\"category\": \"car\", \"corners_3d\": [[906.1, 237.07], [930.97, 242.73], [790.92, 242.47], [777.45, 236.85], [906.1, 180.15], [930.97, 180.79], [790.92, 180.76], [777.45, 180.12]]}, {\"category\": \"car\", \"corners_3d\": [[743.69, 229.34], [731.66, 225.38], [849.3, 224.93], [870.02, 228.82], [743.69, 178.36], [731.66, 177.98], [849.3, 177.93], [870.02, 178.31]]}, {\"category\": \"car\", \"corners_3d\": [[823.56, 217.75], [839.82, 220.67], [719.62, 221.08], [710.63, 218.11], [823.56, 179.26], [839.82, 179.68], [719.62, 179.73], [710.63, 179.31]]}, {\"category\": \"car\", \"corners_3d\": [[807.2, 214.83], [820.22, 217.41], [720.41, 217.53], [713.15, 214.94], [807.2, 171.12], [820.22, 171.01], [720.41, 171.0], [713.15, 171.11]]}, {\"category\": \"car\", \"corners_3d\": [[426.07, 204.42], [433.69, 203.05], [521.45, 203.01], [517.83, 204.38], [426.07, 173.67], [433.69, 173.64], [521.45, 173.64], [517.83, 173.67]]}, {\"category\": \"car\", \"corners_3d\": [[450.74, 204.96], [457.96, 203.64], [532.91, 203.7], [528.92, 205.04], [450.74, 176.67], [457.96, 176.51], [532.91, 176.52], [528.92, 176.68]]}, {\"category\": \"car\", \"corners_3d\": [[635.54, 201.14], [661.23, 201.09], [670.2, 203.99], [641.88, 204.05], [635.54, 177.81], [661.23, 177.8], [670.2, 178.31], [641.88, 178.32]]}, {\"category\": \"van\", \"corners_3d\": [[614.97, 182.75], [628.35, 182.75], [629.47, 183.36], [615.26, 183.36], [614.97, 166.84], [628.35, 166.84], [629.47, 166.48], [615.26, 166.48]]}]\n```" - } - ] - }, - { - "id": 3682, - "data_source": "Mono3DRefer", - "scene_name": "006833", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[164.97, 362.73], [6.27, 429.43], [-557.89, 427.41], [-254.32, 361.62], [164.97, 183.36], [6.27, 187.05], [-557.89, 186.94], [-254.32, 183.3]]}, {\"category\": \"car\", \"corners_3d\": [[241.69, 318.51], [149.9, 351.61], [-157.58, 347.67], [-11.54, 315.89], [241.69, 177.08], [149.9, 178.04], [-157.58, 177.93], [-11.54, 177.01]]}, {\"category\": \"car\", \"corners_3d\": [[348.85, 283.68], [294.84, 304.35], [-34.49, 301.9], [69.62, 281.94], [348.85, 173.49], [294.84, 173.61], [-34.49, 173.6], [69.62, 173.48]]}, {\"category\": \"car\", \"corners_3d\": [[100.45, 270.77], [162.63, 259.0], [336.33, 259.24], [297.75, 271.08], [100.45, 174.66], [162.63, 174.44], [336.33, 174.44], [297.75, 174.66]]}, {\"category\": \"car\", \"corners_3d\": [[197.94, 264.19], [236.65, 254.56], [410.47, 253.38], [392.55, 262.72], [197.94, 186.69], [236.65, 185.23], [410.47, 185.06], [392.55, 186.47]]}, {\"category\": \"car\", \"corners_3d\": [[226.87, 242.86], [261.52, 236.88], [431.77, 237.36], [412.88, 243.44], [226.87, 175.85], [261.52, 175.59], [431.77, 175.61], [412.88, 175.87]]}, {\"category\": \"car\", \"corners_3d\": [[319.35, 227.01], [336.93, 223.02], [461.63, 222.41], [454.08, 226.3], [319.35, 170.44], [336.93, 170.62], [461.63, 170.65], [454.08, 170.47]]}, {\"category\": \"car\", \"corners_3d\": [[480.85, 220.33], [472.18, 223.93], [338.31, 224.12], [356.48, 220.49], [480.85, 161.54], [472.18, 160.68], [338.31, 160.63], [356.48, 161.5]]}, {\"category\": \"cyclist\", \"corners_3d\": [[559.78, 205.78], [548.46, 205.8], [550.16, 204.36], [560.99, 204.34], [559.78, 174.8], [548.46, 174.8], [550.16, 174.71], [560.99, 174.71]]}, {\"category\": \"car\", \"corners_3d\": [[770.63, 297.25], [884.13, 297.46], [1031.59, 366.33], [855.13, 365.82], [770.63, 192.55], [884.13, 192.59], [1031.59, 203.49], [855.13, 203.41]]}, {\"category\": \"car\", \"corners_3d\": [[1055.89, 264.41], [1110.08, 276.72], [866.04, 274.7], [841.21, 262.83], [1055.89, 174.98], [1110.08, 175.26], [866.04, 175.22], [841.21, 174.94]]}, {\"category\": \"car\", \"corners_3d\": [[964.44, 249.42], [1003.75, 258.86], [830.32, 257.85], [810.26, 248.62], [964.44, 179.64], [1003.75, 180.47], [830.32, 180.38], [810.26, 179.56]]}, {\"category\": \"car\", \"corners_3d\": [[906.1, 237.07], [930.97, 242.73], [790.92, 242.47], [777.45, 236.85], [906.1, 180.15], [930.97, 180.79], [790.92, 180.76], [777.45, 180.12]]}, {\"category\": \"car\", \"corners_3d\": [[743.69, 229.34], [731.66, 225.38], [849.3, 224.93], [870.02, 228.82], [743.69, 178.36], [731.66, 177.98], [849.3, 177.93], [870.02, 178.31]]}, {\"category\": \"car\", \"corners_3d\": [[823.56, 217.75], [839.82, 220.67], [719.62, 221.08], [710.63, 218.11], [823.56, 179.26], [839.82, 179.68], [719.62, 179.73], [710.63, 179.31]]}, {\"category\": \"car\", \"corners_3d\": [[807.2, 214.83], [820.22, 217.41], [720.41, 217.53], [713.15, 214.94], [807.2, 171.12], [820.22, 171.01], [720.41, 171.0], [713.15, 171.11]]}, {\"category\": \"car\", \"corners_3d\": [[426.07, 204.42], [433.69, 203.05], [521.45, 203.01], [517.83, 204.38], [426.07, 173.67], [433.69, 173.64], [521.45, 173.64], [517.83, 173.67]]}, {\"category\": \"car\", \"corners_3d\": [[450.74, 204.96], [457.96, 203.64], [532.91, 203.7], [528.92, 205.04], [450.74, 176.67], [457.96, 176.51], [532.91, 176.52], [528.92, 176.68]]}, {\"category\": \"car\", \"corners_3d\": [[635.54, 201.14], [661.23, 201.09], [670.2, 203.99], [641.88, 204.05], [635.54, 177.81], [661.23, 177.8], [670.2, 178.31], [641.88, 178.32]]}]\n```" - } - ] - }, - { - "id": 3683, - "data_source": "Mono3DRefer", - "scene_name": "006833", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[559.78, 205.78], [548.46, 205.8], [550.16, 204.36], [560.99, 204.34], [559.78, 174.8], [548.46, 174.8], [550.16, 174.71], [560.99, 174.71]]}, {\"category\": \"van\", \"corners_3d\": [[614.97, 182.75], [628.35, 182.75], [629.47, 183.36], [615.26, 183.36], [614.97, 166.84], [628.35, 166.84], [629.47, 166.48], [615.26, 166.48]]}]\n```" - } - ] - }, - { - "id": 3684, - "data_source": "Mono3DRefer", - "scene_name": "006833", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[164.97, 362.73], [6.27, 429.43], [-557.89, 427.41], [-254.32, 361.62], [164.97, 183.36], [6.27, 187.05], [-557.89, 186.94], [-254.32, 183.3]]}, {\"category\": \"car\", \"corners_3d\": [[241.69, 318.51], [149.9, 351.61], [-157.58, 347.67], [-11.54, 315.89], [241.69, 177.08], [149.9, 178.04], [-157.58, 177.93], [-11.54, 177.01]]}, {\"category\": \"car\", \"corners_3d\": [[348.85, 283.68], [294.84, 304.35], [-34.49, 301.9], [69.62, 281.94], [348.85, 173.49], [294.84, 173.61], [-34.49, 173.6], [69.62, 173.48]]}, {\"category\": \"car\", \"corners_3d\": [[100.45, 270.77], [162.63, 259.0], [336.33, 259.24], [297.75, 271.08], [100.45, 174.66], [162.63, 174.44], [336.33, 174.44], [297.75, 174.66]]}, {\"category\": \"car\", \"corners_3d\": [[197.94, 264.19], [236.65, 254.56], [410.47, 253.38], [392.55, 262.72], [197.94, 186.69], [236.65, 185.23], [410.47, 185.06], [392.55, 186.47]]}, {\"category\": \"car\", \"corners_3d\": [[226.87, 242.86], [261.52, 236.88], [431.77, 237.36], [412.88, 243.44], [226.87, 175.85], [261.52, 175.59], [431.77, 175.61], [412.88, 175.87]]}, {\"category\": \"car\", \"corners_3d\": [[319.35, 227.01], [336.93, 223.02], [461.63, 222.41], [454.08, 226.3], [319.35, 170.44], [336.93, 170.62], [461.63, 170.65], [454.08, 170.47]]}, {\"category\": \"car\", \"corners_3d\": [[480.85, 220.33], [472.18, 223.93], [338.31, 224.12], [356.48, 220.49], [480.85, 161.54], [472.18, 160.68], [338.31, 160.63], [356.48, 161.5]]}, {\"category\": \"cyclist\", \"corners_3d\": [[559.78, 205.78], [548.46, 205.8], [550.16, 204.36], [560.99, 204.34], [559.78, 174.8], [548.46, 174.8], [550.16, 174.71], [560.99, 174.71]]}, {\"category\": \"car\", \"corners_3d\": [[770.63, 297.25], [884.13, 297.46], [1031.59, 366.33], [855.13, 365.82], [770.63, 192.55], [884.13, 192.59], [1031.59, 203.49], [855.13, 203.41]]}, {\"category\": \"car\", \"corners_3d\": [[1055.89, 264.41], [1110.08, 276.72], [866.04, 274.7], [841.21, 262.83], [1055.89, 174.98], [1110.08, 175.26], [866.04, 175.22], [841.21, 174.94]]}, {\"category\": \"car\", \"corners_3d\": [[964.44, 249.42], [1003.75, 258.86], [830.32, 257.85], [810.26, 248.62], [964.44, 179.64], [1003.75, 180.47], [830.32, 180.38], [810.26, 179.56]]}, {\"category\": \"car\", \"corners_3d\": [[906.1, 237.07], [930.97, 242.73], [790.92, 242.47], [777.45, 236.85], [906.1, 180.15], [930.97, 180.79], [790.92, 180.76], [777.45, 180.12]]}, {\"category\": \"car\", \"corners_3d\": [[743.69, 229.34], [731.66, 225.38], [849.3, 224.93], [870.02, 228.82], [743.69, 178.36], [731.66, 177.98], [849.3, 177.93], [870.02, 178.31]]}, {\"category\": \"car\", \"corners_3d\": [[823.56, 217.75], [839.82, 220.67], [719.62, 221.08], [710.63, 218.11], [823.56, 179.26], [839.82, 179.68], [719.62, 179.73], [710.63, 179.31]]}, {\"category\": \"car\", \"corners_3d\": [[807.2, 214.83], [820.22, 217.41], [720.41, 217.53], [713.15, 214.94], [807.2, 171.12], [820.22, 171.01], [720.41, 171.0], [713.15, 171.11]]}, {\"category\": \"car\", \"corners_3d\": [[426.07, 204.42], [433.69, 203.05], [521.45, 203.01], [517.83, 204.38], [426.07, 173.67], [433.69, 173.64], [521.45, 173.64], [517.83, 173.67]]}, {\"category\": \"car\", \"corners_3d\": [[450.74, 204.96], [457.96, 203.64], [532.91, 203.7], [528.92, 205.04], [450.74, 176.67], [457.96, 176.51], [532.91, 176.52], [528.92, 176.68]]}, {\"category\": \"car\", \"corners_3d\": [[635.54, 201.14], [661.23, 201.09], [670.2, 203.99], [641.88, 204.05], [635.54, 177.81], [661.23, 177.8], [670.2, 178.31], [641.88, 178.32]]}, {\"category\": \"van\", \"corners_3d\": [[614.97, 182.75], [628.35, 182.75], [629.47, 183.36], [615.26, 183.36], [614.97, 166.84], [628.35, 166.84], [629.47, 166.48], [615.26, 166.48]]}]\n```" - } - ] - }, - { - "id": 3685, - "data_source": "Mono3DRefer", - "scene_name": "006838", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[280.48, 377.5], [19.16, 377.56], [330.6, 269.47], [453.91, 269.46], [280.48, 43.41], [19.16, 43.37], [330.6, 111.74], [453.91, 111.75]]}]\n```" - } - ] - }, - { - "id": 3686, - "data_source": "Mono3DRefer", - "scene_name": "006838", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[852.25, 220.09], [859.77, 220.05], [875.43, 222.51], [867.53, 222.55], [852.25, 182.84], [859.77, 182.83], [875.43, 183.35], [867.53, 183.36]]}]\n```" - } - ] - }, - { - "id": 3687, - "data_source": "Mono3DRefer", - "scene_name": "006838", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[555.21, 198.79], [538.9, 198.8], [541.17, 197.59], [556.71, 197.57], [555.21, 182.06], [538.9, 182.07], [541.17, 181.64], [556.71, 181.63]]}, {\"category\": \"car\", \"corners_3d\": [[953.84, 221.44], [941.75, 219.93], [1008.2, 219.66], [1022.34, 221.15], [953.84, 193.39], [941.75, 192.75], [1008.2, 192.63], [1022.34, 193.26]]}, {\"category\": \"car\", \"corners_3d\": [[567.69, 223.8], [612.34, 223.74], [615.76, 234.21], [561.92, 234.3], [567.69, 179.42], [612.34, 179.41], [615.76, 180.76], [561.92, 180.77]]}]\n```" - } - ] - }, - { - "id": 3688, - "data_source": "Mono3DRefer", - "scene_name": "006838", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[280.48, 377.5], [19.16, 377.56], [330.6, 269.47], [453.91, 269.46], [280.48, 43.41], [19.16, 43.37], [330.6, 111.74], [453.91, 111.75]]}, {\"category\": \"cyclist\", \"corners_3d\": [[852.25, 220.09], [859.77, 220.05], [875.43, 222.51], [867.53, 222.55], [852.25, 182.84], [859.77, 182.83], [875.43, 183.35], [867.53, 183.36]]}]\n```" - } - ] - }, - { - "id": 3689, - "data_source": "Mono3DRefer", - "scene_name": "006838", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[280.48, 377.5], [19.16, 377.56], [330.6, 269.47], [453.91, 269.46], [280.48, 43.41], [19.16, 43.37], [330.6, 111.74], [453.91, 111.75]]}, {\"category\": \"car\", \"corners_3d\": [[555.21, 198.79], [538.9, 198.8], [541.17, 197.59], [556.71, 197.57], [555.21, 182.06], [538.9, 182.07], [541.17, 181.64], [556.71, 181.63]]}, {\"category\": \"car\", \"corners_3d\": [[953.84, 221.44], [941.75, 219.93], [1008.2, 219.66], [1022.34, 221.15], [953.84, 193.39], [941.75, 192.75], [1008.2, 192.63], [1022.34, 193.26]]}, {\"category\": \"car\", \"corners_3d\": [[567.69, 223.8], [612.34, 223.74], [615.76, 234.21], [561.92, 234.3], [567.69, 179.42], [612.34, 179.41], [615.76, 180.76], [561.92, 180.77]]}]\n```" - } - ] - }, - { - "id": 3690, - "data_source": "Mono3DRefer", - "scene_name": "006838", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[555.21, 198.79], [538.9, 198.8], [541.17, 197.59], [556.71, 197.57], [555.21, 182.06], [538.9, 182.07], [541.17, 181.64], [556.71, 181.63]]}, {\"category\": \"car\", \"corners_3d\": [[953.84, 221.44], [941.75, 219.93], [1008.2, 219.66], [1022.34, 221.15], [953.84, 193.39], [941.75, 192.75], [1008.2, 192.63], [1022.34, 193.26]]}, {\"category\": \"car\", \"corners_3d\": [[567.69, 223.8], [612.34, 223.74], [615.76, 234.21], [561.92, 234.3], [567.69, 179.42], [612.34, 179.41], [615.76, 180.76], [561.92, 180.77]]}, {\"category\": \"cyclist\", \"corners_3d\": [[852.25, 220.09], [859.77, 220.05], [875.43, 222.51], [867.53, 222.55], [852.25, 182.84], [859.77, 182.83], [875.43, 183.35], [867.53, 183.36]]}]\n```" - } - ] - }, - { - "id": 3691, - "data_source": "Mono3DRefer", - "scene_name": "006838", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[280.48, 377.5], [19.16, 377.56], [330.6, 269.47], [453.91, 269.46], [280.48, 43.41], [19.16, 43.37], [330.6, 111.74], [453.91, 111.75]]}, {\"category\": \"car\", \"corners_3d\": [[555.21, 198.79], [538.9, 198.8], [541.17, 197.59], [556.71, 197.57], [555.21, 182.06], [538.9, 182.07], [541.17, 181.64], [556.71, 181.63]]}, {\"category\": \"car\", \"corners_3d\": [[953.84, 221.44], [941.75, 219.93], [1008.2, 219.66], [1022.34, 221.15], [953.84, 193.39], [941.75, 192.75], [1008.2, 192.63], [1022.34, 193.26]]}, {\"category\": \"car\", \"corners_3d\": [[567.69, 223.8], [612.34, 223.74], [615.76, 234.21], [561.92, 234.3], [567.69, 179.42], [612.34, 179.41], [615.76, 180.76], [561.92, 180.77]]}, {\"category\": \"cyclist\", \"corners_3d\": [[852.25, 220.09], [859.77, 220.05], [875.43, 222.51], [867.53, 222.55], [852.25, 182.84], [859.77, 182.83], [875.43, 183.35], [867.53, 183.36]]}]\n```" - } - ] - }, - { - "id": 3692, - "data_source": "Mono3DRefer", - "scene_name": "006860", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[413.19, 281.92], [299.18, 282.45], [382.28, 250.84], [463.24, 250.58], [413.19, 177.69], [299.18, 177.72], [382.28, 176.31], [463.24, 176.3]]}]\n```" - } - ] - }, - { - "id": 3693, - "data_source": "Mono3DRefer", - "scene_name": "006865", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[582.87, 194.91], [564.25, 194.92], [566.45, 193.63], [583.97, 193.63], [582.87, 179.34], [564.25, 179.34], [566.45, 178.96], [583.97, 178.96]]}]\n```" - } - ] - }, - { - "id": 3694, - "data_source": "Mono3DRefer", - "scene_name": "006871", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[367.8, 269.99], [280.15, 270.46], [343.57, 249.83], [412.51, 249.53], [367.8, 196.57], [280.15, 196.69], [343.57, 191.65], [412.51, 191.58]]}, {\"category\": \"car\", \"corners_3d\": [[-90.09, 339.22], [-259.1, 342.81], [18.9, 283.43], [125.83, 281.9], [-90.09, 209.8], [-259.1, 210.6], [18.9, 197.41], [125.83, 197.07]]}, {\"category\": \"car\", \"corners_3d\": [[182.67, 268.98], [94.44, 270.1], [195.14, 247.95], [262.56, 247.28], [182.67, 201.27], [94.44, 201.6], [195.14, 195.05], [262.56, 194.85]]}, {\"category\": \"car\", \"corners_3d\": [[280.0, 240.56], [232.14, 239.56], [307.87, 229.91], [349.68, 230.64], [280.0, 191.15], [232.14, 190.88], [307.87, 188.27], [349.68, 188.47]]}, {\"category\": \"car\", \"corners_3d\": [[662.43, 200.63], [690.21, 200.77], [688.85, 203.42], [658.47, 203.25], [662.43, 176.39], [690.21, 176.4], [688.85, 176.74], [658.47, 176.72]]}]\n```" - } - ] - }, - { - "id": 3695, - "data_source": "Mono3DRefer", - "scene_name": "006880", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[560.84, 200.66], [539.11, 200.67], [543.66, 198.65], [563.81, 198.64], [560.84, 180.09], [539.11, 180.09], [543.66, 179.56], [563.81, 179.56]]}, {\"category\": \"car\", \"corners_3d\": [[524.09, 222.88], [482.86, 222.85], [499.84, 216.48], [535.83, 216.5], [524.09, 179.2], [482.86, 179.19], [499.84, 178.39], [535.83, 178.39]]}]\n```" - } - ] - }, - { - "id": 3696, - "data_source": "Mono3DRefer", - "scene_name": "006885", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[215.01, 223.58], [161.1, 225.03], [174.45, 218.01], [220.96, 216.92], [215.01, 165.47], [161.1, 165.26], [174.45, 166.28], [220.96, 166.44]]}]\n```" - } - ] - }, - { - "id": 3697, - "data_source": "Mono3DRefer", - "scene_name": "006885", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[311.38, 212.52], [340.31, 211.95], [343.33, 215.31], [311.93, 215.98], [311.38, 187.77], [340.31, 187.55], [343.33, 188.82], [311.93, 189.07]]}, {\"category\": \"car\", \"corners_3d\": [[382.88, 239.06], [442.11, 236.83], [471.53, 247.81], [402.93, 250.88], [382.88, 185.19], [442.11, 184.78], [471.53, 186.82], [402.93, 187.4]]}, {\"category\": \"car\", \"corners_3d\": [[144.31, 347.94], [-109.14, 372.82], [21.28, 292.27], [170.42, 282.93], [144.31, 149.07], [-109.14, 145.69], [21.28, 156.63], [170.42, 157.9]]}, {\"category\": \"car\", \"corners_3d\": [[156.62, 233.72], [207.65, 231.97], [206.49, 240.05], [148.21, 242.31], [156.62, 189.66], [207.65, 189.18], [206.49, 191.41], [148.21, 192.04]]}, {\"category\": \"car\", \"corners_3d\": [[357.18, 229.36], [399.45, 227.96], [416.21, 234.26], [369.44, 236.0], [357.18, 187.86], [399.45, 187.49], [416.21, 189.16], [369.44, 189.62]]}, {\"category\": \"car\", \"corners_3d\": [[334.15, 223.91], [372.83, 222.75], [387.23, 228.32], [344.49, 229.76], [334.15, 187.91], [372.83, 187.56], [387.23, 189.21], [344.49, 189.63]]}]\n```" - } - ] - }, - { - "id": 3698, - "data_source": "Mono3DRefer", - "scene_name": "006885", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[311.38, 212.52], [340.31, 211.95], [343.33, 215.31], [311.93, 215.98], [311.38, 187.77], [340.31, 187.55], [343.33, 188.82], [311.93, 189.07]]}, {\"category\": \"car\", \"corners_3d\": [[382.88, 239.06], [442.11, 236.83], [471.53, 247.81], [402.93, 250.88], [382.88, 185.19], [442.11, 184.78], [471.53, 186.82], [402.93, 187.4]]}, {\"category\": \"car\", \"corners_3d\": [[144.31, 347.94], [-109.14, 372.82], [21.28, 292.27], [170.42, 282.93], [144.31, 149.07], [-109.14, 145.69], [21.28, 156.63], [170.42, 157.9]]}, {\"category\": \"car\", \"corners_3d\": [[156.62, 233.72], [207.65, 231.97], [206.49, 240.05], [148.21, 242.31], [156.62, 189.66], [207.65, 189.18], [206.49, 191.41], [148.21, 192.04]]}, {\"category\": \"car\", \"corners_3d\": [[357.18, 229.36], [399.45, 227.96], [416.21, 234.26], [369.44, 236.0], [357.18, 187.86], [399.45, 187.49], [416.21, 189.16], [369.44, 189.62]]}, {\"category\": \"car\", \"corners_3d\": [[334.15, 223.91], [372.83, 222.75], [387.23, 228.32], [344.49, 229.76], [334.15, 187.91], [372.83, 187.56], [387.23, 189.21], [344.49, 189.63]]}, {\"category\": \"truck\", \"corners_3d\": [[215.01, 223.58], [161.1, 225.03], [174.45, 218.01], [220.96, 216.92], [215.01, 165.47], [161.1, 165.26], [174.45, 166.28], [220.96, 166.44]]}]\n```" - } - ] - }, - { - "id": 3699, - "data_source": "Mono3DRefer", - "scene_name": "006886", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[179.29, 264.38], [119.33, 264.54], [199.87, 249.0], [249.57, 248.89], [179.29, 203.22], [119.33, 203.27], [199.87, 198.12], [249.57, 198.08]]}, {\"category\": \"car\", \"corners_3d\": [[529.03, 205.43], [497.76, 205.47], [505.93, 202.5], [534.34, 202.46], [529.03, 173.21], [497.76, 173.21], [505.93, 173.18], [534.34, 173.18]]}, {\"category\": \"car\", \"corners_3d\": [[413.12, 226.05], [364.78, 226.3], [389.37, 219.51], [431.5, 219.32], [413.12, 184.88], [364.78, 184.94], [389.37, 183.41], [431.5, 183.36]]}, {\"category\": \"car\", \"corners_3d\": [[326.98, 228.79], [285.78, 228.85], [325.98, 221.56], [361.78, 221.51], [326.98, 189.55], [285.78, 189.57], [325.98, 187.39], [361.78, 187.38]]}, {\"category\": \"car\", \"corners_3d\": [[1495.28, 260.63], [1518.12, 272.29], [1096.23, 246.68], [1116.61, 240.05], [1495.28, 143.08], [1518.12, 139.12], [1096.23, 147.81], [1116.61, 150.06]]}, {\"category\": \"car\", \"corners_3d\": [[340.79, 202.75], [338.44, 203.58], [275.46, 204.2], [279.58, 203.34], [340.79, 184.24], [338.44, 184.56], [275.46, 184.79], [279.58, 184.47]]}, {\"category\": \"car\", \"corners_3d\": [[981.9, 177.35], [960.02, 177.26], [975.12, 177.07], [996.32, 177.15], [981.9, 158.31], [960.02, 158.6], [975.12, 159.22], [996.32, 158.95]]}]\n```" - } - ] - }, - { - "id": 3700, - "data_source": "Mono3DRefer", - "scene_name": "006886", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[454.01, 200.2], [429.08, 200.12], [454.29, 197.23], [476.63, 197.29], [454.01, 162.64], [429.08, 162.67], [454.29, 163.75], [476.63, 163.73]]}]\n```" - } - ] - }, - { - "id": 3701, - "data_source": "Mono3DRefer", - "scene_name": "006886", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[268.99, 253.01], [244.18, 253.19], [274.45, 245.46], [296.81, 245.3], [268.99, 178.94], [244.18, 178.96], [274.45, 178.37], [296.81, 178.36]]}]\n```" - } - ] - }, - { - "id": 3702, - "data_source": "Mono3DRefer", - "scene_name": "006886", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[472.18, 200.58], [476.61, 200.84], [469.76, 201.03], [465.37, 200.77], [472.18, 177.58], [476.61, 177.62], [469.76, 177.66], [465.37, 177.61]]}]\n```" - } - ] - }, - { - "id": 3703, - "data_source": "Mono3DRefer", - "scene_name": "006886", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[179.29, 264.38], [119.33, 264.54], [199.87, 249.0], [249.57, 248.89], [179.29, 203.22], [119.33, 203.27], [199.87, 198.12], [249.57, 198.08]]}, {\"category\": \"car\", \"corners_3d\": [[529.03, 205.43], [497.76, 205.47], [505.93, 202.5], [534.34, 202.46], [529.03, 173.21], [497.76, 173.21], [505.93, 173.18], [534.34, 173.18]]}, {\"category\": \"van\", \"corners_3d\": [[454.01, 200.2], [429.08, 200.12], [454.29, 197.23], [476.63, 197.29], [454.01, 162.64], [429.08, 162.67], [454.29, 163.75], [476.63, 163.73]]}, {\"category\": \"car\", \"corners_3d\": [[413.12, 226.05], [364.78, 226.3], [389.37, 219.51], [431.5, 219.32], [413.12, 184.88], [364.78, 184.94], [389.37, 183.41], [431.5, 183.36]]}, {\"category\": \"car\", \"corners_3d\": [[326.98, 228.79], [285.78, 228.85], [325.98, 221.56], [361.78, 221.51], [326.98, 189.55], [285.78, 189.57], [325.98, 187.39], [361.78, 187.38]]}, {\"category\": \"car\", \"corners_3d\": [[1495.28, 260.63], [1518.12, 272.29], [1096.23, 246.68], [1116.61, 240.05], [1495.28, 143.08], [1518.12, 139.12], [1096.23, 147.81], [1116.61, 150.06]]}, {\"category\": \"car\", \"corners_3d\": [[340.79, 202.75], [338.44, 203.58], [275.46, 204.2], [279.58, 203.34], [340.79, 184.24], [338.44, 184.56], [275.46, 184.79], [279.58, 184.47]]}, {\"category\": \"car\", \"corners_3d\": [[981.9, 177.35], [960.02, 177.26], [975.12, 177.07], [996.32, 177.15], [981.9, 158.31], [960.02, 158.6], [975.12, 159.22], [996.32, 158.95]]}]\n```" - } - ] - }, - { - "id": 3704, - "data_source": "Mono3DRefer", - "scene_name": "006886", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[179.29, 264.38], [119.33, 264.54], [199.87, 249.0], [249.57, 248.89], [179.29, 203.22], [119.33, 203.27], [199.87, 198.12], [249.57, 198.08]]}, {\"category\": \"car\", \"corners_3d\": [[529.03, 205.43], [497.76, 205.47], [505.93, 202.5], [534.34, 202.46], [529.03, 173.21], [497.76, 173.21], [505.93, 173.18], [534.34, 173.18]]}, {\"category\": \"car\", \"corners_3d\": [[413.12, 226.05], [364.78, 226.3], [389.37, 219.51], [431.5, 219.32], [413.12, 184.88], [364.78, 184.94], [389.37, 183.41], [431.5, 183.36]]}, {\"category\": \"car\", \"corners_3d\": [[326.98, 228.79], [285.78, 228.85], [325.98, 221.56], [361.78, 221.51], [326.98, 189.55], [285.78, 189.57], [325.98, 187.39], [361.78, 187.38]]}, {\"category\": \"cyclist\", \"corners_3d\": [[268.99, 253.01], [244.18, 253.19], [274.45, 245.46], [296.81, 245.3], [268.99, 178.94], [244.18, 178.96], [274.45, 178.37], [296.81, 178.36]]}, {\"category\": \"car\", \"corners_3d\": [[1495.28, 260.63], [1518.12, 272.29], [1096.23, 246.68], [1116.61, 240.05], [1495.28, 143.08], [1518.12, 139.12], [1096.23, 147.81], [1116.61, 150.06]]}, {\"category\": \"car\", \"corners_3d\": [[340.79, 202.75], [338.44, 203.58], [275.46, 204.2], [279.58, 203.34], [340.79, 184.24], [338.44, 184.56], [275.46, 184.79], [279.58, 184.47]]}, {\"category\": \"car\", \"corners_3d\": [[981.9, 177.35], [960.02, 177.26], [975.12, 177.07], [996.32, 177.15], [981.9, 158.31], [960.02, 158.6], [975.12, 159.22], [996.32, 158.95]]}]\n```" - } - ] - }, - { - "id": 3705, - "data_source": "Mono3DRefer", - "scene_name": "006886", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[179.29, 264.38], [119.33, 264.54], [199.87, 249.0], [249.57, 248.89], [179.29, 203.22], [119.33, 203.27], [199.87, 198.12], [249.57, 198.08]]}, {\"category\": \"car\", \"corners_3d\": [[529.03, 205.43], [497.76, 205.47], [505.93, 202.5], [534.34, 202.46], [529.03, 173.21], [497.76, 173.21], [505.93, 173.18], [534.34, 173.18]]}, {\"category\": \"car\", \"corners_3d\": [[413.12, 226.05], [364.78, 226.3], [389.37, 219.51], [431.5, 219.32], [413.12, 184.88], [364.78, 184.94], [389.37, 183.41], [431.5, 183.36]]}, {\"category\": \"car\", \"corners_3d\": [[326.98, 228.79], [285.78, 228.85], [325.98, 221.56], [361.78, 221.51], [326.98, 189.55], [285.78, 189.57], [325.98, 187.39], [361.78, 187.38]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[472.18, 200.58], [476.61, 200.84], [469.76, 201.03], [465.37, 200.77], [472.18, 177.58], [476.61, 177.62], [469.76, 177.66], [465.37, 177.61]]}, {\"category\": \"car\", \"corners_3d\": [[1495.28, 260.63], [1518.12, 272.29], [1096.23, 246.68], [1116.61, 240.05], [1495.28, 143.08], [1518.12, 139.12], [1096.23, 147.81], [1116.61, 150.06]]}, {\"category\": \"car\", \"corners_3d\": [[340.79, 202.75], [338.44, 203.58], [275.46, 204.2], [279.58, 203.34], [340.79, 184.24], [338.44, 184.56], [275.46, 184.79], [279.58, 184.47]]}, {\"category\": \"car\", \"corners_3d\": [[981.9, 177.35], [960.02, 177.26], [975.12, 177.07], [996.32, 177.15], [981.9, 158.31], [960.02, 158.6], [975.12, 159.22], [996.32, 158.95]]}]\n```" - } - ] - }, - { - "id": 3706, - "data_source": "Mono3DRefer", - "scene_name": "006886", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[454.01, 200.2], [429.08, 200.12], [454.29, 197.23], [476.63, 197.29], [454.01, 162.64], [429.08, 162.67], [454.29, 163.75], [476.63, 163.73]]}, {\"category\": \"cyclist\", \"corners_3d\": [[268.99, 253.01], [244.18, 253.19], [274.45, 245.46], [296.81, 245.3], [268.99, 178.94], [244.18, 178.96], [274.45, 178.37], [296.81, 178.36]]}]\n```" - } - ] - }, - { - "id": 3707, - "data_source": "Mono3DRefer", - "scene_name": "006886", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[454.01, 200.2], [429.08, 200.12], [454.29, 197.23], [476.63, 197.29], [454.01, 162.64], [429.08, 162.67], [454.29, 163.75], [476.63, 163.73]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[472.18, 200.58], [476.61, 200.84], [469.76, 201.03], [465.37, 200.77], [472.18, 177.58], [476.61, 177.62], [469.76, 177.66], [465.37, 177.61]]}]\n```" - } - ] - }, - { - "id": 3708, - "data_source": "Mono3DRefer", - "scene_name": "006886", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[268.99, 253.01], [244.18, 253.19], [274.45, 245.46], [296.81, 245.3], [268.99, 178.94], [244.18, 178.96], [274.45, 178.37], [296.81, 178.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[472.18, 200.58], [476.61, 200.84], [469.76, 201.03], [465.37, 200.77], [472.18, 177.58], [476.61, 177.62], [469.76, 177.66], [465.37, 177.61]]}]\n```" - } - ] - }, - { - "id": 3709, - "data_source": "Mono3DRefer", - "scene_name": "006886", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[179.29, 264.38], [119.33, 264.54], [199.87, 249.0], [249.57, 248.89], [179.29, 203.22], [119.33, 203.27], [199.87, 198.12], [249.57, 198.08]]}, {\"category\": \"car\", \"corners_3d\": [[529.03, 205.43], [497.76, 205.47], [505.93, 202.5], [534.34, 202.46], [529.03, 173.21], [497.76, 173.21], [505.93, 173.18], [534.34, 173.18]]}, {\"category\": \"van\", \"corners_3d\": [[454.01, 200.2], [429.08, 200.12], [454.29, 197.23], [476.63, 197.29], [454.01, 162.64], [429.08, 162.67], [454.29, 163.75], [476.63, 163.73]]}, {\"category\": \"car\", \"corners_3d\": [[413.12, 226.05], [364.78, 226.3], [389.37, 219.51], [431.5, 219.32], [413.12, 184.88], [364.78, 184.94], [389.37, 183.41], [431.5, 183.36]]}, {\"category\": \"car\", \"corners_3d\": [[326.98, 228.79], [285.78, 228.85], [325.98, 221.56], [361.78, 221.51], [326.98, 189.55], [285.78, 189.57], [325.98, 187.39], [361.78, 187.38]]}, {\"category\": \"cyclist\", \"corners_3d\": [[268.99, 253.01], [244.18, 253.19], [274.45, 245.46], [296.81, 245.3], [268.99, 178.94], [244.18, 178.96], [274.45, 178.37], [296.81, 178.36]]}, {\"category\": \"car\", \"corners_3d\": [[1495.28, 260.63], [1518.12, 272.29], [1096.23, 246.68], [1116.61, 240.05], [1495.28, 143.08], [1518.12, 139.12], [1096.23, 147.81], [1116.61, 150.06]]}, {\"category\": \"car\", \"corners_3d\": [[340.79, 202.75], [338.44, 203.58], [275.46, 204.2], [279.58, 203.34], [340.79, 184.24], [338.44, 184.56], [275.46, 184.79], [279.58, 184.47]]}, {\"category\": \"car\", \"corners_3d\": [[981.9, 177.35], [960.02, 177.26], [975.12, 177.07], [996.32, 177.15], [981.9, 158.31], [960.02, 158.6], [975.12, 159.22], [996.32, 158.95]]}]\n```" - } - ] - }, - { - "id": 3710, - "data_source": "Mono3DRefer", - "scene_name": "006886", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[179.29, 264.38], [119.33, 264.54], [199.87, 249.0], [249.57, 248.89], [179.29, 203.22], [119.33, 203.27], [199.87, 198.12], [249.57, 198.08]]}, {\"category\": \"car\", \"corners_3d\": [[529.03, 205.43], [497.76, 205.47], [505.93, 202.5], [534.34, 202.46], [529.03, 173.21], [497.76, 173.21], [505.93, 173.18], [534.34, 173.18]]}, {\"category\": \"van\", \"corners_3d\": [[454.01, 200.2], [429.08, 200.12], [454.29, 197.23], [476.63, 197.29], [454.01, 162.64], [429.08, 162.67], [454.29, 163.75], [476.63, 163.73]]}, {\"category\": \"car\", \"corners_3d\": [[413.12, 226.05], [364.78, 226.3], [389.37, 219.51], [431.5, 219.32], [413.12, 184.88], [364.78, 184.94], [389.37, 183.41], [431.5, 183.36]]}, {\"category\": \"car\", \"corners_3d\": [[326.98, 228.79], [285.78, 228.85], [325.98, 221.56], [361.78, 221.51], [326.98, 189.55], [285.78, 189.57], [325.98, 187.39], [361.78, 187.38]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[472.18, 200.58], [476.61, 200.84], [469.76, 201.03], [465.37, 200.77], [472.18, 177.58], [476.61, 177.62], [469.76, 177.66], [465.37, 177.61]]}, {\"category\": \"car\", \"corners_3d\": [[1495.28, 260.63], [1518.12, 272.29], [1096.23, 246.68], [1116.61, 240.05], [1495.28, 143.08], [1518.12, 139.12], [1096.23, 147.81], [1116.61, 150.06]]}, {\"category\": \"car\", \"corners_3d\": [[340.79, 202.75], [338.44, 203.58], [275.46, 204.2], [279.58, 203.34], [340.79, 184.24], [338.44, 184.56], [275.46, 184.79], [279.58, 184.47]]}, {\"category\": \"car\", \"corners_3d\": [[981.9, 177.35], [960.02, 177.26], [975.12, 177.07], [996.32, 177.15], [981.9, 158.31], [960.02, 158.6], [975.12, 159.22], [996.32, 158.95]]}]\n```" - } - ] - }, - { - "id": 3711, - "data_source": "Mono3DRefer", - "scene_name": "006886", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[179.29, 264.38], [119.33, 264.54], [199.87, 249.0], [249.57, 248.89], [179.29, 203.22], [119.33, 203.27], [199.87, 198.12], [249.57, 198.08]]}, {\"category\": \"car\", \"corners_3d\": [[529.03, 205.43], [497.76, 205.47], [505.93, 202.5], [534.34, 202.46], [529.03, 173.21], [497.76, 173.21], [505.93, 173.18], [534.34, 173.18]]}, {\"category\": \"car\", \"corners_3d\": [[413.12, 226.05], [364.78, 226.3], [389.37, 219.51], [431.5, 219.32], [413.12, 184.88], [364.78, 184.94], [389.37, 183.41], [431.5, 183.36]]}, {\"category\": \"car\", \"corners_3d\": [[326.98, 228.79], [285.78, 228.85], [325.98, 221.56], [361.78, 221.51], [326.98, 189.55], [285.78, 189.57], [325.98, 187.39], [361.78, 187.38]]}, {\"category\": \"cyclist\", \"corners_3d\": [[268.99, 253.01], [244.18, 253.19], [274.45, 245.46], [296.81, 245.3], [268.99, 178.94], [244.18, 178.96], [274.45, 178.37], [296.81, 178.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[472.18, 200.58], [476.61, 200.84], [469.76, 201.03], [465.37, 200.77], [472.18, 177.58], [476.61, 177.62], [469.76, 177.66], [465.37, 177.61]]}, {\"category\": \"car\", \"corners_3d\": [[1495.28, 260.63], [1518.12, 272.29], [1096.23, 246.68], [1116.61, 240.05], [1495.28, 143.08], [1518.12, 139.12], [1096.23, 147.81], [1116.61, 150.06]]}, {\"category\": \"car\", \"corners_3d\": [[340.79, 202.75], [338.44, 203.58], [275.46, 204.2], [279.58, 203.34], [340.79, 184.24], [338.44, 184.56], [275.46, 184.79], [279.58, 184.47]]}, {\"category\": \"car\", \"corners_3d\": [[981.9, 177.35], [960.02, 177.26], [975.12, 177.07], [996.32, 177.15], [981.9, 158.31], [960.02, 158.6], [975.12, 159.22], [996.32, 158.95]]}]\n```" - } - ] - }, - { - "id": 3712, - "data_source": "Mono3DRefer", - "scene_name": "006886", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[454.01, 200.2], [429.08, 200.12], [454.29, 197.23], [476.63, 197.29], [454.01, 162.64], [429.08, 162.67], [454.29, 163.75], [476.63, 163.73]]}, {\"category\": \"cyclist\", \"corners_3d\": [[268.99, 253.01], [244.18, 253.19], [274.45, 245.46], [296.81, 245.3], [268.99, 178.94], [244.18, 178.96], [274.45, 178.37], [296.81, 178.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[472.18, 200.58], [476.61, 200.84], [469.76, 201.03], [465.37, 200.77], [472.18, 177.58], [476.61, 177.62], [469.76, 177.66], [465.37, 177.61]]}]\n```" - } - ] - }, - { - "id": 3713, - "data_source": "Mono3DRefer", - "scene_name": "006886", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[179.29, 264.38], [119.33, 264.54], [199.87, 249.0], [249.57, 248.89], [179.29, 203.22], [119.33, 203.27], [199.87, 198.12], [249.57, 198.08]]}, {\"category\": \"car\", \"corners_3d\": [[529.03, 205.43], [497.76, 205.47], [505.93, 202.5], [534.34, 202.46], [529.03, 173.21], [497.76, 173.21], [505.93, 173.18], [534.34, 173.18]]}, {\"category\": \"van\", \"corners_3d\": [[454.01, 200.2], [429.08, 200.12], [454.29, 197.23], [476.63, 197.29], [454.01, 162.64], [429.08, 162.67], [454.29, 163.75], [476.63, 163.73]]}, {\"category\": \"car\", \"corners_3d\": [[413.12, 226.05], [364.78, 226.3], [389.37, 219.51], [431.5, 219.32], [413.12, 184.88], [364.78, 184.94], [389.37, 183.41], [431.5, 183.36]]}, {\"category\": \"car\", \"corners_3d\": [[326.98, 228.79], [285.78, 228.85], [325.98, 221.56], [361.78, 221.51], [326.98, 189.55], [285.78, 189.57], [325.98, 187.39], [361.78, 187.38]]}, {\"category\": \"cyclist\", \"corners_3d\": [[268.99, 253.01], [244.18, 253.19], [274.45, 245.46], [296.81, 245.3], [268.99, 178.94], [244.18, 178.96], [274.45, 178.37], [296.81, 178.36]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[472.18, 200.58], [476.61, 200.84], [469.76, 201.03], [465.37, 200.77], [472.18, 177.58], [476.61, 177.62], [469.76, 177.66], [465.37, 177.61]]}, {\"category\": \"car\", \"corners_3d\": [[1495.28, 260.63], [1518.12, 272.29], [1096.23, 246.68], [1116.61, 240.05], [1495.28, 143.08], [1518.12, 139.12], [1096.23, 147.81], [1116.61, 150.06]]}, {\"category\": \"car\", \"corners_3d\": [[340.79, 202.75], [338.44, 203.58], [275.46, 204.2], [279.58, 203.34], [340.79, 184.24], [338.44, 184.56], [275.46, 184.79], [279.58, 184.47]]}, {\"category\": \"car\", \"corners_3d\": [[981.9, 177.35], [960.02, 177.26], [975.12, 177.07], [996.32, 177.15], [981.9, 158.31], [960.02, 158.6], [975.12, 159.22], [996.32, 158.95]]}]\n```" - } - ] - }, - { - "id": 3714, - "data_source": "Mono3DRefer", - "scene_name": "006899", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[867.34, 242.88], [941.44, 245.33], [954.1, 263.19], [861.97, 259.41], [867.34, 171.11], [941.44, 171.05], [954.1, 170.6], [861.97, 170.7]]}, {\"category\": \"car\", \"corners_3d\": [[764.97, 207.57], [729.15, 207.03], [742.62, 203.4], [774.77, 203.84], [764.97, 175.9], [729.15, 175.86], [742.62, 175.54], [774.77, 175.58]]}, {\"category\": \"car\", \"corners_3d\": [[803.51, 197.18], [778.2, 196.87], [787.42, 195.01], [810.85, 195.27], [803.51, 175.54], [778.2, 175.5], [787.42, 175.3], [810.85, 175.32]]}]\n```" - } - ] - }, - { - "id": 3715, - "data_source": "Mono3DRefer", - "scene_name": "006917", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[876.16, 296.6], [760.92, 296.42], [720.17, 261.62], [802.89, 261.72], [876.16, 178.65], [760.92, 178.64], [720.17, 177.01], [802.89, 177.02]]}, {\"category\": \"car\", \"corners_3d\": [[509.64, 252.41], [435.76, 252.25], [475.93, 235.26], [534.03, 235.35], [509.64, 187.95], [435.76, 187.92], [475.93, 184.7], [534.03, 184.72]]}, {\"category\": \"car\", \"corners_3d\": [[678.87, 232.82], [734.55, 232.64], [767.37, 245.44], [699.85, 245.71], [678.87, 180.96], [734.55, 180.93], [767.37, 182.66], [699.85, 182.7]]}, {\"category\": \"car\", \"corners_3d\": [[729.48, 230.18], [675.39, 230.14], [665.31, 220.38], [710.17, 220.41], [729.48, 182.73], [675.39, 182.72], [665.31, 181.04], [710.17, 181.05]]}, {\"category\": \"car\", \"corners_3d\": [[559.23, 214.53], [524.94, 214.57], [531.37, 210.72], [562.49, 210.68], [559.23, 181.45], [524.94, 181.46], [531.37, 180.67], [562.49, 180.66]]}]\n```" - } - ] - }, - { - "id": 3716, - "data_source": "Mono3DRefer", - "scene_name": "006920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-54.43, 453.3], [-306.17, 456.23], [99.09, 327.7], [235.0, 326.82], [-54.43, 239.87], [-306.17, 240.58], [99.09, 209.86], [235.0, 209.65]]}, {\"category\": \"car\", \"corners_3d\": [[2514.94, 1074.62], [1549.4, 1031.73], [885.78, 406.45], [1130.47, 409.51], [2514.94, 242.03], [1549.4, 238.74], [885.78, 190.77], [1130.47, 191.01]]}, {\"category\": \"car\", \"corners_3d\": [[665.53, 223.63], [718.25, 223.63], [740.07, 233.88], [676.71, 233.87], [665.53, 178.31], [718.25, 178.31], [740.07, 179.41], [676.71, 179.41]]}, {\"category\": \"car\", \"corners_3d\": [[540.13, 204.36], [511.09, 204.36], [519.24, 201.74], [545.86, 201.74], [540.13, 180.77], [511.09, 180.77], [519.24, 180.11], [545.86, 180.11]]}, {\"category\": \"car\", \"corners_3d\": [[380.49, 272.39], [303.65, 272.18], [374.93, 250.04], [434.73, 250.17], [380.49, 208.15], [303.65, 208.07], [374.93, 200.22], [434.73, 200.27]]}]\n```" - } - ] - }, - { - "id": 3717, - "data_source": "Mono3DRefer", - "scene_name": "006920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[410.93, 375.25], [178.39, 376.6], [366.11, 284.7], [493.45, 284.29], [410.93, 142.11], [178.39, 141.9], [366.11, 155.86], [493.45, 155.93]]}]\n```" - } - ] - }, - { - "id": 3718, - "data_source": "Mono3DRefer", - "scene_name": "006920", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[410.93, 375.25], [178.39, 376.6], [366.11, 284.7], [493.45, 284.29], [410.93, 142.11], [178.39, 141.9], [366.11, 155.86], [493.45, 155.93]]}, {\"category\": \"car\", \"corners_3d\": [[-54.43, 453.3], [-306.17, 456.23], [99.09, 327.7], [235.0, 326.82], [-54.43, 239.87], [-306.17, 240.58], [99.09, 209.86], [235.0, 209.65]]}, {\"category\": \"car\", \"corners_3d\": [[2514.94, 1074.62], [1549.4, 1031.73], [885.78, 406.45], [1130.47, 409.51], [2514.94, 242.03], [1549.4, 238.74], [885.78, 190.77], [1130.47, 191.01]]}, {\"category\": \"car\", \"corners_3d\": [[665.53, 223.63], [718.25, 223.63], [740.07, 233.88], [676.71, 233.87], [665.53, 178.31], [718.25, 178.31], [740.07, 179.41], [676.71, 179.41]]}, {\"category\": \"car\", \"corners_3d\": [[540.13, 204.36], [511.09, 204.36], [519.24, 201.74], [545.86, 201.74], [540.13, 180.77], [511.09, 180.77], [519.24, 180.11], [545.86, 180.11]]}, {\"category\": \"car\", \"corners_3d\": [[380.49, 272.39], [303.65, 272.18], [374.93, 250.04], [434.73, 250.17], [380.49, 208.15], [303.65, 208.07], [374.93, 200.22], [434.73, 200.27]]}]\n```" - } - ] - }, - { - "id": 3719, - "data_source": "Mono3DRefer", - "scene_name": "006925", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[118.9, 323.84], [185.89, 315.76], [198.55, 321.01], [129.57, 329.75], [118.9, 159.53], [185.89, 160.78], [198.55, 159.97], [129.57, 158.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[542.13, 274.0], [494.24, 270.22], [527.38, 264.63], [573.45, 267.95], [542.13, 183.23], [494.24, 183.16], [527.38, 183.05], [573.45, 183.11]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[843.31, 245.09], [875.87, 246.64], [868.44, 249.11], [834.86, 247.44], [843.31, 170.13], [875.87, 169.85], [868.44, 169.41], [834.86, 169.71]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[457.83, 262.71], [452.09, 266.81], [411.52, 267.0], [419.22, 262.89], [457.83, 174.53], [452.09, 174.17], [411.52, 174.16], [419.22, 174.51]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[541.0, 247.5], [502.44, 248.14], [500.73, 244.72], [537.35, 244.15], [541.0, 172.76], [502.44, 172.67], [500.73, 173.12], [537.35, 173.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[767.98, 230.64], [800.07, 230.6], [814.27, 233.87], [780.06, 233.91], [767.98, 161.78], [800.07, 161.8], [814.27, 160.48], [780.06, 160.46]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[593.67, 226.15], [623.69, 226.19], [624.05, 228.11], [592.74, 228.07], [593.67, 169.66], [623.69, 169.65], [624.05, 169.13], [592.74, 169.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[597.7, 218.31], [580.09, 218.17], [583.88, 217.09], [600.98, 217.22], [597.7, 172.13], [580.09, 172.16], [583.88, 172.44], [600.98, 172.41]]}]\n```" - } - ] - }, - { - "id": 3720, - "data_source": "Mono3DRefer", - "scene_name": "006931", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[553.26, 196.35], [534.82, 196.35], [538.91, 194.92], [556.22, 194.91], [553.26, 176.86], [534.82, 176.87], [538.91, 176.62], [556.22, 176.62]]}]\n```" - } - ] - }, - { - "id": 3721, - "data_source": "Mono3DRefer", - "scene_name": "006935", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1021.41, 479.13], [1152.91, 478.62], [1375.18, 597.94], [1191.45, 598.95], [1021.41, 121.52], [1152.91, 121.62], [1375.18, 98.05], [1191.45, 97.85]]}]\n```" - } - ] - }, - { - "id": 3722, - "data_source": "Mono3DRefer", - "scene_name": "006948", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[53.71, 372.71], [-120.02, 372.74], [153.18, 297.83], [261.78, 297.82], [53.71, 189.65], [-120.02, 189.65], [153.18, 183.36], [261.78, 183.36]]}, {\"category\": \"car\", \"corners_3d\": [[534.04, 216.56], [491.2, 216.51], [508.32, 210.86], [545.62, 210.89], [534.04, 182.13], [491.2, 182.11], [508.32, 180.92], [545.62, 180.92]]}, {\"category\": \"car\", \"corners_3d\": [[304.39, 274.94], [216.23, 274.83], [308.33, 251.35], [376.25, 251.42], [304.39, 187.96], [216.23, 187.95], [308.33, 184.47], [376.25, 184.48]]}, {\"category\": \"car\", \"corners_3d\": [[434.88, 230.48], [384.47, 230.52], [413.35, 222.86], [457.05, 222.82], [434.88, 188.21], [384.47, 188.22], [413.35, 186.18], [457.05, 186.17]]}, {\"category\": \"car\", \"corners_3d\": [[629.45, 196.46], [653.35, 196.46], [657.25, 198.59], [631.19, 198.59], [629.45, 174.14], [653.35, 174.14], [657.25, 174.26], [631.19, 174.26]]}, {\"category\": \"car\", \"corners_3d\": [[708.26, 191.74], [703.91, 190.97], [783.95, 190.95], [791.69, 191.72], [708.26, 165.29], [703.91, 165.6], [783.95, 165.61], [791.69, 165.3]]}, {\"category\": \"car\", \"corners_3d\": [[702.91, 191.57], [699.49, 190.82], [769.14, 190.84], [775.48, 191.59], [702.91, 159.64], [699.49, 160.17], [769.14, 160.16], [775.48, 159.63]]}]\n```" - } - ] - }, - { - "id": 3723, - "data_source": "Mono3DRefer", - "scene_name": "006948", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[947.35, 178.16], [962.37, 178.39], [891.25, 178.4], [879.05, 178.18], [947.35, 149.35], [962.37, 148.37], [891.25, 148.29], [879.05, 149.28]]}]\n```" - } - ] - }, - { - "id": 3724, - "data_source": "Mono3DRefer", - "scene_name": "006948", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[53.71, 372.71], [-120.02, 372.74], [153.18, 297.83], [261.78, 297.82], [53.71, 189.65], [-120.02, 189.65], [153.18, 183.36], [261.78, 183.36]]}, {\"category\": \"car\", \"corners_3d\": [[534.04, 216.56], [491.2, 216.51], [508.32, 210.86], [545.62, 210.89], [534.04, 182.13], [491.2, 182.11], [508.32, 180.92], [545.62, 180.92]]}, {\"category\": \"car\", \"corners_3d\": [[304.39, 274.94], [216.23, 274.83], [308.33, 251.35], [376.25, 251.42], [304.39, 187.96], [216.23, 187.95], [308.33, 184.47], [376.25, 184.48]]}, {\"category\": \"car\", \"corners_3d\": [[434.88, 230.48], [384.47, 230.52], [413.35, 222.86], [457.05, 222.82], [434.88, 188.21], [384.47, 188.22], [413.35, 186.18], [457.05, 186.17]]}, {\"category\": \"car\", \"corners_3d\": [[629.45, 196.46], [653.35, 196.46], [657.25, 198.59], [631.19, 198.59], [629.45, 174.14], [653.35, 174.14], [657.25, 174.26], [631.19, 174.26]]}, {\"category\": \"car\", \"corners_3d\": [[708.26, 191.74], [703.91, 190.97], [783.95, 190.95], [791.69, 191.72], [708.26, 165.29], [703.91, 165.6], [783.95, 165.61], [791.69, 165.3]]}, {\"category\": \"car\", \"corners_3d\": [[702.91, 191.57], [699.49, 190.82], [769.14, 190.84], [775.48, 191.59], [702.91, 159.64], [699.49, 160.17], [769.14, 160.16], [775.48, 159.63]]}, {\"category\": \"van\", \"corners_3d\": [[947.35, 178.16], [962.37, 178.39], [891.25, 178.4], [879.05, 178.18], [947.35, 149.35], [962.37, 148.37], [891.25, 148.29], [879.05, 149.28]]}]\n```" - } - ] - }, - { - "id": 3725, - "data_source": "Mono3DRefer", - "scene_name": "006949", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[298.82, 206.33], [296.79, 208.53], [62.79, 212.37], [80.91, 209.69], [298.82, 145.49], [296.79, 143.69], [62.79, 140.55], [80.91, 142.74]]}]\n```" - } - ] - }, - { - "id": 3726, - "data_source": "Mono3DRefer", - "scene_name": "006949", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[779.0, 216.74], [823.07, 216.66], [868.11, 225.07], [815.66, 225.18], [779.0, 176.57], [823.07, 176.57], [868.11, 177.28], [815.66, 177.29]]}]\n```" - } - ] - }, - { - "id": 3727, - "data_source": "Mono3DRefer", - "scene_name": "006949", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[779.0, 216.74], [823.07, 216.66], [868.11, 225.07], [815.66, 225.18], [779.0, 176.57], [823.07, 176.57], [868.11, 177.28], [815.66, 177.29]]}, {\"category\": \"truck\", \"corners_3d\": [[298.82, 206.33], [296.79, 208.53], [62.79, 212.37], [80.91, 209.69], [298.82, 145.49], [296.79, 143.69], [62.79, 140.55], [80.91, 142.74]]}]\n```" - } - ] - }, - { - "id": 3728, - "data_source": "Mono3DRefer", - "scene_name": "006951", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[408.52, 300.81], [419.1, 294.77], [504.03, 295.07], [497.92, 301.14], [408.52, 158.76], [419.1, 159.85], [504.03, 159.8], [497.92, 158.7]]}]\n```" - } - ] - }, - { - "id": 3729, - "data_source": "Mono3DRefer", - "scene_name": "006957", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[481.43, 221.51], [435.42, 221.55], [460.53, 214.19], [499.59, 214.17], [481.43, 174.92], [435.42, 174.92], [460.53, 174.61], [499.59, 174.61]]}]\n```" - } - ] - }, - { - "id": 3730, - "data_source": "Mono3DRefer", - "scene_name": "006957", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[897.9, 236.02], [921.25, 237.85], [900.88, 239.75], [877.46, 237.8], [897.9, 170.85], [921.25, 170.79], [900.88, 170.73], [877.46, 170.8]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[986.63, 227.9], [966.49, 226.84], [976.77, 225.51], [996.6, 226.51], [986.63, 175.65], [966.49, 175.59], [976.77, 175.53], [996.6, 175.58]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1135.64, 232.52], [1122.0, 231.0], [1134.18, 230.98], [1148.14, 232.51], [1135.64, 165.43], [1122.0, 165.62], [1134.18, 165.62], [1148.14, 165.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1053.5, 213.59], [1044.53, 213.07], [1047.97, 212.98], [1056.96, 213.51], [1053.5, 171.68], [1044.53, 171.7], [1047.97, 171.7], [1056.96, 171.68]]}]\n```" - } - ] - }, - { - "id": 3731, - "data_source": "Mono3DRefer", - "scene_name": "006957", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[579.35, 251.32], [653.35, 251.33], [666.21, 274.69], [570.18, 274.68], [579.35, 181.36], [653.35, 181.36], [666.21, 183.89], [570.18, 183.89]]}, {\"category\": \"car\", \"corners_3d\": [[74.6, 230.11], [52.4, 232.9], [-40.99, 233.69], [-14.11, 230.83], [74.6, 196.82], [52.4, 197.99], [-40.99, 198.32], [-14.11, 197.12]]}]\n```" - } - ] - }, - { - "id": 3732, - "data_source": "Mono3DRefer", - "scene_name": "006957", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[481.43, 221.51], [435.42, 221.55], [460.53, 214.19], [499.59, 214.17], [481.43, 174.92], [435.42, 174.92], [460.53, 174.61], [499.59, 174.61]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[897.9, 236.02], [921.25, 237.85], [900.88, 239.75], [877.46, 237.8], [897.9, 170.85], [921.25, 170.79], [900.88, 170.73], [877.46, 170.8]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[986.63, 227.9], [966.49, 226.84], [976.77, 225.51], [996.6, 226.51], [986.63, 175.65], [966.49, 175.59], [976.77, 175.53], [996.6, 175.58]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1135.64, 232.52], [1122.0, 231.0], [1134.18, 230.98], [1148.14, 232.51], [1135.64, 165.43], [1122.0, 165.62], [1134.18, 165.62], [1148.14, 165.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1053.5, 213.59], [1044.53, 213.07], [1047.97, 212.98], [1056.96, 213.51], [1053.5, 171.68], [1044.53, 171.7], [1047.97, 171.7], [1056.96, 171.68]]}]\n```" - } - ] - }, - { - "id": 3733, - "data_source": "Mono3DRefer", - "scene_name": "006957", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[579.35, 251.32], [653.35, 251.33], [666.21, 274.69], [570.18, 274.68], [579.35, 181.36], [653.35, 181.36], [666.21, 183.89], [570.18, 183.89]]}, {\"category\": \"van\", \"corners_3d\": [[481.43, 221.51], [435.42, 221.55], [460.53, 214.19], [499.59, 214.17], [481.43, 174.92], [435.42, 174.92], [460.53, 174.61], [499.59, 174.61]]}, {\"category\": \"car\", \"corners_3d\": [[74.6, 230.11], [52.4, 232.9], [-40.99, 233.69], [-14.11, 230.83], [74.6, 196.82], [52.4, 197.99], [-40.99, 198.32], [-14.11, 197.12]]}]\n```" - } - ] - }, - { - "id": 3734, - "data_source": "Mono3DRefer", - "scene_name": "006957", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[579.35, 251.32], [653.35, 251.33], [666.21, 274.69], [570.18, 274.68], [579.35, 181.36], [653.35, 181.36], [666.21, 183.89], [570.18, 183.89]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[897.9, 236.02], [921.25, 237.85], [900.88, 239.75], [877.46, 237.8], [897.9, 170.85], [921.25, 170.79], [900.88, 170.73], [877.46, 170.8]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[986.63, 227.9], [966.49, 226.84], [976.77, 225.51], [996.6, 226.51], [986.63, 175.65], [966.49, 175.59], [976.77, 175.53], [996.6, 175.58]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1135.64, 232.52], [1122.0, 231.0], [1134.18, 230.98], [1148.14, 232.51], [1135.64, 165.43], [1122.0, 165.62], [1134.18, 165.62], [1148.14, 165.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1053.5, 213.59], [1044.53, 213.07], [1047.97, 212.98], [1056.96, 213.51], [1053.5, 171.68], [1044.53, 171.7], [1047.97, 171.7], [1056.96, 171.68]]}, {\"category\": \"car\", \"corners_3d\": [[74.6, 230.11], [52.4, 232.9], [-40.99, 233.69], [-14.11, 230.83], [74.6, 196.82], [52.4, 197.99], [-40.99, 198.32], [-14.11, 197.12]]}]\n```" - } - ] - }, - { - "id": 3735, - "data_source": "Mono3DRefer", - "scene_name": "006957", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[579.35, 251.32], [653.35, 251.33], [666.21, 274.69], [570.18, 274.68], [579.35, 181.36], [653.35, 181.36], [666.21, 183.89], [570.18, 183.89]]}, {\"category\": \"van\", \"corners_3d\": [[481.43, 221.51], [435.42, 221.55], [460.53, 214.19], [499.59, 214.17], [481.43, 174.92], [435.42, 174.92], [460.53, 174.61], [499.59, 174.61]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[897.9, 236.02], [921.25, 237.85], [900.88, 239.75], [877.46, 237.8], [897.9, 170.85], [921.25, 170.79], [900.88, 170.73], [877.46, 170.8]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[986.63, 227.9], [966.49, 226.84], [976.77, 225.51], [996.6, 226.51], [986.63, 175.65], [966.49, 175.59], [976.77, 175.53], [996.6, 175.58]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1135.64, 232.52], [1122.0, 231.0], [1134.18, 230.98], [1148.14, 232.51], [1135.64, 165.43], [1122.0, 165.62], [1134.18, 165.62], [1148.14, 165.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1053.5, 213.59], [1044.53, 213.07], [1047.97, 212.98], [1056.96, 213.51], [1053.5, 171.68], [1044.53, 171.7], [1047.97, 171.7], [1056.96, 171.68]]}, {\"category\": \"car\", \"corners_3d\": [[74.6, 230.11], [52.4, 232.9], [-40.99, 233.69], [-14.11, 230.83], [74.6, 196.82], [52.4, 197.99], [-40.99, 198.32], [-14.11, 197.12]]}]\n```" - } - ] - }, - { - "id": 3736, - "data_source": "Mono3DRefer", - "scene_name": "006960", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[639.11, 193.21], [657.69, 193.21], [660.47, 194.61], [640.61, 194.61], [639.11, 175.87], [657.69, 175.87], [660.47, 176.08], [640.61, 176.08]]}, {\"category\": \"car\", \"corners_3d\": [[696.84, 189.28], [694.45, 188.79], [754.35, 188.8], [758.61, 189.3], [696.84, 169.8], [694.45, 169.89], [754.35, 169.89], [758.61, 169.8]]}, {\"category\": \"car\", \"corners_3d\": [[130.63, 348.35], [-12.21, 348.73], [177.38, 294.42], [275.89, 294.24], [130.63, 217.85], [-12.21, 217.95], [177.38, 204.02], [275.89, 203.98]]}, {\"category\": \"car\", \"corners_3d\": [[716.02, 248.11], [800.48, 248.12], [864.88, 273.58], [751.84, 273.57], [716.02, 175.25], [800.48, 175.25], [864.88, 176.06], [751.84, 176.06]]}, {\"category\": \"car\", \"corners_3d\": [[329.25, 275.09], [247.5, 274.98], [333.17, 251.25], [395.98, 251.31], [329.25, 201.12], [247.5, 201.09], [333.17, 194.53], [395.98, 194.54]]}, {\"category\": \"car\", \"corners_3d\": [[440.96, 238.62], [385.57, 238.63], [421.43, 228.07], [467.93, 228.07], [440.96, 180.23], [385.57, 180.23], [421.43, 179.04], [467.93, 179.04]]}, {\"category\": \"car\", \"corners_3d\": [[476.31, 225.49], [434.4, 225.43], [457.93, 218.89], [494.64, 218.93], [476.31, 184.14], [434.4, 184.13], [457.93, 182.73], [494.64, 182.74]]}, {\"category\": \"car\", \"corners_3d\": [[512.19, 210.33], [480.7, 210.31], [492.58, 207.03], [521.32, 207.04], [512.19, 183.8], [480.7, 183.79], [492.58, 182.83], [521.32, 182.84]]}]\n```" - } - ] - }, - { - "id": 3737, - "data_source": "Mono3DRefer", - "scene_name": "006967", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-158.27, 308.32], [-253.16, 305.13], [5.3, 269.93], [79.03, 271.63], [-158.27, 202.44], [-253.16, 201.75], [5.3, 194.06], [79.03, 194.43]]}, {\"category\": \"car\", \"corners_3d\": [[639.44, 209.29], [675.57, 209.44], [676.54, 214.11], [635.81, 213.92], [639.44, 177.69], [675.57, 177.71], [676.54, 178.33], [635.81, 178.3]]}, {\"category\": \"car\", \"corners_3d\": [[598.05, 201.23], [629.1, 201.36], [621.69, 204.84], [586.92, 204.67], [598.05, 171.56], [629.1, 171.55], [621.69, 171.39], [586.92, 171.4]]}, {\"category\": \"car\", \"corners_3d\": [[656.93, 198.14], [684.93, 198.23], [686.08, 201.25], [654.75, 201.14], [656.93, 171.99], [684.93, 171.99], [686.08, 171.89], [654.75, 171.89]]}, {\"category\": \"car\", \"corners_3d\": [[1028.74, 192.83], [1058.45, 193.88], [976.86, 194.43], [950.59, 193.32], [1028.74, 158.84], [1058.45, 158.1], [976.86, 157.71], [950.59, 158.49]]}, {\"category\": \"car\", \"corners_3d\": [[978.7, 188.68], [1003.0, 189.44], [931.65, 189.79], [910.18, 189.0], [978.7, 156.38], [1003.0, 155.6], [931.65, 155.23], [910.18, 156.05]]}]\n```" - } - ] - }, - { - "id": 3738, - "data_source": "Mono3DRefer", - "scene_name": "006968", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[492.96, 239.16], [558.33, 239.34], [537.88, 257.85], [454.44, 257.55], [492.96, 188.17], [558.33, 188.22], [537.88, 192.49], [454.44, 192.42]]}, {\"category\": \"car\", \"corners_3d\": [[677.03, 206.82], [711.34, 206.79], [729.01, 211.97], [689.47, 212.01], [677.03, 173.95], [711.34, 173.95], [729.01, 174.12], [689.47, 174.12]]}, {\"category\": \"car\", \"corners_3d\": [[541.97, 218.07], [581.64, 218.17], [571.63, 226.08], [525.08, 225.94], [541.97, 181.16], [581.64, 181.18], [571.63, 182.63], [525.08, 182.61]]}, {\"category\": \"car\", \"corners_3d\": [[708.62, 230.76], [765.12, 230.9], [796.55, 244.58], [726.68, 244.37], [708.62, 174.61], [765.12, 174.61], [796.55, 175.03], [726.68, 175.02]]}, {\"category\": \"car\", \"corners_3d\": [[662.91, 195.4], [687.06, 195.47], [687.96, 197.21], [661.96, 197.13], [662.91, 175.19], [687.06, 175.2], [687.96, 175.38], [661.96, 175.37]]}]\n```" - } - ] - }, - { - "id": 3739, - "data_source": "Mono3DRefer", - "scene_name": "006968", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[906.56, 287.51], [886.91, 280.3], [920.03, 280.19], [941.88, 287.39], [906.56, 137.14], [886.91, 139.39], [920.03, 139.42], [941.88, 137.18]]}]\n```" - } - ] - }, - { - "id": 3740, - "data_source": "Mono3DRefer", - "scene_name": "006968", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[724.33, 240.6], [810.78, 240.69], [946.28, 288.2], [799.1, 287.95], [724.33, 121.67], [810.78, 121.6], [946.28, 85.7], [799.1, 85.89]]}]\n```" - } - ] - }, - { - "id": 3741, - "data_source": "Mono3DRefer", - "scene_name": "006968", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[492.96, 239.16], [558.33, 239.34], [537.88, 257.85], [454.44, 257.55], [492.96, 188.17], [558.33, 188.22], [537.88, 192.49], [454.44, 192.42]]}, {\"category\": \"car\", \"corners_3d\": [[677.03, 206.82], [711.34, 206.79], [729.01, 211.97], [689.47, 212.01], [677.03, 173.95], [711.34, 173.95], [729.01, 174.12], [689.47, 174.12]]}, {\"category\": \"car\", \"corners_3d\": [[541.97, 218.07], [581.64, 218.17], [571.63, 226.08], [525.08, 225.94], [541.97, 181.16], [581.64, 181.18], [571.63, 182.63], [525.08, 182.61]]}, {\"category\": \"car\", \"corners_3d\": [[708.62, 230.76], [765.12, 230.9], [796.55, 244.58], [726.68, 244.37], [708.62, 174.61], [765.12, 174.61], [796.55, 175.03], [726.68, 175.02]]}, {\"category\": \"car\", \"corners_3d\": [[662.91, 195.4], [687.06, 195.47], [687.96, 197.21], [661.96, 197.13], [662.91, 175.19], [687.06, 175.2], [687.96, 175.38], [661.96, 175.37]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[906.56, 287.51], [886.91, 280.3], [920.03, 280.19], [941.88, 287.39], [906.56, 137.14], [886.91, 139.39], [920.03, 139.42], [941.88, 137.18]]}]\n```" - } - ] - }, - { - "id": 3742, - "data_source": "Mono3DRefer", - "scene_name": "006968", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[492.96, 239.16], [558.33, 239.34], [537.88, 257.85], [454.44, 257.55], [492.96, 188.17], [558.33, 188.22], [537.88, 192.49], [454.44, 192.42]]}, {\"category\": \"van\", \"corners_3d\": [[724.33, 240.6], [810.78, 240.69], [946.28, 288.2], [799.1, 287.95], [724.33, 121.67], [810.78, 121.6], [946.28, 85.7], [799.1, 85.89]]}, {\"category\": \"car\", \"corners_3d\": [[677.03, 206.82], [711.34, 206.79], [729.01, 211.97], [689.47, 212.01], [677.03, 173.95], [711.34, 173.95], [729.01, 174.12], [689.47, 174.12]]}, {\"category\": \"car\", \"corners_3d\": [[541.97, 218.07], [581.64, 218.17], [571.63, 226.08], [525.08, 225.94], [541.97, 181.16], [581.64, 181.18], [571.63, 182.63], [525.08, 182.61]]}, {\"category\": \"car\", \"corners_3d\": [[708.62, 230.76], [765.12, 230.9], [796.55, 244.58], [726.68, 244.37], [708.62, 174.61], [765.12, 174.61], [796.55, 175.03], [726.68, 175.02]]}, {\"category\": \"car\", \"corners_3d\": [[662.91, 195.4], [687.06, 195.47], [687.96, 197.21], [661.96, 197.13], [662.91, 175.19], [687.06, 175.2], [687.96, 175.38], [661.96, 175.37]]}]\n```" - } - ] - }, - { - "id": 3743, - "data_source": "Mono3DRefer", - "scene_name": "006968", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[724.33, 240.6], [810.78, 240.69], [946.28, 288.2], [799.1, 287.95], [724.33, 121.67], [810.78, 121.6], [946.28, 85.7], [799.1, 85.89]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[906.56, 287.51], [886.91, 280.3], [920.03, 280.19], [941.88, 287.39], [906.56, 137.14], [886.91, 139.39], [920.03, 139.42], [941.88, 137.18]]}]\n```" - } - ] - }, - { - "id": 3744, - "data_source": "Mono3DRefer", - "scene_name": "006968", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[492.96, 239.16], [558.33, 239.34], [537.88, 257.85], [454.44, 257.55], [492.96, 188.17], [558.33, 188.22], [537.88, 192.49], [454.44, 192.42]]}, {\"category\": \"van\", \"corners_3d\": [[724.33, 240.6], [810.78, 240.69], [946.28, 288.2], [799.1, 287.95], [724.33, 121.67], [810.78, 121.6], [946.28, 85.7], [799.1, 85.89]]}, {\"category\": \"car\", \"corners_3d\": [[677.03, 206.82], [711.34, 206.79], [729.01, 211.97], [689.47, 212.01], [677.03, 173.95], [711.34, 173.95], [729.01, 174.12], [689.47, 174.12]]}, {\"category\": \"car\", \"corners_3d\": [[541.97, 218.07], [581.64, 218.17], [571.63, 226.08], [525.08, 225.94], [541.97, 181.16], [581.64, 181.18], [571.63, 182.63], [525.08, 182.61]]}, {\"category\": \"car\", \"corners_3d\": [[708.62, 230.76], [765.12, 230.9], [796.55, 244.58], [726.68, 244.37], [708.62, 174.61], [765.12, 174.61], [796.55, 175.03], [726.68, 175.02]]}, {\"category\": \"car\", \"corners_3d\": [[662.91, 195.4], [687.06, 195.47], [687.96, 197.21], [661.96, 197.13], [662.91, 175.19], [687.06, 175.2], [687.96, 175.38], [661.96, 175.37]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[906.56, 287.51], [886.91, 280.3], [920.03, 280.19], [941.88, 287.39], [906.56, 137.14], [886.91, 139.39], [920.03, 139.42], [941.88, 137.18]]}]\n```" - } - ] - }, - { - "id": 3745, - "data_source": "Mono3DRefer", - "scene_name": "006975", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[552.95, 217.23], [591.76, 216.95], [603.36, 224.41], [558.03, 224.8], [552.95, 181.62], [591.76, 181.56], [603.36, 183.04], [558.03, 183.11]]}, {\"category\": \"car\", \"corners_3d\": [[591.6, 247.69], [667.17, 246.83], [706.5, 268.11], [609.45, 269.53], [591.6, 173.76], [667.17, 173.75], [706.5, 174.01], [609.45, 174.03]]}, {\"category\": \"car\", \"corners_3d\": [[247.88, 202.98], [224.17, 203.07], [248.89, 200.73], [270.7, 200.66], [247.88, 182.38], [224.17, 182.41], [248.89, 181.67], [270.7, 181.64]]}]\n```" - } - ] - }, - { - "id": 3746, - "data_source": "Mono3DRefer", - "scene_name": "006976", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[446.75, 204.12], [452.79, 202.95], [513.66, 202.94], [509.98, 204.11], [446.75, 178.66], [452.79, 178.44], [513.66, 178.44], [509.98, 178.66]]}]\n```" - } - ] - }, - { - "id": 3747, - "data_source": "Mono3DRefer", - "scene_name": "006982", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[72.2, 223.71], [9.23, 225.81], [18.73, 218.95], [73.5, 217.35], [72.2, 162.48], [9.23, 162.06], [18.73, 163.45], [73.5, 163.78]]}]\n```" - } - ] - }, - { - "id": 3748, - "data_source": "Mono3DRefer", - "scene_name": "006982", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[47.08, 343.58], [-213.23, 373.78], [-121.48, 300.1], [43.75, 287.28], [47.08, 158.05], [-213.23, 155.43], [-121.48, 161.82], [43.75, 162.93]]}, {\"category\": \"car\", \"corners_3d\": [[226.57, 230.81], [272.11, 228.8], [294.51, 234.46], [245.07, 236.9], [226.57, 188.25], [272.11, 187.71], [294.51, 189.21], [245.07, 189.86]]}, {\"category\": \"car\", \"corners_3d\": [[196.89, 226.31], [239.41, 224.59], [259.41, 229.75], [213.22, 231.83], [196.89, 188.96], [239.41, 188.44], [259.41, 189.99], [213.22, 190.62]]}, {\"category\": \"car\", \"corners_3d\": [[175.05, 214.94], [207.57, 214.12], [209.25, 217.57], [173.98, 218.54], [175.05, 189.01], [207.57, 188.69], [209.25, 190.02], [173.98, 190.39]]}, {\"category\": \"car\", \"corners_3d\": [[257.74, 241.4], [319.96, 238.22], [355.24, 248.01], [285.17, 252.24], [257.74, 187.04], [319.96, 186.38], [355.24, 188.41], [285.17, 189.28]]}]\n```" - } - ] - }, - { - "id": 3749, - "data_source": "Mono3DRefer", - "scene_name": "006982", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[47.08, 343.58], [-213.23, 373.78], [-121.48, 300.1], [43.75, 287.28], [47.08, 158.05], [-213.23, 155.43], [-121.48, 161.82], [43.75, 162.93]]}, {\"category\": \"car\", \"corners_3d\": [[226.57, 230.81], [272.11, 228.8], [294.51, 234.46], [245.07, 236.9], [226.57, 188.25], [272.11, 187.71], [294.51, 189.21], [245.07, 189.86]]}, {\"category\": \"car\", \"corners_3d\": [[196.89, 226.31], [239.41, 224.59], [259.41, 229.75], [213.22, 231.83], [196.89, 188.96], [239.41, 188.44], [259.41, 189.99], [213.22, 190.62]]}, {\"category\": \"truck\", \"corners_3d\": [[72.2, 223.71], [9.23, 225.81], [18.73, 218.95], [73.5, 217.35], [72.2, 162.48], [9.23, 162.06], [18.73, 163.45], [73.5, 163.78]]}, {\"category\": \"car\", \"corners_3d\": [[175.05, 214.94], [207.57, 214.12], [209.25, 217.57], [173.98, 218.54], [175.05, 189.01], [207.57, 188.69], [209.25, 190.02], [173.98, 190.39]]}, {\"category\": \"car\", \"corners_3d\": [[257.74, 241.4], [319.96, 238.22], [355.24, 248.01], [285.17, 252.24], [257.74, 187.04], [319.96, 186.38], [355.24, 188.41], [285.17, 189.28]]}]\n```" - } - ] - }, - { - "id": 3750, - "data_source": "Mono3DRefer", - "scene_name": "006988", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[412.27, 392.58], [176.03, 398.97], [316.64, 312.53], [461.69, 310.07], [412.27, 180.52], [176.03, 180.74], [316.64, 177.73], [461.69, 177.64]]}, {\"category\": \"car\", \"corners_3d\": [[502.33, 271.58], [414.64, 271.47], [463.14, 247.74], [529.75, 247.81], [502.33, 181.97], [414.64, 181.96], [463.14, 179.77], [529.75, 179.77]]}]\n```" - } - ] - }, - { - "id": 3751, - "data_source": "Mono3DRefer", - "scene_name": "006990", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[353.72, 277.34], [469.43, 277.02], [391.72, 341.13], [204.06, 341.97], [353.72, 187.66], [469.43, 187.62], [391.72, 196.71], [204.06, 196.82]]}, {\"category\": \"car\", \"corners_3d\": [[653.25, 211.47], [698.23, 211.36], [724.38, 219.46], [669.99, 219.64], [653.25, 168.2], [698.23, 168.21], [724.38, 167.23], [669.99, 167.21]]}, {\"category\": \"car\", \"corners_3d\": [[465.07, 225.13], [521.47, 225.09], [499.8, 238.98], [428.37, 239.04], [465.07, 172.84], [521.47, 172.84], [499.8, 172.84], [428.37, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[711.15, 263.22], [804.91, 262.88], [906.41, 305.09], [769.0, 305.84], [711.15, 168.67], [804.91, 168.69], [906.41, 166.73], [769.0, 166.7]]}, {\"category\": \"car\", \"corners_3d\": [[630.98, 191.39], [659.74, 191.45], [658.97, 193.18], [627.55, 193.11], [630.98, 167.25], [659.74, 167.23], [658.97, 166.71], [627.55, 166.73]]}]\n```" - } - ] - }, - { - "id": 3752, - "data_source": "Mono3DRefer", - "scene_name": "006990", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[775.81, 309.75], [944.14, 308.81], [2448.94, 884.06], [1597.93, 910.54], [775.81, 74.23], [944.14, 74.91], [2448.94, -339.51], [1597.93, -358.58]]}, {\"category\": \"van\", \"corners_3d\": [[688.02, 175.12], [705.76, 175.15], [687.04, 175.3], [668.37, 175.27], [688.02, 150.65], [705.76, 150.38], [687.04, 148.88], [668.37, 149.18]]}]\n```" - } - ] - }, - { - "id": 3753, - "data_source": "Mono3DRefer", - "scene_name": "006990", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[353.72, 277.34], [469.43, 277.02], [391.72, 341.13], [204.06, 341.97], [353.72, 187.66], [469.43, 187.62], [391.72, 196.71], [204.06, 196.82]]}, {\"category\": \"van\", \"corners_3d\": [[775.81, 309.75], [944.14, 308.81], [2448.94, 884.06], [1597.93, 910.54], [775.81, 74.23], [944.14, 74.91], [2448.94, -339.51], [1597.93, -358.58]]}, {\"category\": \"car\", \"corners_3d\": [[653.25, 211.47], [698.23, 211.36], [724.38, 219.46], [669.99, 219.64], [653.25, 168.2], [698.23, 168.21], [724.38, 167.23], [669.99, 167.21]]}, {\"category\": \"car\", \"corners_3d\": [[465.07, 225.13], [521.47, 225.09], [499.8, 238.98], [428.37, 239.04], [465.07, 172.84], [521.47, 172.84], [499.8, 172.84], [428.37, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[711.15, 263.22], [804.91, 262.88], [906.41, 305.09], [769.0, 305.84], [711.15, 168.67], [804.91, 168.69], [906.41, 166.73], [769.0, 166.7]]}, {\"category\": \"car\", \"corners_3d\": [[630.98, 191.39], [659.74, 191.45], [658.97, 193.18], [627.55, 193.11], [630.98, 167.25], [659.74, 167.23], [658.97, 166.71], [627.55, 166.73]]}, {\"category\": \"van\", \"corners_3d\": [[688.02, 175.12], [705.76, 175.15], [687.04, 175.3], [668.37, 175.27], [688.02, 150.65], [705.76, 150.38], [687.04, 148.88], [668.37, 149.18]]}]\n```" - } - ] - }, - { - "id": 3754, - "data_source": "Mono3DRefer", - "scene_name": "007004", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[433.28, 215.73], [371.8, 216.55], [376.21, 210.95], [428.23, 210.36], [433.28, 160.59], [371.8, 160.13], [376.21, 163.29], [428.23, 163.63]]}, {\"category\": \"car\", \"corners_3d\": [[442.36, 199.85], [401.44, 200.05], [410.66, 197.59], [446.41, 197.44], [442.36, 165.61], [401.44, 165.45], [410.66, 167.35], [446.41, 167.46]]}, {\"category\": \"car\", \"corners_3d\": [[424.42, 182.78], [439.1, 182.73], [467.78, 182.83], [453.02, 182.88], [424.42, 167.17], [439.1, 167.42], [467.78, 166.85], [453.02, 166.58]]}]\n```" - } - ] - }, - { - "id": 3755, - "data_source": "Mono3DRefer", - "scene_name": "007004", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1173.84, 225.03], [1118.8, 221.12], [1262.63, 220.4], [1330.15, 224.16], [1173.84, 154.59], [1118.8, 156.86], [1262.63, 157.28], [1330.15, 155.09]]}]\n```" - } - ] - }, - { - "id": 3756, - "data_source": "Mono3DRefer", - "scene_name": "007004", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[845.72, 222.47], [854.85, 221.66], [895.55, 223.43], [886.85, 224.31], [845.72, 155.66], [854.85, 156.14], [895.55, 155.09], [886.85, 154.57]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[860.67, 220.23], [869.31, 219.46], [912.5, 221.18], [904.36, 222.03], [860.67, 155.44], [869.31, 155.93], [912.5, 154.84], [904.36, 154.31]]}]\n```" - } - ] - }, - { - "id": 3757, - "data_source": "Mono3DRefer", - "scene_name": "007004", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[433.28, 215.73], [371.8, 216.55], [376.21, 210.95], [428.23, 210.36], [433.28, 160.59], [371.8, 160.13], [376.21, 163.29], [428.23, 163.63]]}, {\"category\": \"van\", \"corners_3d\": [[1173.84, 225.03], [1118.8, 221.12], [1262.63, 220.4], [1330.15, 224.16], [1173.84, 154.59], [1118.8, 156.86], [1262.63, 157.28], [1330.15, 155.09]]}, {\"category\": \"car\", \"corners_3d\": [[442.36, 199.85], [401.44, 200.05], [410.66, 197.59], [446.41, 197.44], [442.36, 165.61], [401.44, 165.45], [410.66, 167.35], [446.41, 167.46]]}, {\"category\": \"car\", \"corners_3d\": [[424.42, 182.78], [439.1, 182.73], [467.78, 182.83], [453.02, 182.88], [424.42, 167.17], [439.1, 167.42], [467.78, 166.85], [453.02, 166.58]]}]\n```" - } - ] - }, - { - "id": 3758, - "data_source": "Mono3DRefer", - "scene_name": "007004", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[433.28, 215.73], [371.8, 216.55], [376.21, 210.95], [428.23, 210.36], [433.28, 160.59], [371.8, 160.13], [376.21, 163.29], [428.23, 163.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[845.72, 222.47], [854.85, 221.66], [895.55, 223.43], [886.85, 224.31], [845.72, 155.66], [854.85, 156.14], [895.55, 155.09], [886.85, 154.57]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[860.67, 220.23], [869.31, 219.46], [912.5, 221.18], [904.36, 222.03], [860.67, 155.44], [869.31, 155.93], [912.5, 154.84], [904.36, 154.31]]}, {\"category\": \"car\", \"corners_3d\": [[442.36, 199.85], [401.44, 200.05], [410.66, 197.59], [446.41, 197.44], [442.36, 165.61], [401.44, 165.45], [410.66, 167.35], [446.41, 167.46]]}, {\"category\": \"car\", \"corners_3d\": [[424.42, 182.78], [439.1, 182.73], [467.78, 182.83], [453.02, 182.88], [424.42, 167.17], [439.1, 167.42], [467.78, 166.85], [453.02, 166.58]]}]\n```" - } - ] - }, - { - "id": 3759, - "data_source": "Mono3DRefer", - "scene_name": "007004", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[845.72, 222.47], [854.85, 221.66], [895.55, 223.43], [886.85, 224.31], [845.72, 155.66], [854.85, 156.14], [895.55, 155.09], [886.85, 154.57]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[860.67, 220.23], [869.31, 219.46], [912.5, 221.18], [904.36, 222.03], [860.67, 155.44], [869.31, 155.93], [912.5, 154.84], [904.36, 154.31]]}, {\"category\": \"van\", \"corners_3d\": [[1173.84, 225.03], [1118.8, 221.12], [1262.63, 220.4], [1330.15, 224.16], [1173.84, 154.59], [1118.8, 156.86], [1262.63, 157.28], [1330.15, 155.09]]}]\n```" - } - ] - }, - { - "id": 3760, - "data_source": "Mono3DRefer", - "scene_name": "007004", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[433.28, 215.73], [371.8, 216.55], [376.21, 210.95], [428.23, 210.36], [433.28, 160.59], [371.8, 160.13], [376.21, 163.29], [428.23, 163.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[845.72, 222.47], [854.85, 221.66], [895.55, 223.43], [886.85, 224.31], [845.72, 155.66], [854.85, 156.14], [895.55, 155.09], [886.85, 154.57]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[860.67, 220.23], [869.31, 219.46], [912.5, 221.18], [904.36, 222.03], [860.67, 155.44], [869.31, 155.93], [912.5, 154.84], [904.36, 154.31]]}, {\"category\": \"van\", \"corners_3d\": [[1173.84, 225.03], [1118.8, 221.12], [1262.63, 220.4], [1330.15, 224.16], [1173.84, 154.59], [1118.8, 156.86], [1262.63, 157.28], [1330.15, 155.09]]}, {\"category\": \"car\", \"corners_3d\": [[442.36, 199.85], [401.44, 200.05], [410.66, 197.59], [446.41, 197.44], [442.36, 165.61], [401.44, 165.45], [410.66, 167.35], [446.41, 167.46]]}, {\"category\": \"car\", \"corners_3d\": [[424.42, 182.78], [439.1, 182.73], [467.78, 182.83], [453.02, 182.88], [424.42, 167.17], [439.1, 167.42], [467.78, 166.85], [453.02, 166.58]]}]\n```" - } - ] - }, - { - "id": 3761, - "data_source": "Mono3DRefer", - "scene_name": "007009", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[622.62, 209.31], [583.18, 209.31], [586.62, 204.45], [620.81, 204.45], [622.62, 174.33], [583.18, 174.33], [586.62, 174.14], [620.81, 174.14]]}]\n```" - } - ] - }, - { - "id": 3762, - "data_source": "Mono3DRefer", - "scene_name": "007014", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[432.65, 242.0], [374.42, 242.01], [413.68, 230.43], [462.16, 230.43], [432.65, 180.6], [374.42, 180.6], [413.68, 179.31], [462.16, 179.31]]}, {\"category\": \"car\", \"corners_3d\": [[578.04, 199.51], [553.12, 199.49], [558.71, 197.37], [581.65, 197.39], [578.04, 179.51], [553.12, 179.51], [558.71, 178.98], [581.65, 178.98]]}, {\"category\": \"car\", \"corners_3d\": [[469.71, 225.53], [426.07, 225.47], [451.52, 218.69], [489.56, 218.73], [469.71, 182.47], [426.07, 182.46], [451.52, 181.22], [489.56, 181.23]]}, {\"category\": \"car\", \"corners_3d\": [[509.03, 210.83], [476.64, 210.81], [489.2, 207.4], [518.67, 207.41], [509.03, 183.54], [476.64, 183.53], [489.2, 182.57], [518.67, 182.57]]}, {\"category\": \"car\", \"corners_3d\": [[640.07, 193.33], [658.97, 193.33], [661.88, 194.77], [641.66, 194.76], [640.07, 175.69], [658.97, 175.69], [661.88, 175.89], [641.66, 175.89]]}, {\"category\": \"car\", \"corners_3d\": [[698.16, 190.31], [695.68, 189.77], [756.83, 189.79], [761.26, 190.33], [698.16, 170.41], [695.68, 170.49], [756.83, 170.49], [761.26, 170.41]]}, {\"category\": \"car\", \"corners_3d\": [[57.19, 374.98], [-108.42, 375.48], [133.73, 306.39], [242.57, 306.17], [57.19, 223.89], [-108.42, 224.01], [133.73, 206.57], [242.57, 206.52]]}, {\"category\": \"car\", \"corners_3d\": [[725.6, 256.23], [816.83, 256.24], [894.48, 287.56], [768.97, 287.55], [725.6, 177.53], [816.83, 177.53], [894.48, 179.28], [768.97, 179.28]]}, {\"category\": \"car\", \"corners_3d\": [[305.5, 286.33], [216.69, 286.2], [315.58, 258.15], [382.47, 258.22], [305.5, 205.94], [216.69, 205.9], [315.58, 197.72], [382.47, 197.74]]}]\n```" - } - ] - }, - { - "id": 3763, - "data_source": "Mono3DRefer", - "scene_name": "007017", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[943.28, 318.42], [913.74, 307.15], [981.55, 306.76], [1017.0, 317.96], [943.28, 144.65], [913.74, 147.58], [981.55, 147.68], [1017.0, 144.77]]}]\n```" - } - ] - }, - { - "id": 3764, - "data_source": "Mono3DRefer", - "scene_name": "007026", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[123.76, 302.17], [-19.45, 311.18], [52.0, 277.95], [158.94, 272.67], [123.76, 209.69], [-19.45, 212.26], [52.0, 202.79], [158.94, 201.29]]}, {\"category\": \"car\", \"corners_3d\": [[1331.65, 254.6], [1201.23, 250.57], [1121.4, 231.23], [1215.03, 233.47], [1331.65, 150.15], [1201.23, 151.27], [1121.4, 156.64], [1215.03, 156.02]]}, {\"category\": \"car\", \"corners_3d\": [[1117.91, 253.67], [1000.5, 250.44], [959.86, 232.54], [1048.04, 234.44], [1117.91, 160.79], [1000.5, 161.28], [959.86, 163.95], [1048.04, 163.66]]}, {\"category\": \"car\", \"corners_3d\": [[413.61, 214.84], [455.01, 214.06], [468.67, 219.69], [421.79, 220.69], [413.61, 182.28], [455.01, 182.11], [468.67, 183.37], [421.79, 183.6]]}, {\"category\": \"car\", \"corners_3d\": [[584.89, 207.19], [596.91, 205.59], [682.89, 206.85], [674.77, 208.59], [584.89, 166.26], [596.91, 166.57], [682.89, 166.32], [674.77, 165.99]]}, {\"category\": \"car\", \"corners_3d\": [[255.51, 214.59], [220.36, 215.2], [231.69, 211.6], [263.74, 211.09], [255.51, 185.15], [220.36, 185.33], [231.69, 184.27], [263.74, 184.12]]}, {\"category\": \"car\", \"corners_3d\": [[377.82, 195.6], [398.79, 195.4], [401.05, 197.25], [378.38, 197.48], [377.82, 175.47], [398.79, 175.44], [401.05, 175.66], [378.38, 175.68]]}, {\"category\": \"car\", \"corners_3d\": [[508.44, 250.51], [577.8, 247.61], [647.34, 267.3], [562.26, 271.96], [508.44, 189.9], [577.8, 189.26], [647.34, 193.58], [562.26, 194.61]]}]\n```" - } - ] - }, - { - "id": 3765, - "data_source": "Mono3DRefer", - "scene_name": "007026", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[657.18, 300.6], [779.63, 291.73], [1118.01, 379.63], [939.07, 408.04], [657.18, 150.5], [779.63, 152.06], [1118.01, 136.68], [939.07, 131.71]]}]\n```" - } - ] - }, - { - "id": 3766, - "data_source": "Mono3DRefer", - "scene_name": "007026", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[123.76, 302.17], [-19.45, 311.18], [52.0, 277.95], [158.94, 272.67], [123.76, 209.69], [-19.45, 212.26], [52.0, 202.79], [158.94, 201.29]]}, {\"category\": \"car\", \"corners_3d\": [[1331.65, 254.6], [1201.23, 250.57], [1121.4, 231.23], [1215.03, 233.47], [1331.65, 150.15], [1201.23, 151.27], [1121.4, 156.64], [1215.03, 156.02]]}, {\"category\": \"car\", \"corners_3d\": [[1117.91, 253.67], [1000.5, 250.44], [959.86, 232.54], [1048.04, 234.44], [1117.91, 160.79], [1000.5, 161.28], [959.86, 163.95], [1048.04, 163.66]]}, {\"category\": \"car\", \"corners_3d\": [[413.61, 214.84], [455.01, 214.06], [468.67, 219.69], [421.79, 220.69], [413.61, 182.28], [455.01, 182.11], [468.67, 183.37], [421.79, 183.6]]}, {\"category\": \"car\", \"corners_3d\": [[584.89, 207.19], [596.91, 205.59], [682.89, 206.85], [674.77, 208.59], [584.89, 166.26], [596.91, 166.57], [682.89, 166.32], [674.77, 165.99]]}, {\"category\": \"car\", \"corners_3d\": [[255.51, 214.59], [220.36, 215.2], [231.69, 211.6], [263.74, 211.09], [255.51, 185.15], [220.36, 185.33], [231.69, 184.27], [263.74, 184.12]]}, {\"category\": \"car\", \"corners_3d\": [[377.82, 195.6], [398.79, 195.4], [401.05, 197.25], [378.38, 197.48], [377.82, 175.47], [398.79, 175.44], [401.05, 175.66], [378.38, 175.68]]}, {\"category\": \"van\", \"corners_3d\": [[657.18, 300.6], [779.63, 291.73], [1118.01, 379.63], [939.07, 408.04], [657.18, 150.5], [779.63, 152.06], [1118.01, 136.68], [939.07, 131.71]]}, {\"category\": \"car\", \"corners_3d\": [[508.44, 250.51], [577.8, 247.61], [647.34, 267.3], [562.26, 271.96], [508.44, 189.9], [577.8, 189.26], [647.34, 193.58], [562.26, 194.61]]}]\n```" - } - ] - }, - { - "id": 3767, - "data_source": "Mono3DRefer", - "scene_name": "007034", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[546.08, 216.02], [502.53, 215.82], [529.84, 208.68], [566.23, 208.82], [546.08, 149.2], [502.53, 149.32], [529.84, 153.23], [566.23, 153.15]]}]\n```" - } - ] - }, - { - "id": 3768, - "data_source": "Mono3DRefer", - "scene_name": "007034", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[602.22, 214.37], [611.37, 213.96], [617.36, 214.4], [608.17, 214.82], [602.22, 175.4], [611.37, 175.37], [617.36, 175.4], [608.17, 175.42]]}]\n```" - } - ] - }, - { - "id": 3769, - "data_source": "Mono3DRefer", - "scene_name": "007034", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[317.43, 368.36], [162.63, 369.23], [305.56, 304.25], [408.86, 303.86], [317.43, 242.3], [162.63, 242.61], [305.56, 219.53], [408.86, 219.39]]}, {\"category\": \"car\", \"corners_3d\": [[448.55, 276.57], [358.3, 276.84], [412.12, 253.15], [481.74, 252.99], [448.55, 188.04], [358.3, 188.08], [412.12, 184.61], [481.74, 184.59]]}, {\"category\": \"car\", \"corners_3d\": [[613.0, 213.55], [646.98, 213.51], [652.53, 217.91], [614.87, 217.95], [613.0, 181.07], [646.98, 181.07], [652.53, 181.96], [614.87, 181.97]]}]\n```" - } - ] - }, - { - "id": 3770, - "data_source": "Mono3DRefer", - "scene_name": "007034", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[546.08, 216.02], [502.53, 215.82], [529.84, 208.68], [566.23, 208.82], [546.08, 149.2], [502.53, 149.32], [529.84, 153.23], [566.23, 153.15]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[602.22, 214.37], [611.37, 213.96], [617.36, 214.4], [608.17, 214.82], [602.22, 175.4], [611.37, 175.37], [617.36, 175.4], [608.17, 175.42]]}]\n```" - } - ] - }, - { - "id": 3771, - "data_source": "Mono3DRefer", - "scene_name": "007034", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[317.43, 368.36], [162.63, 369.23], [305.56, 304.25], [408.86, 303.86], [317.43, 242.3], [162.63, 242.61], [305.56, 219.53], [408.86, 219.39]]}, {\"category\": \"car\", \"corners_3d\": [[448.55, 276.57], [358.3, 276.84], [412.12, 253.15], [481.74, 252.99], [448.55, 188.04], [358.3, 188.08], [412.12, 184.61], [481.74, 184.59]]}, {\"category\": \"truck\", \"corners_3d\": [[546.08, 216.02], [502.53, 215.82], [529.84, 208.68], [566.23, 208.82], [546.08, 149.2], [502.53, 149.32], [529.84, 153.23], [566.23, 153.15]]}, {\"category\": \"car\", \"corners_3d\": [[613.0, 213.55], [646.98, 213.51], [652.53, 217.91], [614.87, 217.95], [613.0, 181.07], [646.98, 181.07], [652.53, 181.96], [614.87, 181.97]]}]\n```" - } - ] - }, - { - "id": 3772, - "data_source": "Mono3DRefer", - "scene_name": "007034", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[317.43, 368.36], [162.63, 369.23], [305.56, 304.25], [408.86, 303.86], [317.43, 242.3], [162.63, 242.61], [305.56, 219.53], [408.86, 219.39]]}, {\"category\": \"car\", \"corners_3d\": [[448.55, 276.57], [358.3, 276.84], [412.12, 253.15], [481.74, 252.99], [448.55, 188.04], [358.3, 188.08], [412.12, 184.61], [481.74, 184.59]]}, {\"category\": \"car\", \"corners_3d\": [[613.0, 213.55], [646.98, 213.51], [652.53, 217.91], [614.87, 217.95], [613.0, 181.07], [646.98, 181.07], [652.53, 181.96], [614.87, 181.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[602.22, 214.37], [611.37, 213.96], [617.36, 214.4], [608.17, 214.82], [602.22, 175.4], [611.37, 175.37], [617.36, 175.4], [608.17, 175.42]]}]\n```" - } - ] - }, - { - "id": 3773, - "data_source": "Mono3DRefer", - "scene_name": "007034", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[317.43, 368.36], [162.63, 369.23], [305.56, 304.25], [408.86, 303.86], [317.43, 242.3], [162.63, 242.61], [305.56, 219.53], [408.86, 219.39]]}, {\"category\": \"car\", \"corners_3d\": [[448.55, 276.57], [358.3, 276.84], [412.12, 253.15], [481.74, 252.99], [448.55, 188.04], [358.3, 188.08], [412.12, 184.61], [481.74, 184.59]]}, {\"category\": \"truck\", \"corners_3d\": [[546.08, 216.02], [502.53, 215.82], [529.84, 208.68], [566.23, 208.82], [546.08, 149.2], [502.53, 149.32], [529.84, 153.23], [566.23, 153.15]]}, {\"category\": \"car\", \"corners_3d\": [[613.0, 213.55], [646.98, 213.51], [652.53, 217.91], [614.87, 217.95], [613.0, 181.07], [646.98, 181.07], [652.53, 181.96], [614.87, 181.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[602.22, 214.37], [611.37, 213.96], [617.36, 214.4], [608.17, 214.82], [602.22, 175.4], [611.37, 175.37], [617.36, 175.4], [608.17, 175.42]]}]\n```" - } - ] - }, - { - "id": 3774, - "data_source": "Mono3DRefer", - "scene_name": "007037", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[551.04, 215.24], [510.22, 215.26], [521.22, 210.17], [557.13, 210.15], [551.04, 181.23], [510.22, 181.24], [521.22, 180.23], [557.13, 180.22]]}, {\"category\": \"car\", \"corners_3d\": [[347.37, 336.34], [208.45, 337.6], [329.33, 284.02], [422.67, 283.44], [347.37, 208.65], [208.45, 208.92], [329.33, 197.19], [422.67, 197.07]]}]\n```" - } - ] - }, - { - "id": 3775, - "data_source": "Mono3DRefer", - "scene_name": "007037", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[457.9, 266.42], [369.83, 266.54], [422.96, 245.08], [490.83, 245.01], [457.9, 170.72], [369.83, 170.72], [422.96, 171.21], [490.83, 171.21]]}]\n```" - } - ] - }, - { - "id": 3776, - "data_source": "Mono3DRefer", - "scene_name": "007037", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[457.9, 266.42], [369.83, 266.54], [422.96, 245.08], [490.83, 245.01], [457.9, 170.72], [369.83, 170.72], [422.96, 171.21], [490.83, 171.21]]}, {\"category\": \"car\", \"corners_3d\": [[551.04, 215.24], [510.22, 215.26], [521.22, 210.17], [557.13, 210.15], [551.04, 181.23], [510.22, 181.24], [521.22, 180.23], [557.13, 180.22]]}, {\"category\": \"car\", \"corners_3d\": [[347.37, 336.34], [208.45, 337.6], [329.33, 284.02], [422.67, 283.44], [347.37, 208.65], [208.45, 208.92], [329.33, 197.19], [422.67, 197.07]]}]\n```" - } - ] - }, - { - "id": 3777, - "data_source": "Mono3DRefer", - "scene_name": "007043", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[352.16, 276.75], [441.07, 267.12], [602.66, 293.69], [507.1, 309.98], [352.16, 170.66], [441.07, 170.86], [602.66, 170.3], [507.1, 169.96]]}]\n```" - } - ] - }, - { - "id": 3778, - "data_source": "Mono3DRefer", - "scene_name": "007043", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[761.87, 241.5], [701.54, 241.97], [674.97, 229.76], [724.85, 229.44], [761.87, 181.52], [701.54, 181.58], [674.97, 180.04], [724.85, 180.0]]}, {\"category\": \"car\", \"corners_3d\": [[659.53, 239.49], [591.48, 239.31], [598.27, 227.79], [654.51, 227.91], [659.53, 179.51], [591.48, 179.49], [598.27, 178.34], [654.51, 178.35]]}, {\"category\": \"car\", \"corners_3d\": [[275.28, 220.73], [286.49, 218.48], [373.73, 217.79], [366.93, 219.97], [275.28, 181.24], [286.49, 180.85], [373.73, 180.73], [366.93, 181.11]]}, {\"category\": \"car\", \"corners_3d\": [[-7.47, 211.6], [22.85, 210.76], [28.04, 213.53], [-4.43, 214.49], [-7.47, 188.03], [22.85, 187.7], [28.04, 188.79], [-4.43, 189.16]]}, {\"category\": \"car\", \"corners_3d\": [[208.82, 254.66], [275.39, 249.53], [344.17, 261.16], [272.07, 268.03], [208.82, 200.78], [275.39, 199.03], [344.17, 203.0], [272.07, 205.34]]}]\n```" - } - ] - }, - { - "id": 3779, - "data_source": "Mono3DRefer", - "scene_name": "007043", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[761.87, 241.5], [701.54, 241.97], [674.97, 229.76], [724.85, 229.44], [761.87, 181.52], [701.54, 181.58], [674.97, 180.04], [724.85, 180.0]]}, {\"category\": \"car\", \"corners_3d\": [[659.53, 239.49], [591.48, 239.31], [598.27, 227.79], [654.51, 227.91], [659.53, 179.51], [591.48, 179.49], [598.27, 178.34], [654.51, 178.35]]}, {\"category\": \"car\", \"corners_3d\": [[275.28, 220.73], [286.49, 218.48], [373.73, 217.79], [366.93, 219.97], [275.28, 181.24], [286.49, 180.85], [373.73, 180.73], [366.93, 181.11]]}, {\"category\": \"car\", \"corners_3d\": [[-7.47, 211.6], [22.85, 210.76], [28.04, 213.53], [-4.43, 214.49], [-7.47, 188.03], [22.85, 187.7], [28.04, 188.79], [-4.43, 189.16]]}, {\"category\": \"van\", \"corners_3d\": [[352.16, 276.75], [441.07, 267.12], [602.66, 293.69], [507.1, 309.98], [352.16, 170.66], [441.07, 170.86], [602.66, 170.3], [507.1, 169.96]]}, {\"category\": \"car\", \"corners_3d\": [[208.82, 254.66], [275.39, 249.53], [344.17, 261.16], [272.07, 268.03], [208.82, 200.78], [275.39, 199.03], [344.17, 203.0], [272.07, 205.34]]}]\n```" - } - ] - }, - { - "id": 3780, - "data_source": "Mono3DRefer", - "scene_name": "007047", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[663.15, 207.55], [695.73, 207.57], [706.12, 212.17], [669.22, 212.15], [663.15, 179.75], [695.73, 179.75], [706.12, 180.67], [669.22, 180.66]]}, {\"category\": \"car\", \"corners_3d\": [[534.3, 214.84], [529.45, 217.55], [414.72, 217.55], [426.52, 214.84], [534.3, 179.89], [529.45, 180.34], [414.72, 180.34], [426.52, 179.89]]}, {\"category\": \"car\", \"corners_3d\": [[505.19, 224.74], [493.83, 228.73], [358.05, 228.09], [378.88, 224.19], [505.19, 181.73], [493.83, 182.41], [358.05, 182.31], [378.88, 181.64]]}, {\"category\": \"car\", \"corners_3d\": [[500.21, 229.02], [484.8, 233.99], [340.41, 232.86], [367.08, 228.08], [500.21, 179.05], [484.8, 179.6], [340.41, 179.47], [367.08, 178.94]]}, {\"category\": \"car\", \"corners_3d\": [[485.17, 238.33], [469.33, 243.6], [318.57, 242.08], [345.1, 237.03], [485.17, 182.83], [469.33, 183.63], [318.57, 183.4], [345.1, 182.63]]}, {\"category\": \"car\", \"corners_3d\": [[447.97, 244.85], [430.34, 252.37], [265.63, 252.19], [298.78, 244.7], [447.97, 181.16], [430.34, 182.03], [265.63, 182.01], [298.78, 181.15]]}, {\"category\": \"car\", \"corners_3d\": [[54.37, 288.8], [124.77, 274.31], [325.6, 274.64], [283.72, 289.23], [54.37, 184.43], [124.77, 182.98], [325.6, 183.01], [283.72, 184.47]]}, {\"category\": \"car\", \"corners_3d\": [[-109.25, 326.71], [9.54, 300.72], [292.61, 299.66], [231.97, 325.18], [-109.25, 199.82], [9.54, 195.27], [292.61, 195.08], [231.97, 199.55]]}]\n```" - } - ] - }, - { - "id": 3781, - "data_source": "Mono3DRefer", - "scene_name": "007047", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[643.41, 197.31], [676.82, 197.31], [686.02, 200.68], [647.99, 200.68], [643.41, 165.23], [676.82, 165.23], [686.02, 164.18], [647.99, 164.18]]}]\n```" - } - ] - }, - { - "id": 3782, - "data_source": "Mono3DRefer", - "scene_name": "007047", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[544.1, 211.65], [541.5, 213.95], [422.79, 214.15], [432.07, 211.83], [544.1, 167.01], [541.5, 166.67], [422.79, 166.64], [432.07, 166.99]]}]\n```" - } - ] - }, - { - "id": 3783, - "data_source": "Mono3DRefer", - "scene_name": "007047", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[663.15, 207.55], [695.73, 207.57], [706.12, 212.17], [669.22, 212.15], [663.15, 179.75], [695.73, 179.75], [706.12, 180.67], [669.22, 180.66]]}, {\"category\": \"car\", \"corners_3d\": [[534.3, 214.84], [529.45, 217.55], [414.72, 217.55], [426.52, 214.84], [534.3, 179.89], [529.45, 180.34], [414.72, 180.34], [426.52, 179.89]]}, {\"category\": \"car\", \"corners_3d\": [[505.19, 224.74], [493.83, 228.73], [358.05, 228.09], [378.88, 224.19], [505.19, 181.73], [493.83, 182.41], [358.05, 182.31], [378.88, 181.64]]}, {\"category\": \"car\", \"corners_3d\": [[500.21, 229.02], [484.8, 233.99], [340.41, 232.86], [367.08, 228.08], [500.21, 179.05], [484.8, 179.6], [340.41, 179.47], [367.08, 178.94]]}, {\"category\": \"car\", \"corners_3d\": [[485.17, 238.33], [469.33, 243.6], [318.57, 242.08], [345.1, 237.03], [485.17, 182.83], [469.33, 183.63], [318.57, 183.4], [345.1, 182.63]]}, {\"category\": \"car\", \"corners_3d\": [[447.97, 244.85], [430.34, 252.37], [265.63, 252.19], [298.78, 244.7], [447.97, 181.16], [430.34, 182.03], [265.63, 182.01], [298.78, 181.15]]}, {\"category\": \"car\", \"corners_3d\": [[54.37, 288.8], [124.77, 274.31], [325.6, 274.64], [283.72, 289.23], [54.37, 184.43], [124.77, 182.98], [325.6, 183.01], [283.72, 184.47]]}, {\"category\": \"car\", \"corners_3d\": [[-109.25, 326.71], [9.54, 300.72], [292.61, 299.66], [231.97, 325.18], [-109.25, 199.82], [9.54, 195.27], [292.61, 195.08], [231.97, 199.55]]}, {\"category\": \"truck\", \"corners_3d\": [[643.41, 197.31], [676.82, 197.31], [686.02, 200.68], [647.99, 200.68], [643.41, 165.23], [676.82, 165.23], [686.02, 164.18], [647.99, 164.18]]}]\n```" - } - ] - }, - { - "id": 3784, - "data_source": "Mono3DRefer", - "scene_name": "007047", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[663.15, 207.55], [695.73, 207.57], [706.12, 212.17], [669.22, 212.15], [663.15, 179.75], [695.73, 179.75], [706.12, 180.67], [669.22, 180.66]]}, {\"category\": \"van\", \"corners_3d\": [[544.1, 211.65], [541.5, 213.95], [422.79, 214.15], [432.07, 211.83], [544.1, 167.01], [541.5, 166.67], [422.79, 166.64], [432.07, 166.99]]}, {\"category\": \"car\", \"corners_3d\": [[534.3, 214.84], [529.45, 217.55], [414.72, 217.55], [426.52, 214.84], [534.3, 179.89], [529.45, 180.34], [414.72, 180.34], [426.52, 179.89]]}, {\"category\": \"car\", \"corners_3d\": [[505.19, 224.74], [493.83, 228.73], [358.05, 228.09], [378.88, 224.19], [505.19, 181.73], [493.83, 182.41], [358.05, 182.31], [378.88, 181.64]]}, {\"category\": \"car\", \"corners_3d\": [[500.21, 229.02], [484.8, 233.99], [340.41, 232.86], [367.08, 228.08], [500.21, 179.05], [484.8, 179.6], [340.41, 179.47], [367.08, 178.94]]}, {\"category\": \"car\", \"corners_3d\": [[485.17, 238.33], [469.33, 243.6], [318.57, 242.08], [345.1, 237.03], [485.17, 182.83], [469.33, 183.63], [318.57, 183.4], [345.1, 182.63]]}, {\"category\": \"car\", \"corners_3d\": [[447.97, 244.85], [430.34, 252.37], [265.63, 252.19], [298.78, 244.7], [447.97, 181.16], [430.34, 182.03], [265.63, 182.01], [298.78, 181.15]]}, {\"category\": \"car\", \"corners_3d\": [[54.37, 288.8], [124.77, 274.31], [325.6, 274.64], [283.72, 289.23], [54.37, 184.43], [124.77, 182.98], [325.6, 183.01], [283.72, 184.47]]}, {\"category\": \"car\", \"corners_3d\": [[-109.25, 326.71], [9.54, 300.72], [292.61, 299.66], [231.97, 325.18], [-109.25, 199.82], [9.54, 195.27], [292.61, 195.08], [231.97, 199.55]]}]\n```" - } - ] - }, - { - "id": 3785, - "data_source": "Mono3DRefer", - "scene_name": "007047", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[544.1, 211.65], [541.5, 213.95], [422.79, 214.15], [432.07, 211.83], [544.1, 167.01], [541.5, 166.67], [422.79, 166.64], [432.07, 166.99]]}, {\"category\": \"truck\", \"corners_3d\": [[643.41, 197.31], [676.82, 197.31], [686.02, 200.68], [647.99, 200.68], [643.41, 165.23], [676.82, 165.23], [686.02, 164.18], [647.99, 164.18]]}]\n```" - } - ] - }, - { - "id": 3786, - "data_source": "Mono3DRefer", - "scene_name": "007047", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[663.15, 207.55], [695.73, 207.57], [706.12, 212.17], [669.22, 212.15], [663.15, 179.75], [695.73, 179.75], [706.12, 180.67], [669.22, 180.66]]}, {\"category\": \"van\", \"corners_3d\": [[544.1, 211.65], [541.5, 213.95], [422.79, 214.15], [432.07, 211.83], [544.1, 167.01], [541.5, 166.67], [422.79, 166.64], [432.07, 166.99]]}, {\"category\": \"car\", \"corners_3d\": [[534.3, 214.84], [529.45, 217.55], [414.72, 217.55], [426.52, 214.84], [534.3, 179.89], [529.45, 180.34], [414.72, 180.34], [426.52, 179.89]]}, {\"category\": \"car\", \"corners_3d\": [[505.19, 224.74], [493.83, 228.73], [358.05, 228.09], [378.88, 224.19], [505.19, 181.73], [493.83, 182.41], [358.05, 182.31], [378.88, 181.64]]}, {\"category\": \"car\", \"corners_3d\": [[500.21, 229.02], [484.8, 233.99], [340.41, 232.86], [367.08, 228.08], [500.21, 179.05], [484.8, 179.6], [340.41, 179.47], [367.08, 178.94]]}, {\"category\": \"car\", \"corners_3d\": [[485.17, 238.33], [469.33, 243.6], [318.57, 242.08], [345.1, 237.03], [485.17, 182.83], [469.33, 183.63], [318.57, 183.4], [345.1, 182.63]]}, {\"category\": \"car\", \"corners_3d\": [[447.97, 244.85], [430.34, 252.37], [265.63, 252.19], [298.78, 244.7], [447.97, 181.16], [430.34, 182.03], [265.63, 182.01], [298.78, 181.15]]}, {\"category\": \"car\", \"corners_3d\": [[54.37, 288.8], [124.77, 274.31], [325.6, 274.64], [283.72, 289.23], [54.37, 184.43], [124.77, 182.98], [325.6, 183.01], [283.72, 184.47]]}, {\"category\": \"car\", \"corners_3d\": [[-109.25, 326.71], [9.54, 300.72], [292.61, 299.66], [231.97, 325.18], [-109.25, 199.82], [9.54, 195.27], [292.61, 195.08], [231.97, 199.55]]}, {\"category\": \"truck\", \"corners_3d\": [[643.41, 197.31], [676.82, 197.31], [686.02, 200.68], [647.99, 200.68], [643.41, 165.23], [676.82, 165.23], [686.02, 164.18], [647.99, 164.18]]}]\n```" - } - ] - }, - { - "id": 3787, - "data_source": "Mono3DRefer", - "scene_name": "007058", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[442.4, 238.12], [379.74, 238.07], [418.44, 227.4], [470.87, 227.43], [442.4, 185.69], [379.74, 185.68], [418.44, 183.58], [470.87, 183.59]]}]\n```" - } - ] - }, - { - "id": 3788, - "data_source": "Mono3DRefer", - "scene_name": "007060", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[728.56, 263.91], [791.71, 275.32], [587.48, 293.2], [540.93, 277.76], [728.56, 184.64], [791.71, 186.12], [587.48, 188.43], [540.93, 186.43]]}, {\"category\": \"car\", \"corners_3d\": [[640.45, 251.89], [695.58, 261.49], [478.26, 278.31], [441.45, 264.99], [640.45, 170.06], [695.58, 169.72], [478.26, 169.13], [441.45, 169.6]]}, {\"category\": \"car\", \"corners_3d\": [[349.98, 219.38], [309.85, 220.63], [295.68, 215.17], [331.66, 214.18], [349.98, 187.34], [309.85, 187.73], [295.68, 186.03], [331.66, 185.72]]}, {\"category\": \"car\", \"corners_3d\": [[596.91, 244.77], [627.29, 250.64], [454.23, 260.72], [434.86, 253.3], [596.91, 187.55], [627.29, 188.75], [454.23, 190.81], [434.86, 189.3]]}]\n```" - } - ] - }, - { - "id": 3789, - "data_source": "Mono3DRefer", - "scene_name": "007060", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[121.02, 244.23], [45.14, 246.91], [81.23, 234.98], [144.15, 233.08], [121.02, 181.55], [45.14, 181.87], [81.23, 180.42], [144.15, 180.19]]}]\n```" - } - ] - }, - { - "id": 3790, - "data_source": "Mono3DRefer", - "scene_name": "007060", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[728.56, 263.91], [791.71, 275.32], [587.48, 293.2], [540.93, 277.76], [728.56, 184.64], [791.71, 186.12], [587.48, 188.43], [540.93, 186.43]]}, {\"category\": \"car\", \"corners_3d\": [[640.45, 251.89], [695.58, 261.49], [478.26, 278.31], [441.45, 264.99], [640.45, 170.06], [695.58, 169.72], [478.26, 169.13], [441.45, 169.6]]}, {\"category\": \"van\", \"corners_3d\": [[121.02, 244.23], [45.14, 246.91], [81.23, 234.98], [144.15, 233.08], [121.02, 181.55], [45.14, 181.87], [81.23, 180.42], [144.15, 180.19]]}, {\"category\": \"car\", \"corners_3d\": [[349.98, 219.38], [309.85, 220.63], [295.68, 215.17], [331.66, 214.18], [349.98, 187.34], [309.85, 187.73], [295.68, 186.03], [331.66, 185.72]]}, {\"category\": \"car\", \"corners_3d\": [[596.91, 244.77], [627.29, 250.64], [454.23, 260.72], [434.86, 253.3], [596.91, 187.55], [627.29, 188.75], [454.23, 190.81], [434.86, 189.3]]}]\n```" - } - ] - }, - { - "id": 3791, - "data_source": "Mono3DRefer", - "scene_name": "007063", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[656.74, 216.79], [705.35, 216.79], [727.86, 227.17], [667.76, 227.17], [656.74, 158.36], [705.35, 158.36], [727.86, 154.94], [667.76, 154.94]]}]\n```" - } - ] - }, - { - "id": 3792, - "data_source": "Mono3DRefer", - "scene_name": "007063", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[711.75, 251.67], [781.64, 251.68], [834.77, 276.1], [743.23, 276.09], [711.75, 184.04], [781.64, 184.04], [834.77, 187.5], [743.23, 187.5]]}, {\"category\": \"car\", \"corners_3d\": [[682.7, 232.68], [739.29, 232.59], [770.63, 245.63], [701.72, 245.77], [682.7, 179.6], [739.29, 179.59], [770.63, 181.06], [701.72, 181.08]]}, {\"category\": \"car\", \"corners_3d\": [[552.15, 215.55], [515.51, 215.7], [520.43, 210.79], [552.87, 210.67], [552.15, 178.67], [515.51, 178.69], [520.43, 178.02], [552.87, 178.0]]}]\n```" - } - ] - }, - { - "id": 3793, - "data_source": "Mono3DRefer", - "scene_name": "007063", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[711.75, 251.67], [781.64, 251.68], [834.77, 276.1], [743.23, 276.09], [711.75, 184.04], [781.64, 184.04], [834.77, 187.5], [743.23, 187.5]]}, {\"category\": \"car\", \"corners_3d\": [[682.7, 232.68], [739.29, 232.59], [770.63, 245.63], [701.72, 245.77], [682.7, 179.6], [739.29, 179.59], [770.63, 181.06], [701.72, 181.08]]}, {\"category\": \"van\", \"corners_3d\": [[656.74, 216.79], [705.35, 216.79], [727.86, 227.17], [667.76, 227.17], [656.74, 158.36], [705.35, 158.36], [727.86, 154.94], [667.76, 154.94]]}, {\"category\": \"car\", \"corners_3d\": [[552.15, 215.55], [515.51, 215.7], [520.43, 210.79], [552.87, 210.67], [552.15, 178.67], [515.51, 178.69], [520.43, 178.02], [552.87, 178.0]]}]\n```" - } - ] - }, - { - "id": 3794, - "data_source": "Mono3DRefer", - "scene_name": "007070", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[914.51, 222.9], [895.13, 222.93], [882.93, 220.9], [901.53, 220.87], [914.51, 162.21], [895.13, 162.21], [882.93, 162.64], [901.53, 162.64]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[551.03, 209.36], [551.67, 210.01], [538.49, 210.09], [538.08, 209.44], [551.03, 172.25], [551.67, 172.24], [538.49, 172.24], [538.08, 172.25]]}]\n```" - } - ] - }, - { - "id": 3795, - "data_source": "Mono3DRefer", - "scene_name": "007070", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[647.95, 200.75], [678.03, 200.79], [682.3, 203.85], [648.92, 203.79], [647.95, 166.56], [678.03, 166.55], [682.3, 165.86], [648.92, 165.87]]}]\n```" - } - ] - }, - { - "id": 3796, - "data_source": "Mono3DRefer", - "scene_name": "007070", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.87, 225.26], [659.24, 225.27], [668.14, 234.77], [602.73, 234.76], [603.87, 180.71], [659.24, 180.71], [668.14, 182.13], [602.73, 182.13]]}, {\"category\": \"car\", \"corners_3d\": [[1014.64, 204.83], [991.05, 203.1], [1089.47, 202.92], [1118.53, 204.64], [1014.64, 168.88], [991.05, 169.09], [1089.47, 169.12], [1118.53, 168.9]]}, {\"category\": \"car\", \"corners_3d\": [[423.49, 210.94], [393.66, 210.92], [412.6, 207.68], [439.9, 207.7], [423.49, 183.28], [393.66, 183.28], [412.6, 182.39], [439.9, 182.4]]}, {\"category\": \"car\", \"corners_3d\": [[588.74, 210.75], [625.98, 210.75], [627.97, 215.51], [586.05, 215.51], [588.74, 176.86], [625.98, 176.86], [627.97, 177.36], [586.05, 177.36]]}]\n```" - } - ] - }, - { - "id": 3797, - "data_source": "Mono3DRefer", - "scene_name": "007070", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[746.97, 206.83], [744.68, 205.99], [786.86, 206.11], [790.23, 206.95], [746.97, 171.15], [744.68, 171.19], [786.86, 171.18], [790.23, 171.14]]}]\n```" - } - ] - }, - { - "id": 3798, - "data_source": "Mono3DRefer", - "scene_name": "007070", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[914.51, 222.9], [895.13, 222.93], [882.93, 220.9], [901.53, 220.87], [914.51, 162.21], [895.13, 162.21], [882.93, 162.64], [901.53, 162.64]]}, {\"category\": \"van\", \"corners_3d\": [[647.95, 200.75], [678.03, 200.79], [682.3, 203.85], [648.92, 203.79], [647.95, 166.56], [678.03, 166.55], [682.3, 165.86], [648.92, 165.87]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[551.03, 209.36], [551.67, 210.01], [538.49, 210.09], [538.08, 209.44], [551.03, 172.25], [551.67, 172.24], [538.49, 172.24], [538.08, 172.25]]}]\n```" - } - ] - }, - { - "id": 3799, - "data_source": "Mono3DRefer", - "scene_name": "007070", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.87, 225.26], [659.24, 225.27], [668.14, 234.77], [602.73, 234.76], [603.87, 180.71], [659.24, 180.71], [668.14, 182.13], [602.73, 182.13]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[914.51, 222.9], [895.13, 222.93], [882.93, 220.9], [901.53, 220.87], [914.51, 162.21], [895.13, 162.21], [882.93, 162.64], [901.53, 162.64]]}, {\"category\": \"car\", \"corners_3d\": [[1014.64, 204.83], [991.05, 203.1], [1089.47, 202.92], [1118.53, 204.64], [1014.64, 168.88], [991.05, 169.09], [1089.47, 169.12], [1118.53, 168.9]]}, {\"category\": \"car\", \"corners_3d\": [[423.49, 210.94], [393.66, 210.92], [412.6, 207.68], [439.9, 207.7], [423.49, 183.28], [393.66, 183.28], [412.6, 182.39], [439.9, 182.4]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[551.03, 209.36], [551.67, 210.01], [538.49, 210.09], [538.08, 209.44], [551.03, 172.25], [551.67, 172.24], [538.49, 172.24], [538.08, 172.25]]}, {\"category\": \"car\", \"corners_3d\": [[588.74, 210.75], [625.98, 210.75], [627.97, 215.51], [586.05, 215.51], [588.74, 176.86], [625.98, 176.86], [627.97, 177.36], [586.05, 177.36]]}]\n```" - } - ] - }, - { - "id": 3800, - "data_source": "Mono3DRefer", - "scene_name": "007070", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[914.51, 222.9], [895.13, 222.93], [882.93, 220.9], [901.53, 220.87], [914.51, 162.21], [895.13, 162.21], [882.93, 162.64], [901.53, 162.64]]}, {\"category\": \"cyclist\", \"corners_3d\": [[746.97, 206.83], [744.68, 205.99], [786.86, 206.11], [790.23, 206.95], [746.97, 171.15], [744.68, 171.19], [786.86, 171.18], [790.23, 171.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[551.03, 209.36], [551.67, 210.01], [538.49, 210.09], [538.08, 209.44], [551.03, 172.25], [551.67, 172.24], [538.49, 172.24], [538.08, 172.25]]}]\n```" - } - ] - }, - { - "id": 3801, - "data_source": "Mono3DRefer", - "scene_name": "007070", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.87, 225.26], [659.24, 225.27], [668.14, 234.77], [602.73, 234.76], [603.87, 180.71], [659.24, 180.71], [668.14, 182.13], [602.73, 182.13]]}, {\"category\": \"car\", \"corners_3d\": [[1014.64, 204.83], [991.05, 203.1], [1089.47, 202.92], [1118.53, 204.64], [1014.64, 168.88], [991.05, 169.09], [1089.47, 169.12], [1118.53, 168.9]]}, {\"category\": \"car\", \"corners_3d\": [[423.49, 210.94], [393.66, 210.92], [412.6, 207.68], [439.9, 207.7], [423.49, 183.28], [393.66, 183.28], [412.6, 182.39], [439.9, 182.4]]}, {\"category\": \"van\", \"corners_3d\": [[647.95, 200.75], [678.03, 200.79], [682.3, 203.85], [648.92, 203.79], [647.95, 166.56], [678.03, 166.55], [682.3, 165.86], [648.92, 165.87]]}, {\"category\": \"car\", \"corners_3d\": [[588.74, 210.75], [625.98, 210.75], [627.97, 215.51], [586.05, 215.51], [588.74, 176.86], [625.98, 176.86], [627.97, 177.36], [586.05, 177.36]]}]\n```" - } - ] - }, - { - "id": 3802, - "data_source": "Mono3DRefer", - "scene_name": "007070", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[746.97, 206.83], [744.68, 205.99], [786.86, 206.11], [790.23, 206.95], [746.97, 171.15], [744.68, 171.19], [786.86, 171.18], [790.23, 171.14]]}, {\"category\": \"van\", \"corners_3d\": [[647.95, 200.75], [678.03, 200.79], [682.3, 203.85], [648.92, 203.79], [647.95, 166.56], [678.03, 166.55], [682.3, 165.86], [648.92, 165.87]]}]\n```" - } - ] - }, - { - "id": 3803, - "data_source": "Mono3DRefer", - "scene_name": "007070", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.87, 225.26], [659.24, 225.27], [668.14, 234.77], [602.73, 234.76], [603.87, 180.71], [659.24, 180.71], [668.14, 182.13], [602.73, 182.13]]}, {\"category\": \"car\", \"corners_3d\": [[1014.64, 204.83], [991.05, 203.1], [1089.47, 202.92], [1118.53, 204.64], [1014.64, 168.88], [991.05, 169.09], [1089.47, 169.12], [1118.53, 168.9]]}, {\"category\": \"car\", \"corners_3d\": [[423.49, 210.94], [393.66, 210.92], [412.6, 207.68], [439.9, 207.7], [423.49, 183.28], [393.66, 183.28], [412.6, 182.39], [439.9, 182.4]]}, {\"category\": \"cyclist\", \"corners_3d\": [[746.97, 206.83], [744.68, 205.99], [786.86, 206.11], [790.23, 206.95], [746.97, 171.15], [744.68, 171.19], [786.86, 171.18], [790.23, 171.14]]}, {\"category\": \"car\", \"corners_3d\": [[588.74, 210.75], [625.98, 210.75], [627.97, 215.51], [586.05, 215.51], [588.74, 176.86], [625.98, 176.86], [627.97, 177.36], [586.05, 177.36]]}]\n```" - } - ] - }, - { - "id": 3804, - "data_source": "Mono3DRefer", - "scene_name": "007070", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.87, 225.26], [659.24, 225.27], [668.14, 234.77], [602.73, 234.76], [603.87, 180.71], [659.24, 180.71], [668.14, 182.13], [602.73, 182.13]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[914.51, 222.9], [895.13, 222.93], [882.93, 220.9], [901.53, 220.87], [914.51, 162.21], [895.13, 162.21], [882.93, 162.64], [901.53, 162.64]]}, {\"category\": \"car\", \"corners_3d\": [[1014.64, 204.83], [991.05, 203.1], [1089.47, 202.92], [1118.53, 204.64], [1014.64, 168.88], [991.05, 169.09], [1089.47, 169.12], [1118.53, 168.9]]}, {\"category\": \"car\", \"corners_3d\": [[423.49, 210.94], [393.66, 210.92], [412.6, 207.68], [439.9, 207.7], [423.49, 183.28], [393.66, 183.28], [412.6, 182.39], [439.9, 182.4]]}, {\"category\": \"van\", \"corners_3d\": [[647.95, 200.75], [678.03, 200.79], [682.3, 203.85], [648.92, 203.79], [647.95, 166.56], [678.03, 166.55], [682.3, 165.86], [648.92, 165.87]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[551.03, 209.36], [551.67, 210.01], [538.49, 210.09], [538.08, 209.44], [551.03, 172.25], [551.67, 172.24], [538.49, 172.24], [538.08, 172.25]]}, {\"category\": \"car\", \"corners_3d\": [[588.74, 210.75], [625.98, 210.75], [627.97, 215.51], [586.05, 215.51], [588.74, 176.86], [625.98, 176.86], [627.97, 177.36], [586.05, 177.36]]}]\n```" - } - ] - }, - { - "id": 3805, - "data_source": "Mono3DRefer", - "scene_name": "007070", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[914.51, 222.9], [895.13, 222.93], [882.93, 220.9], [901.53, 220.87], [914.51, 162.21], [895.13, 162.21], [882.93, 162.64], [901.53, 162.64]]}, {\"category\": \"cyclist\", \"corners_3d\": [[746.97, 206.83], [744.68, 205.99], [786.86, 206.11], [790.23, 206.95], [746.97, 171.15], [744.68, 171.19], [786.86, 171.18], [790.23, 171.14]]}, {\"category\": \"van\", \"corners_3d\": [[647.95, 200.75], [678.03, 200.79], [682.3, 203.85], [648.92, 203.79], [647.95, 166.56], [678.03, 166.55], [682.3, 165.86], [648.92, 165.87]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[551.03, 209.36], [551.67, 210.01], [538.49, 210.09], [538.08, 209.44], [551.03, 172.25], [551.67, 172.24], [538.49, 172.24], [538.08, 172.25]]}]\n```" - } - ] - }, - { - "id": 3806, - "data_source": "Mono3DRefer", - "scene_name": "007070", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.87, 225.26], [659.24, 225.27], [668.14, 234.77], [602.73, 234.76], [603.87, 180.71], [659.24, 180.71], [668.14, 182.13], [602.73, 182.13]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[914.51, 222.9], [895.13, 222.93], [882.93, 220.9], [901.53, 220.87], [914.51, 162.21], [895.13, 162.21], [882.93, 162.64], [901.53, 162.64]]}, {\"category\": \"car\", \"corners_3d\": [[1014.64, 204.83], [991.05, 203.1], [1089.47, 202.92], [1118.53, 204.64], [1014.64, 168.88], [991.05, 169.09], [1089.47, 169.12], [1118.53, 168.9]]}, {\"category\": \"car\", \"corners_3d\": [[423.49, 210.94], [393.66, 210.92], [412.6, 207.68], [439.9, 207.7], [423.49, 183.28], [393.66, 183.28], [412.6, 182.39], [439.9, 182.4]]}, {\"category\": \"cyclist\", \"corners_3d\": [[746.97, 206.83], [744.68, 205.99], [786.86, 206.11], [790.23, 206.95], [746.97, 171.15], [744.68, 171.19], [786.86, 171.18], [790.23, 171.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[551.03, 209.36], [551.67, 210.01], [538.49, 210.09], [538.08, 209.44], [551.03, 172.25], [551.67, 172.24], [538.49, 172.24], [538.08, 172.25]]}, {\"category\": \"car\", \"corners_3d\": [[588.74, 210.75], [625.98, 210.75], [627.97, 215.51], [586.05, 215.51], [588.74, 176.86], [625.98, 176.86], [627.97, 177.36], [586.05, 177.36]]}]\n```" - } - ] - }, - { - "id": 3807, - "data_source": "Mono3DRefer", - "scene_name": "007070", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.87, 225.26], [659.24, 225.27], [668.14, 234.77], [602.73, 234.76], [603.87, 180.71], [659.24, 180.71], [668.14, 182.13], [602.73, 182.13]]}, {\"category\": \"car\", \"corners_3d\": [[1014.64, 204.83], [991.05, 203.1], [1089.47, 202.92], [1118.53, 204.64], [1014.64, 168.88], [991.05, 169.09], [1089.47, 169.12], [1118.53, 168.9]]}, {\"category\": \"car\", \"corners_3d\": [[423.49, 210.94], [393.66, 210.92], [412.6, 207.68], [439.9, 207.7], [423.49, 183.28], [393.66, 183.28], [412.6, 182.39], [439.9, 182.4]]}, {\"category\": \"cyclist\", \"corners_3d\": [[746.97, 206.83], [744.68, 205.99], [786.86, 206.11], [790.23, 206.95], [746.97, 171.15], [744.68, 171.19], [786.86, 171.18], [790.23, 171.14]]}, {\"category\": \"van\", \"corners_3d\": [[647.95, 200.75], [678.03, 200.79], [682.3, 203.85], [648.92, 203.79], [647.95, 166.56], [678.03, 166.55], [682.3, 165.86], [648.92, 165.87]]}, {\"category\": \"car\", \"corners_3d\": [[588.74, 210.75], [625.98, 210.75], [627.97, 215.51], [586.05, 215.51], [588.74, 176.86], [625.98, 176.86], [627.97, 177.36], [586.05, 177.36]]}]\n```" - } - ] - }, - { - "id": 3808, - "data_source": "Mono3DRefer", - "scene_name": "007070", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van, car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[603.87, 225.26], [659.24, 225.27], [668.14, 234.77], [602.73, 234.76], [603.87, 180.71], [659.24, 180.71], [668.14, 182.13], [602.73, 182.13]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[914.51, 222.9], [895.13, 222.93], [882.93, 220.9], [901.53, 220.87], [914.51, 162.21], [895.13, 162.21], [882.93, 162.64], [901.53, 162.64]]}, {\"category\": \"car\", \"corners_3d\": [[1014.64, 204.83], [991.05, 203.1], [1089.47, 202.92], [1118.53, 204.64], [1014.64, 168.88], [991.05, 169.09], [1089.47, 169.12], [1118.53, 168.9]]}, {\"category\": \"car\", \"corners_3d\": [[423.49, 210.94], [393.66, 210.92], [412.6, 207.68], [439.9, 207.7], [423.49, 183.28], [393.66, 183.28], [412.6, 182.39], [439.9, 182.4]]}, {\"category\": \"cyclist\", \"corners_3d\": [[746.97, 206.83], [744.68, 205.99], [786.86, 206.11], [790.23, 206.95], [746.97, 171.15], [744.68, 171.19], [786.86, 171.18], [790.23, 171.14]]}, {\"category\": \"van\", \"corners_3d\": [[647.95, 200.75], [678.03, 200.79], [682.3, 203.85], [648.92, 203.79], [647.95, 166.56], [678.03, 166.55], [682.3, 165.86], [648.92, 165.87]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[551.03, 209.36], [551.67, 210.01], [538.49, 210.09], [538.08, 209.44], [551.03, 172.25], [551.67, 172.24], [538.49, 172.24], [538.08, 172.25]]}, {\"category\": \"car\", \"corners_3d\": [[588.74, 210.75], [625.98, 210.75], [627.97, 215.51], [586.05, 215.51], [588.74, 176.86], [625.98, 176.86], [627.97, 177.36], [586.05, 177.36]]}]\n```" - } - ] - }, - { - "id": 3809, - "data_source": "Mono3DRefer", - "scene_name": "007076", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[474.35, 244.34], [401.07, 244.49], [441.2, 229.63], [499.23, 229.54], [474.35, 176.89], [401.07, 176.89], [441.2, 176.06], [499.23, 176.05]]}]\n```" - } - ] - }, - { - "id": 3810, - "data_source": "Mono3DRefer", - "scene_name": "007082", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[647.69, 214.26], [692.59, 214.16], [715.72, 222.74], [661.54, 222.89], [647.69, 173.99], [692.59, 173.98], [715.72, 174.22], [661.54, 174.22]]}]\n```" - } - ] - }, - { - "id": 3811, - "data_source": "Mono3DRefer", - "scene_name": "007090", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[451.71, 219.73], [438.75, 222.65], [322.17, 222.07], [341.74, 219.22], [451.71, 185.45], [438.75, 186.23], [322.17, 186.08], [341.74, 185.31]]}]\n```" - } - ] - }, - { - "id": 3812, - "data_source": "Mono3DRefer", - "scene_name": "007091", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[939.72, 324.01], [798.52, 323.45], [741.92, 274.77], [837.23, 275.03], [939.72, 179.94], [798.52, 179.91], [741.92, 177.63], [837.23, 177.64]]}, {\"category\": \"car\", \"corners_3d\": [[503.44, 262.5], [420.92, 262.2], [469.9, 241.34], [533.22, 241.51], [503.44, 190.27], [420.92, 190.21], [469.9, 186.16], [533.22, 186.19]]}, {\"category\": \"car\", \"corners_3d\": [[689.39, 240.12], [750.35, 239.9], [790.84, 255.95], [715.39, 256.3], [689.39, 183.28], [750.35, 183.25], [790.84, 185.73], [715.39, 185.79]]}, {\"category\": \"car\", \"corners_3d\": [[740.94, 235.68], [682.01, 235.63], [669.97, 224.15], [718.12, 224.18], [740.94, 183.97], [682.01, 183.96], [669.97, 181.93], [718.12, 181.94]]}, {\"category\": \"car\", \"corners_3d\": [[562.38, 217.58], [525.98, 217.63], [532.67, 213.26], [565.51, 213.22], [562.38, 182.47], [525.98, 182.48], [532.67, 181.54], [565.51, 181.53]]}, {\"category\": \"car\", \"corners_3d\": [[699.57, 211.3], [666.1, 211.25], [662.25, 207.08], [692.08, 207.12], [699.57, 182.3], [666.1, 182.29], [662.25, 181.27], [692.08, 181.28]]}]\n```" - } - ] - }, - { - "id": 3813, - "data_source": "Mono3DRefer", - "scene_name": "007093", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[203.43, 485.35], [-51.05, 478.59], [282.54, 333.59], [418.78, 335.44], [203.43, 235.64], [-51.05, 234.28], [282.54, 205.15], [418.78, 205.52]]}, {\"category\": \"car\", \"corners_3d\": [[463.1, 308.33], [373.25, 305.16], [462.9, 272.84], [532.02, 274.64], [463.1, 202.49], [373.25, 201.8], [462.9, 194.73], [532.02, 195.12]]}, {\"category\": \"car\", \"corners_3d\": [[577.79, 265.32], [515.25, 263.35], [560.88, 248.07], [613.5, 249.43], [577.79, 200.58], [515.25, 199.99], [560.88, 195.41], [613.5, 195.82]]}, {\"category\": \"car\", \"corners_3d\": [[783.24, 203.27], [810.32, 203.78], [794.01, 206.81], [764.62, 206.19], [783.24, 176.35], [810.32, 176.41], [794.01, 176.75], [764.62, 176.68]]}, {\"category\": \"car\", \"corners_3d\": [[829.56, 196.22], [806.06, 195.9], [816.04, 194.46], [838.18, 194.75], [829.56, 174.22], [806.06, 174.2], [816.04, 174.12], [838.18, 174.14]]}]\n```" - } - ] - }, - { - "id": 3814, - "data_source": "Mono3DRefer", - "scene_name": "007095", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[988.87, 468.06], [1272.94, 478.16], [4065.09, 1904.43], [2215.91, 1623.07], [988.87, 187.98], [1272.94, 188.5], [4065.09, 261.57], [2215.91, 247.16]]}, {\"category\": \"car\", \"corners_3d\": [[559.33, 228.05], [500.43, 227.92], [523.03, 218.37], [571.73, 218.46], [559.33, 175.73], [500.43, 175.72], [523.03, 175.22], [571.73, 175.23]]}, {\"category\": \"car\", \"corners_3d\": [[739.65, 214.91], [701.69, 214.8], [695.26, 210.03], [728.87, 210.11], [739.65, 176.19], [701.69, 176.18], [695.26, 175.8], [728.87, 175.81]]}, {\"category\": \"car\", \"corners_3d\": [[606.26, 204.57], [573.88, 204.43], [585.32, 201.07], [614.28, 201.18], [606.26, 176.18], [573.88, 176.16], [585.32, 175.81], [614.28, 175.82]]}]\n```" - } - ] - }, - { - "id": 3815, - "data_source": "Mono3DRefer", - "scene_name": "007097", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[43.44, 368.21], [-192.44, 369.47], [139.01, 289.38], [273.5, 288.97], [43.44, 142.3], [-192.44, 142.04], [139.01, 158.86], [273.5, 158.95]]}]\n```" - } - ] - }, - { - "id": 3816, - "data_source": "Mono3DRefer", - "scene_name": "007097", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[488.78, 230.7], [546.18, 230.82], [526.68, 243.42], [454.71, 243.22], [488.78, 177.48], [546.18, 177.47], [526.68, 176.43], [454.71, 176.45]]}, {\"category\": \"car\", \"corners_3d\": [[529.81, 209.96], [566.67, 209.96], [562.34, 213.55], [520.84, 213.54], [529.81, 176.21], [566.67, 176.21], [562.34, 175.54], [520.84, 175.54]]}, {\"category\": \"car\", \"corners_3d\": [[605.21, 202.64], [639.56, 202.68], [640.97, 205.71], [601.71, 205.66], [605.21, 171.42], [639.56, 171.4], [640.97, 169.95], [601.71, 169.97]]}, {\"category\": \"car\", \"corners_3d\": [[571.23, 194.54], [596.82, 194.56], [593.88, 195.72], [566.02, 195.7], [571.23, 169.41], [596.82, 169.39], [593.88, 168.31], [566.02, 168.33]]}]\n```" - } - ] - }, - { - "id": 3817, - "data_source": "Mono3DRefer", - "scene_name": "007097", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[488.78, 230.7], [546.18, 230.82], [526.68, 243.42], [454.71, 243.22], [488.78, 177.48], [546.18, 177.47], [526.68, 176.43], [454.71, 176.45]]}, {\"category\": \"car\", \"corners_3d\": [[529.81, 209.96], [566.67, 209.96], [562.34, 213.55], [520.84, 213.54], [529.81, 176.21], [566.67, 176.21], [562.34, 175.54], [520.84, 175.54]]}, {\"category\": \"car\", \"corners_3d\": [[605.21, 202.64], [639.56, 202.68], [640.97, 205.71], [601.71, 205.66], [605.21, 171.42], [639.56, 171.4], [640.97, 169.95], [601.71, 169.97]]}, {\"category\": \"van\", \"corners_3d\": [[43.44, 368.21], [-192.44, 369.47], [139.01, 289.38], [273.5, 288.97], [43.44, 142.3], [-192.44, 142.04], [139.01, 158.86], [273.5, 158.95]]}, {\"category\": \"car\", \"corners_3d\": [[571.23, 194.54], [596.82, 194.56], [593.88, 195.72], [566.02, 195.7], [571.23, 169.41], [596.82, 169.39], [593.88, 168.31], [566.02, 168.33]]}]\n```" - } - ] - }, - { - "id": 3818, - "data_source": "Mono3DRefer", - "scene_name": "007123", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[501.01, 218.38], [516.09, 217.91], [526.33, 219.5], [510.84, 220.0], [501.01, 173.56], [516.09, 173.55], [526.33, 173.58], [510.84, 173.59]]}]\n```" - } - ] - }, - { - "id": 3819, - "data_source": "Mono3DRefer", - "scene_name": "007123", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[254.27, 401.86], [74.17, 499.97], [-616.36, 469.69], [-254.97, 386.6], [254.27, 221.24], [74.17, 241.96], [-616.36, 235.57], [-254.97, 218.01]]}, {\"category\": \"car\", \"corners_3d\": [[1700.52, 423.1], [2330.26, 566.61], [1250.7, 572.64], [1012.11, 425.53], [1700.52, 147.45], [2330.26, 132.88], [1250.7, 132.27], [1012.11, 147.2]]}, {\"category\": \"car\", \"corners_3d\": [[237.87, 260.97], [282.76, 252.08], [465.38, 254.23], [440.23, 263.62], [237.87, 189.39], [282.76, 187.72], [465.38, 188.13], [440.23, 189.89]]}, {\"category\": \"car\", \"corners_3d\": [[340.26, 240.54], [355.55, 235.25], [505.19, 233.98], [502.64, 239.04], [340.26, 189.77], [355.55, 188.44], [505.19, 188.13], [502.64, 189.39]]}, {\"category\": \"car\", \"corners_3d\": [[1046.85, 273.88], [1115.37, 289.97], [881.68, 289.59], [845.37, 273.6], [1046.85, 177.22], [1115.37, 177.92], [881.68, 177.9], [845.37, 177.21]]}, {\"category\": \"car\", \"corners_3d\": [[1203.86, 296.19], [1339.17, 324.62], [947.72, 323.8], [886.0, 295.65], [1203.86, 177.54], [1339.17, 178.62], [947.72, 178.59], [886.0, 177.52]]}]\n```" - } - ] - }, - { - "id": 3820, - "data_source": "Mono3DRefer", - "scene_name": "007123", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[254.27, 401.86], [74.17, 499.97], [-616.36, 469.69], [-254.97, 386.6], [254.27, 221.24], [74.17, 241.96], [-616.36, 235.57], [-254.97, 218.01]]}, {\"category\": \"car\", \"corners_3d\": [[1700.52, 423.1], [2330.26, 566.61], [1250.7, 572.64], [1012.11, 425.53], [1700.52, 147.45], [2330.26, 132.88], [1250.7, 132.27], [1012.11, 147.2]]}, {\"category\": \"car\", \"corners_3d\": [[237.87, 260.97], [282.76, 252.08], [465.38, 254.23], [440.23, 263.62], [237.87, 189.39], [282.76, 187.72], [465.38, 188.13], [440.23, 189.89]]}, {\"category\": \"car\", \"corners_3d\": [[340.26, 240.54], [355.55, 235.25], [505.19, 233.98], [502.64, 239.04], [340.26, 189.77], [355.55, 188.44], [505.19, 188.13], [502.64, 189.39]]}, {\"category\": \"car\", \"corners_3d\": [[1046.85, 273.88], [1115.37, 289.97], [881.68, 289.59], [845.37, 273.6], [1046.85, 177.22], [1115.37, 177.92], [881.68, 177.9], [845.37, 177.21]]}, {\"category\": \"car\", \"corners_3d\": [[1203.86, 296.19], [1339.17, 324.62], [947.72, 323.8], [886.0, 295.65], [1203.86, 177.54], [1339.17, 178.62], [947.72, 178.59], [886.0, 177.52]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[501.01, 218.38], [516.09, 217.91], [526.33, 219.5], [510.84, 220.0], [501.01, 173.56], [516.09, 173.55], [526.33, 173.58], [510.84, 173.59]]}]\n```" - } - ] - }, - { - "id": 3821, - "data_source": "Mono3DRefer", - "scene_name": "007126", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[860.5, 297.51], [847.85, 291.71], [932.0, 291.73], [949.05, 297.53], [860.5, 157.42], [847.85, 158.57], [932.0, 158.56], [949.05, 157.42]]}]\n```" - } - ] - }, - { - "id": 3822, - "data_source": "Mono3DRefer", - "scene_name": "007136", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[558.21, 194.74], [571.08, 195.46], [370.98, 196.95], [368.19, 196.07], [558.21, 142.3], [571.08, 140.17], [370.98, 135.77], [368.19, 138.37]]}]\n```" - } - ] - }, - { - "id": 3823, - "data_source": "Mono3DRefer", - "scene_name": "007136", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[870.72, 355.84], [1065.84, 354.93], [1770.55, 609.92], [1294.01, 615.56], [870.72, 143.2], [1065.84, 143.4], [1770.55, 87.37], [1294.01, 86.13]]}]\n```" - } - ] - }, - { - "id": 3824, - "data_source": "Mono3DRefer", - "scene_name": "007136", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[376.61, 322.86], [441.33, 322.87], [424.27, 337.98], [352.63, 337.97], [376.61, 200.79], [441.33, 200.8], [424.27, 202.86], [352.63, 202.85]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[317.97, 320.74], [380.71, 320.75], [356.88, 335.82], [287.34, 335.81], [317.97, 200.76], [380.71, 200.76], [356.88, 202.84], [287.34, 202.84]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[395.72, 317.55], [344.12, 317.26], [364.32, 307.37], [412.2, 307.62], [395.72, 203.49], [344.12, 203.44], [364.32, 201.84], [412.2, 201.88]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[393.17, 292.58], [433.04, 292.65], [420.5, 300.6], [377.8, 300.53], [393.17, 198.84], [433.04, 198.85], [420.5, 200.1], [377.8, 200.08]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[441.21, 284.58], [401.31, 284.53], [412.52, 278.91], [450.25, 278.96], [441.21, 192.64], [401.31, 192.63], [412.52, 192.03], [450.25, 192.03]]}]\n```" - } - ] - }, - { - "id": 3825, - "data_source": "Mono3DRefer", - "scene_name": "007136", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[148.05, 327.19], [212.27, 324.86], [193.44, 335.96], [123.84, 338.67], [148.05, 159.23], [212.27, 159.58], [193.44, 157.88], [123.84, 157.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[190.73, 323.08], [141.46, 322.79], [170.12, 314.35], [216.5, 314.61], [190.73, 159.85], [141.46, 159.9], [170.12, 161.19], [216.5, 161.15]]}]\n```" - } - ] - }, - { - "id": 3826, - "data_source": "Mono3DRefer", - "scene_name": "007136", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[870.72, 355.84], [1065.84, 354.93], [1770.55, 609.92], [1294.01, 615.56], [870.72, 143.2], [1065.84, 143.4], [1770.55, 87.37], [1294.01, 86.13]]}, {\"category\": \"tram\", \"corners_3d\": [[558.21, 194.74], [571.08, 195.46], [370.98, 196.95], [368.19, 196.07], [558.21, 142.3], [571.08, 140.17], [370.98, 135.77], [368.19, 138.37]]}]\n```" - } - ] - }, - { - "id": 3827, - "data_source": "Mono3DRefer", - "scene_name": "007136", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[558.21, 194.74], [571.08, 195.46], [370.98, 196.95], [368.19, 196.07], [558.21, 142.3], [571.08, 140.17], [370.98, 135.77], [368.19, 138.37]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[376.61, 322.86], [441.33, 322.87], [424.27, 337.98], [352.63, 337.97], [376.61, 200.79], [441.33, 200.8], [424.27, 202.86], [352.63, 202.85]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[317.97, 320.74], [380.71, 320.75], [356.88, 335.82], [287.34, 335.81], [317.97, 200.76], [380.71, 200.76], [356.88, 202.84], [287.34, 202.84]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[395.72, 317.55], [344.12, 317.26], [364.32, 307.37], [412.2, 307.62], [395.72, 203.49], [344.12, 203.44], [364.32, 201.84], [412.2, 201.88]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[393.17, 292.58], [433.04, 292.65], [420.5, 300.6], [377.8, 300.53], [393.17, 198.84], [433.04, 198.85], [420.5, 200.1], [377.8, 200.08]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[441.21, 284.58], [401.31, 284.53], [412.52, 278.91], [450.25, 278.96], [441.21, 192.64], [401.31, 192.63], [412.52, 192.03], [450.25, 192.03]]}]\n```" - } - ] - }, - { - "id": 3828, - "data_source": "Mono3DRefer", - "scene_name": "007136", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[558.21, 194.74], [571.08, 195.46], [370.98, 196.95], [368.19, 196.07], [558.21, 142.3], [571.08, 140.17], [370.98, 135.77], [368.19, 138.37]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[148.05, 327.19], [212.27, 324.86], [193.44, 335.96], [123.84, 338.67], [148.05, 159.23], [212.27, 159.58], [193.44, 157.88], [123.84, 157.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[190.73, 323.08], [141.46, 322.79], [170.12, 314.35], [216.5, 314.61], [190.73, 159.85], [141.46, 159.9], [170.12, 161.19], [216.5, 161.15]]}]\n```" - } - ] - }, - { - "id": 3829, - "data_source": "Mono3DRefer", - "scene_name": "007136", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[870.72, 355.84], [1065.84, 354.93], [1770.55, 609.92], [1294.01, 615.56], [870.72, 143.2], [1065.84, 143.4], [1770.55, 87.37], [1294.01, 86.13]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[376.61, 322.86], [441.33, 322.87], [424.27, 337.98], [352.63, 337.97], [376.61, 200.79], [441.33, 200.8], [424.27, 202.86], [352.63, 202.85]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[317.97, 320.74], [380.71, 320.75], [356.88, 335.82], [287.34, 335.81], [317.97, 200.76], [380.71, 200.76], [356.88, 202.84], [287.34, 202.84]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[395.72, 317.55], [344.12, 317.26], [364.32, 307.37], [412.2, 307.62], [395.72, 203.49], [344.12, 203.44], [364.32, 201.84], [412.2, 201.88]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[393.17, 292.58], [433.04, 292.65], [420.5, 300.6], [377.8, 300.53], [393.17, 198.84], [433.04, 198.85], [420.5, 200.1], [377.8, 200.08]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[441.21, 284.58], [401.31, 284.53], [412.52, 278.91], [450.25, 278.96], [441.21, 192.64], [401.31, 192.63], [412.52, 192.03], [450.25, 192.03]]}]\n```" - } - ] - }, - { - "id": 3830, - "data_source": "Mono3DRefer", - "scene_name": "007136", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[870.72, 355.84], [1065.84, 354.93], [1770.55, 609.92], [1294.01, 615.56], [870.72, 143.2], [1065.84, 143.4], [1770.55, 87.37], [1294.01, 86.13]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[148.05, 327.19], [212.27, 324.86], [193.44, 335.96], [123.84, 338.67], [148.05, 159.23], [212.27, 159.58], [193.44, 157.88], [123.84, 157.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[190.73, 323.08], [141.46, 322.79], [170.12, 314.35], [216.5, 314.61], [190.73, 159.85], [141.46, 159.9], [170.12, 161.19], [216.5, 161.15]]}]\n```" - } - ] - }, - { - "id": 3831, - "data_source": "Mono3DRefer", - "scene_name": "007136", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[376.61, 322.86], [441.33, 322.87], [424.27, 337.98], [352.63, 337.97], [376.61, 200.79], [441.33, 200.8], [424.27, 202.86], [352.63, 202.85]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[317.97, 320.74], [380.71, 320.75], [356.88, 335.82], [287.34, 335.81], [317.97, 200.76], [380.71, 200.76], [356.88, 202.84], [287.34, 202.84]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[395.72, 317.55], [344.12, 317.26], [364.32, 307.37], [412.2, 307.62], [395.72, 203.49], [344.12, 203.44], [364.32, 201.84], [412.2, 201.88]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[393.17, 292.58], [433.04, 292.65], [420.5, 300.6], [377.8, 300.53], [393.17, 198.84], [433.04, 198.85], [420.5, 200.1], [377.8, 200.08]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[441.21, 284.58], [401.31, 284.53], [412.52, 278.91], [450.25, 278.96], [441.21, 192.64], [401.31, 192.63], [412.52, 192.03], [450.25, 192.03]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[148.05, 327.19], [212.27, 324.86], [193.44, 335.96], [123.84, 338.67], [148.05, 159.23], [212.27, 159.58], [193.44, 157.88], [123.84, 157.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[190.73, 323.08], [141.46, 322.79], [170.12, 314.35], [216.5, 314.61], [190.73, 159.85], [141.46, 159.9], [170.12, 161.19], [216.5, 161.15]]}]\n```" - } - ] - }, - { - "id": 3832, - "data_source": "Mono3DRefer", - "scene_name": "007136", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, van, person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[870.72, 355.84], [1065.84, 354.93], [1770.55, 609.92], [1294.01, 615.56], [870.72, 143.2], [1065.84, 143.4], [1770.55, 87.37], [1294.01, 86.13]]}, {\"category\": \"tram\", \"corners_3d\": [[558.21, 194.74], [571.08, 195.46], [370.98, 196.95], [368.19, 196.07], [558.21, 142.3], [571.08, 140.17], [370.98, 135.77], [368.19, 138.37]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[376.61, 322.86], [441.33, 322.87], [424.27, 337.98], [352.63, 337.97], [376.61, 200.79], [441.33, 200.8], [424.27, 202.86], [352.63, 202.85]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[317.97, 320.74], [380.71, 320.75], [356.88, 335.82], [287.34, 335.81], [317.97, 200.76], [380.71, 200.76], [356.88, 202.84], [287.34, 202.84]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[395.72, 317.55], [344.12, 317.26], [364.32, 307.37], [412.2, 307.62], [395.72, 203.49], [344.12, 203.44], [364.32, 201.84], [412.2, 201.88]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[393.17, 292.58], [433.04, 292.65], [420.5, 300.6], [377.8, 300.53], [393.17, 198.84], [433.04, 198.85], [420.5, 200.1], [377.8, 200.08]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[441.21, 284.58], [401.31, 284.53], [412.52, 278.91], [450.25, 278.96], [441.21, 192.64], [401.31, 192.63], [412.52, 192.03], [450.25, 192.03]]}]\n```" - } - ] - }, - { - "id": 3833, - "data_source": "Mono3DRefer", - "scene_name": "007136", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[870.72, 355.84], [1065.84, 354.93], [1770.55, 609.92], [1294.01, 615.56], [870.72, 143.2], [1065.84, 143.4], [1770.55, 87.37], [1294.01, 86.13]]}, {\"category\": \"tram\", \"corners_3d\": [[558.21, 194.74], [571.08, 195.46], [370.98, 196.95], [368.19, 196.07], [558.21, 142.3], [571.08, 140.17], [370.98, 135.77], [368.19, 138.37]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[148.05, 327.19], [212.27, 324.86], [193.44, 335.96], [123.84, 338.67], [148.05, 159.23], [212.27, 159.58], [193.44, 157.88], [123.84, 157.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[190.73, 323.08], [141.46, 322.79], [170.12, 314.35], [216.5, 314.61], [190.73, 159.85], [141.46, 159.9], [170.12, 161.19], [216.5, 161.15]]}]\n```" - } - ] - }, - { - "id": 3834, - "data_source": "Mono3DRefer", - "scene_name": "007136", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, person_sitting, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[558.21, 194.74], [571.08, 195.46], [370.98, 196.95], [368.19, 196.07], [558.21, 142.3], [571.08, 140.17], [370.98, 135.77], [368.19, 138.37]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[376.61, 322.86], [441.33, 322.87], [424.27, 337.98], [352.63, 337.97], [376.61, 200.79], [441.33, 200.8], [424.27, 202.86], [352.63, 202.85]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[317.97, 320.74], [380.71, 320.75], [356.88, 335.82], [287.34, 335.81], [317.97, 200.76], [380.71, 200.76], [356.88, 202.84], [287.34, 202.84]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[395.72, 317.55], [344.12, 317.26], [364.32, 307.37], [412.2, 307.62], [395.72, 203.49], [344.12, 203.44], [364.32, 201.84], [412.2, 201.88]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[393.17, 292.58], [433.04, 292.65], [420.5, 300.6], [377.8, 300.53], [393.17, 198.84], [433.04, 198.85], [420.5, 200.1], [377.8, 200.08]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[441.21, 284.58], [401.31, 284.53], [412.52, 278.91], [450.25, 278.96], [441.21, 192.64], [401.31, 192.63], [412.52, 192.03], [450.25, 192.03]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[148.05, 327.19], [212.27, 324.86], [193.44, 335.96], [123.84, 338.67], [148.05, 159.23], [212.27, 159.58], [193.44, 157.88], [123.84, 157.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[190.73, 323.08], [141.46, 322.79], [170.12, 314.35], [216.5, 314.61], [190.73, 159.85], [141.46, 159.9], [170.12, 161.19], [216.5, 161.15]]}]\n```" - } - ] - }, - { - "id": 3835, - "data_source": "Mono3DRefer", - "scene_name": "007136", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, person_sitting, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[870.72, 355.84], [1065.84, 354.93], [1770.55, 609.92], [1294.01, 615.56], [870.72, 143.2], [1065.84, 143.4], [1770.55, 87.37], [1294.01, 86.13]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[376.61, 322.86], [441.33, 322.87], [424.27, 337.98], [352.63, 337.97], [376.61, 200.79], [441.33, 200.8], [424.27, 202.86], [352.63, 202.85]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[317.97, 320.74], [380.71, 320.75], [356.88, 335.82], [287.34, 335.81], [317.97, 200.76], [380.71, 200.76], [356.88, 202.84], [287.34, 202.84]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[395.72, 317.55], [344.12, 317.26], [364.32, 307.37], [412.2, 307.62], [395.72, 203.49], [344.12, 203.44], [364.32, 201.84], [412.2, 201.88]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[393.17, 292.58], [433.04, 292.65], [420.5, 300.6], [377.8, 300.53], [393.17, 198.84], [433.04, 198.85], [420.5, 200.1], [377.8, 200.08]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[441.21, 284.58], [401.31, 284.53], [412.52, 278.91], [450.25, 278.96], [441.21, 192.64], [401.31, 192.63], [412.52, 192.03], [450.25, 192.03]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[148.05, 327.19], [212.27, 324.86], [193.44, 335.96], [123.84, 338.67], [148.05, 159.23], [212.27, 159.58], [193.44, 157.88], [123.84, 157.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[190.73, 323.08], [141.46, 322.79], [170.12, 314.35], [216.5, 314.61], [190.73, 159.85], [141.46, 159.9], [170.12, 161.19], [216.5, 161.15]]}]\n```" - } - ] - }, - { - "id": 3836, - "data_source": "Mono3DRefer", - "scene_name": "007136", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (tram, van, person_sitting, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[870.72, 355.84], [1065.84, 354.93], [1770.55, 609.92], [1294.01, 615.56], [870.72, 143.2], [1065.84, 143.4], [1770.55, 87.37], [1294.01, 86.13]]}, {\"category\": \"tram\", \"corners_3d\": [[558.21, 194.74], [571.08, 195.46], [370.98, 196.95], [368.19, 196.07], [558.21, 142.3], [571.08, 140.17], [370.98, 135.77], [368.19, 138.37]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[376.61, 322.86], [441.33, 322.87], [424.27, 337.98], [352.63, 337.97], [376.61, 200.79], [441.33, 200.8], [424.27, 202.86], [352.63, 202.85]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[317.97, 320.74], [380.71, 320.75], [356.88, 335.82], [287.34, 335.81], [317.97, 200.76], [380.71, 200.76], [356.88, 202.84], [287.34, 202.84]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[395.72, 317.55], [344.12, 317.26], [364.32, 307.37], [412.2, 307.62], [395.72, 203.49], [344.12, 203.44], [364.32, 201.84], [412.2, 201.88]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[393.17, 292.58], [433.04, 292.65], [420.5, 300.6], [377.8, 300.53], [393.17, 198.84], [433.04, 198.85], [420.5, 200.1], [377.8, 200.08]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[441.21, 284.58], [401.31, 284.53], [412.52, 278.91], [450.25, 278.96], [441.21, 192.64], [401.31, 192.63], [412.52, 192.03], [450.25, 192.03]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[148.05, 327.19], [212.27, 324.86], [193.44, 335.96], [123.84, 338.67], [148.05, 159.23], [212.27, 159.58], [193.44, 157.88], [123.84, 157.47]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[190.73, 323.08], [141.46, 322.79], [170.12, 314.35], [216.5, 314.61], [190.73, 159.85], [141.46, 159.9], [170.12, 161.19], [216.5, 161.15]]}]\n```" - } - ] - }, - { - "id": 3837, - "data_source": "Mono3DRefer", - "scene_name": "007140", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[470.29, 245.34], [540.12, 245.84], [497.33, 271.74], [403.36, 270.81], [470.29, 174.05], [540.12, 174.06], [497.33, 174.48], [403.36, 174.47]]}, {\"category\": \"car\", \"corners_3d\": [[1781.79, 324.67], [2230.82, 381.77], [1468.14, 388.88], [1221.67, 328.39], [1781.79, 133.87], [2230.82, 119.2], [1468.14, 117.38], [1221.67, 132.91]]}, {\"category\": \"car\", \"corners_3d\": [[568.76, 221.69], [519.4, 221.52], [536.82, 214.78], [579.37, 214.9], [568.76, 174.32], [519.4, 174.32], [536.82, 174.12], [579.37, 174.12]]}, {\"category\": \"car\", \"corners_3d\": [[611.3, 192.24], [588.84, 192.21], [592.35, 190.83], [613.21, 190.85], [611.3, 172.98], [588.84, 172.98], [592.35, 172.98], [613.21, 172.98]]}, {\"category\": \"car\", \"corners_3d\": [[1171.97, 227.28], [1185.89, 232.05], [950.27, 224.58], [954.75, 220.9], [1171.97, 131.13], [1185.89, 127.47], [950.27, 133.2], [954.75, 136.02]]}]\n```" - } - ] - }, - { - "id": 3838, - "data_source": "Mono3DRefer", - "scene_name": "007147", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[522.45, 249.84], [511.51, 241.65], [711.1, 237.53], [745.9, 244.59], [522.45, 156.82], [511.51, 159.62], [711.1, 161.02], [745.9, 158.61]]}, {\"category\": \"car\", \"corners_3d\": [[197.36, 250.14], [226.23, 242.01], [468.44, 237.78], [471.33, 244.76], [197.36, 145.51], [226.23, 149.59], [468.44, 151.72], [471.33, 148.21]]}, {\"category\": \"car\", \"corners_3d\": [[-266.66, 263.26], [-163.47, 251.38], [85.87, 247.03], [28.83, 257.38], [-266.66, 147.53], [-163.47, 152.26], [85.87, 154.0], [28.83, 149.87]]}]\n```" - } - ] - }, - { - "id": 3839, - "data_source": "Mono3DRefer", - "scene_name": "007147", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[620.73, 307.34], [587.79, 300.8], [691.51, 285.58], [724.97, 290.54], [620.73, 152.77], [587.79, 154.2], [691.51, 157.52], [724.97, 156.44]]}]\n```" - } - ] - }, - { - "id": 3840, - "data_source": "Mono3DRefer", - "scene_name": "007147", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[522.45, 249.84], [511.51, 241.65], [711.1, 237.53], [745.9, 244.59], [522.45, 156.82], [511.51, 159.62], [711.1, 161.02], [745.9, 158.61]]}, {\"category\": \"car\", \"corners_3d\": [[197.36, 250.14], [226.23, 242.01], [468.44, 237.78], [471.33, 244.76], [197.36, 145.51], [226.23, 149.59], [468.44, 151.72], [471.33, 148.21]]}, {\"category\": \"car\", \"corners_3d\": [[-266.66, 263.26], [-163.47, 251.38], [85.87, 247.03], [28.83, 257.38], [-266.66, 147.53], [-163.47, 152.26], [85.87, 154.0], [28.83, 149.87]]}, {\"category\": \"cyclist\", \"corners_3d\": [[620.73, 307.34], [587.79, 300.8], [691.51, 285.58], [724.97, 290.54], [620.73, 152.77], [587.79, 154.2], [691.51, 157.52], [724.97, 156.44]]}]\n```" - } - ] - }, - { - "id": 3841, - "data_source": "Mono3DRefer", - "scene_name": "007153", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[21716.68, -17749.62], [14984.34, -6712.7], [-126.09, 499.56], [183.65, 490.28], [21716.68, 1062.83], [14984.34, 514.77], [-126.09, 156.63], [183.65, 157.09]]}, {\"category\": \"car\", \"corners_3d\": [[561.09, 240.3], [508.46, 239.29], [546.6, 229.02], [591.47, 229.74], [561.09, 189.11], [508.46, 188.87], [546.6, 186.39], [591.47, 186.57]]}, {\"category\": \"car\", \"corners_3d\": [[600.15, 227.55], [565.25, 226.67], [595.87, 220.85], [627.38, 221.54], [600.15, 186.65], [565.25, 186.43], [595.87, 184.96], [627.38, 185.14]]}, {\"category\": \"car\", \"corners_3d\": [[650.89, 219.41], [619.75, 218.71], [641.58, 214.62], [670.19, 215.2], [650.89, 187.42], [619.75, 187.21], [641.58, 185.93], [670.19, 186.11]]}, {\"category\": \"car\", \"corners_3d\": [[813.04, 191.29], [833.28, 191.58], [815.8, 192.84], [794.52, 192.51], [813.04, 170.89], [833.28, 170.86], [815.8, 170.73], [794.52, 170.76]]}]\n```" - } - ] - }, - { - "id": 3842, - "data_source": "Mono3DRefer", - "scene_name": "007155", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[385.68, 402.22], [180.56, 398.96], [379.91, 301.93], [497.91, 302.99], [385.68, 186.54], [180.56, 186.35], [379.91, 180.56], [497.91, 180.62]]}]\n```" - } - ] - }, - { - "id": 3843, - "data_source": "Mono3DRefer", - "scene_name": "007158", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[653.94, 185.02], [676.68, 185.02], [682.04, 186.0], [657.47, 186.0], [653.94, 163.46], [676.68, 163.46], [682.04, 162.71], [657.47, 162.71]]}, {\"category\": \"car\", \"corners_3d\": [[902.01, 209.02], [921.14, 211.38], [808.22, 211.38], [796.02, 209.02], [902.01, 163.59], [921.14, 162.98], [808.22, 162.98], [796.02, 163.59]]}]\n```" - } - ] - }, - { - "id": 3844, - "data_source": "Mono3DRefer", - "scene_name": "007159", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1073.53, 320.06], [1029.22, 308.64], [1110.09, 307.05], [1160.91, 318.18], [1073.53, 124.03], [1029.22, 128.66], [1110.09, 129.3], [1160.91, 124.8]]}]\n```" - } - ] - }, - { - "id": 3845, - "data_source": "Mono3DRefer", - "scene_name": "007175", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[573.38, 203.97], [545.64, 204.04], [547.58, 201.01], [572.63, 200.95], [573.38, 179.45], [545.64, 179.47], [547.58, 178.83], [572.63, 178.81]]}]\n```" - } - ] - }, - { - "id": 3846, - "data_source": "Mono3DRefer", - "scene_name": "007177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[549.38, 209.71], [537.3, 209.7], [541.42, 207.78], [552.87, 207.78], [549.38, 174.19], [537.3, 174.19], [541.42, 174.12], [552.87, 174.12]]}, {\"category\": \"cyclist\", \"corners_3d\": [[564.03, 208.14], [551.55, 208.16], [553.18, 206.55], [565.1, 206.53], [564.03, 174.92], [551.55, 174.92], [553.18, 174.83], [565.1, 174.83]]}]\n```" - } - ] - }, - { - "id": 3847, - "data_source": "Mono3DRefer", - "scene_name": "007177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[194.67, 315.24], [332.46, 315.27], [112.93, 427.85], [-133.68, 427.78], [194.67, 183.55], [332.46, 183.55], [112.93, 192.0], [-133.68, 192.0]]}, {\"category\": \"car\", \"corners_3d\": [[451.34, 244.1], [389.07, 244.11], [425.42, 232.33], [477.4, 232.32], [451.34, 183.34], [389.07, 183.34], [425.42, 181.61], [477.4, 181.61]]}, {\"category\": \"car\", \"corners_3d\": [[510.59, 212.36], [476.04, 212.37], [492.47, 207.48], [522.76, 207.48], [510.59, 182.16], [476.04, 182.16], [492.47, 181.01], [522.76, 181.01]]}, {\"category\": \"car\", \"corners_3d\": [[632.01, 193.8], [654.08, 193.83], [654.65, 195.37], [630.97, 195.33], [632.01, 175.72], [654.08, 175.73], [654.65, 175.94], [630.97, 175.93]]}, {\"category\": \"car\", \"corners_3d\": [[671.72, 232.86], [733.12, 232.81], [762.59, 246.39], [687.32, 246.46], [671.72, 173.58], [733.12, 173.58], [762.59, 173.74], [687.32, 173.74]]}]\n```" - } - ] - }, - { - "id": 3848, - "data_source": "Mono3DRefer", - "scene_name": "007177", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[194.67, 315.24], [332.46, 315.27], [112.93, 427.85], [-133.68, 427.78], [194.67, 183.55], [332.46, 183.55], [112.93, 192.0], [-133.68, 192.0]]}, {\"category\": \"car\", \"corners_3d\": [[451.34, 244.1], [389.07, 244.11], [425.42, 232.33], [477.4, 232.32], [451.34, 183.34], [389.07, 183.34], [425.42, 181.61], [477.4, 181.61]]}, {\"category\": \"cyclist\", \"corners_3d\": [[549.38, 209.71], [537.3, 209.7], [541.42, 207.78], [552.87, 207.78], [549.38, 174.19], [537.3, 174.19], [541.42, 174.12], [552.87, 174.12]]}, {\"category\": \"cyclist\", \"corners_3d\": [[564.03, 208.14], [551.55, 208.16], [553.18, 206.55], [565.1, 206.53], [564.03, 174.92], [551.55, 174.92], [553.18, 174.83], [565.1, 174.83]]}, {\"category\": \"car\", \"corners_3d\": [[510.59, 212.36], [476.04, 212.37], [492.47, 207.48], [522.76, 207.48], [510.59, 182.16], [476.04, 182.16], [492.47, 181.01], [522.76, 181.01]]}, {\"category\": \"car\", \"corners_3d\": [[632.01, 193.8], [654.08, 193.83], [654.65, 195.37], [630.97, 195.33], [632.01, 175.72], [654.08, 175.73], [654.65, 175.94], [630.97, 175.93]]}, {\"category\": \"car\", \"corners_3d\": [[671.72, 232.86], [733.12, 232.81], [762.59, 246.39], [687.32, 246.46], [671.72, 173.58], [733.12, 173.58], [762.59, 173.74], [687.32, 173.74]]}]\n```" - } - ] - }, - { - "id": 3849, - "data_source": "Mono3DRefer", - "scene_name": "007184", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[771.1, 222.26], [829.56, 223.0], [846.52, 233.54], [775.68, 232.46], [771.1, 169.09], [829.56, 169.04], [846.52, 168.23], [775.68, 168.32]]}, {\"category\": \"car\", \"corners_3d\": [[414.93, 247.94], [351.93, 245.65], [452.4, 230.66], [504.61, 232.1], [414.93, 184.46], [351.93, 184.11], [452.4, 181.79], [504.61, 182.01]]}, {\"category\": \"car\", \"corners_3d\": [[1117.82, 192.35], [1137.67, 193.32], [1047.8, 192.98], [1032.46, 192.04], [1117.82, 159.7], [1137.67, 159.05], [1047.8, 159.28], [1032.46, 159.91]]}]\n```" - } - ] - }, - { - "id": 3850, - "data_source": "Mono3DRefer", - "scene_name": "007185", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[232.24, 406.36], [15.1, 401.31], [286.08, 305.25], [414.2, 306.93], [232.24, 212.18], [15.1, 211.33], [286.08, 195.15], [414.2, 195.44]]}, {\"category\": \"car\", \"corners_3d\": [[776.87, 256.99], [856.87, 257.27], [921.44, 281.49], [818.31, 281.02], [776.87, 182.24], [856.87, 182.27], [921.44, 184.98], [818.31, 184.92]]}, {\"category\": \"car\", \"corners_3d\": [[743.69, 228.02], [805.81, 228.49], [836.46, 242.33], [758.73, 241.59], [743.69, 177.38], [805.81, 177.42], [836.46, 178.56], [758.73, 178.5]]}, {\"category\": \"car\", \"corners_3d\": [[516.45, 250.13], [448.46, 249.55], [487.07, 235.99], [543.21, 236.38], [516.45, 186.48], [448.46, 186.38], [487.07, 183.98], [543.21, 184.05]]}, {\"category\": \"car\", \"corners_3d\": [[552.34, 229.1], [505.33, 228.88], [523.86, 221.82], [565.0, 221.98], [552.34, 182.53], [505.33, 182.49], [523.86, 181.28], [565.0, 181.31]]}, {\"category\": \"car\", \"corners_3d\": [[716.16, 219.67], [759.66, 219.79], [782.42, 228.64], [730.67, 228.48], [716.16, 179.65], [759.66, 179.66], [782.42, 180.95], [730.67, 180.92]]}, {\"category\": \"car\", \"corners_3d\": [[575.08, 220.49], [535.88, 220.34], [550.0, 214.58], [584.47, 214.7], [575.08, 178.83], [535.88, 178.81], [550.0, 178.09], [584.47, 178.1]]}, {\"category\": \"car\", \"corners_3d\": [[604.82, 210.49], [578.8, 209.99], [602.23, 207.03], [626.44, 207.44], [604.82, 178.67], [578.8, 178.59], [602.23, 178.13], [626.44, 178.2]]}, {\"category\": \"car\", \"corners_3d\": [[686.89, 199.66], [664.93, 199.15], [693.1, 197.41], [714.09, 197.86], [686.89, 176.07], [664.93, 176.0], [693.1, 175.8], [714.09, 175.85]]}]\n```" - } - ] - }, - { - "id": 3851, - "data_source": "Mono3DRefer", - "scene_name": "007185", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[858.26, 234.13], [829.24, 233.8], [822.73, 230.45], [850.11, 230.75], [858.26, 157.62], [829.24, 157.7], [822.73, 158.53], [850.11, 158.46]]}]\n```" - } - ] - }, - { - "id": 3852, - "data_source": "Mono3DRefer", - "scene_name": "007185", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[232.24, 406.36], [15.1, 401.31], [286.08, 305.25], [414.2, 306.93], [232.24, 212.18], [15.1, 211.33], [286.08, 195.15], [414.2, 195.44]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[858.26, 234.13], [829.24, 233.8], [822.73, 230.45], [850.11, 230.75], [858.26, 157.62], [829.24, 157.7], [822.73, 158.53], [850.11, 158.46]]}, {\"category\": \"car\", \"corners_3d\": [[776.87, 256.99], [856.87, 257.27], [921.44, 281.49], [818.31, 281.02], [776.87, 182.24], [856.87, 182.27], [921.44, 184.98], [818.31, 184.92]]}, {\"category\": \"car\", \"corners_3d\": [[743.69, 228.02], [805.81, 228.49], [836.46, 242.33], [758.73, 241.59], [743.69, 177.38], [805.81, 177.42], [836.46, 178.56], [758.73, 178.5]]}, {\"category\": \"car\", \"corners_3d\": [[516.45, 250.13], [448.46, 249.55], [487.07, 235.99], [543.21, 236.38], [516.45, 186.48], [448.46, 186.38], [487.07, 183.98], [543.21, 184.05]]}, {\"category\": \"car\", \"corners_3d\": [[552.34, 229.1], [505.33, 228.88], [523.86, 221.82], [565.0, 221.98], [552.34, 182.53], [505.33, 182.49], [523.86, 181.28], [565.0, 181.31]]}, {\"category\": \"car\", \"corners_3d\": [[716.16, 219.67], [759.66, 219.79], [782.42, 228.64], [730.67, 228.48], [716.16, 179.65], [759.66, 179.66], [782.42, 180.95], [730.67, 180.92]]}, {\"category\": \"car\", \"corners_3d\": [[575.08, 220.49], [535.88, 220.34], [550.0, 214.58], [584.47, 214.7], [575.08, 178.83], [535.88, 178.81], [550.0, 178.09], [584.47, 178.1]]}, {\"category\": \"car\", \"corners_3d\": [[604.82, 210.49], [578.8, 209.99], [602.23, 207.03], [626.44, 207.44], [604.82, 178.67], [578.8, 178.59], [602.23, 178.13], [626.44, 178.2]]}, {\"category\": \"car\", \"corners_3d\": [[686.89, 199.66], [664.93, 199.15], [693.1, 197.41], [714.09, 197.86], [686.89, 176.07], [664.93, 176.0], [693.1, 175.8], [714.09, 175.85]]}]\n```" - } - ] - }, - { - "id": 3853, - "data_source": "Mono3DRefer", - "scene_name": "007188", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[710.29, 320.43], [703.62, 310.15], [814.41, 310.59], [829.92, 320.94], [710.29, 144.59], [703.62, 147.23], [814.41, 147.11], [829.92, 144.46]]}]\n```" - } - ] - }, - { - "id": 3854, - "data_source": "Mono3DRefer", - "scene_name": "007190", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[807.59, 225.14], [859.94, 224.95], [928.17, 237.38], [863.58, 237.67], [807.59, 176.97], [859.94, 176.96], [928.17, 177.93], [863.58, 177.96]]}]\n```" - } - ] - }, - { - "id": 3855, - "data_source": "Mono3DRefer", - "scene_name": "007190", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[284.65, 203.42], [285.52, 205.22], [65.24, 209.06], [77.93, 206.82], [284.65, 146.71], [285.52, 145.17], [65.24, 141.88], [77.93, 143.81]]}]\n```" - } - ] - }, - { - "id": 3856, - "data_source": "Mono3DRefer", - "scene_name": "007190", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, truck). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[807.59, 225.14], [859.94, 224.95], [928.17, 237.38], [863.58, 237.67], [807.59, 176.97], [859.94, 176.96], [928.17, 177.93], [863.58, 177.96]]}, {\"category\": \"truck\", \"corners_3d\": [[284.65, 203.42], [285.52, 205.22], [65.24, 209.06], [77.93, 206.82], [284.65, 146.71], [285.52, 145.17], [65.24, 141.88], [77.93, 143.81]]}]\n```" - } - ] - }, - { - "id": 3857, - "data_source": "Mono3DRefer", - "scene_name": "007191", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[484.05, 267.22], [389.13, 267.23], [442.25, 244.43], [514.23, 244.42], [484.05, 167.25], [389.13, 167.25], [442.25, 168.61], [514.23, 168.61]]}, {\"category\": \"car\", \"corners_3d\": [[517.17, 232.51], [465.8, 232.38], [489.41, 223.86], [533.45, 223.95], [517.17, 185.6], [465.8, 185.58], [489.41, 183.76], [533.45, 183.78]]}, {\"category\": \"car\", \"corners_3d\": [[553.63, 208.98], [525.16, 208.97], [533.33, 205.73], [559.24, 205.74], [553.63, 177.32], [525.16, 177.32], [533.33, 176.92], [559.24, 176.92]]}]\n```" - } - ] - }, - { - "id": 3858, - "data_source": "Mono3DRefer", - "scene_name": "007198", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[903.56, 240.26], [975.47, 243.03], [977.18, 258.62], [889.99, 254.51], [903.56, 172.02], [975.47, 171.98], [977.18, 171.79], [889.99, 171.84]]}, {\"category\": \"car\", \"corners_3d\": [[805.63, 207.29], [770.33, 206.64], [785.7, 203.23], [817.64, 203.75], [805.63, 176.27], [770.33, 176.2], [785.7, 175.87], [817.64, 175.92]]}, {\"category\": \"car\", \"corners_3d\": [[844.85, 196.38], [819.51, 196.03], [830.42, 194.31], [854.01, 194.61], [844.85, 174.92], [819.51, 174.89], [830.42, 174.74], [854.01, 174.77]]}]\n```" - } - ] - }, - { - "id": 3859, - "data_source": "Mono3DRefer", - "scene_name": "007218", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[276.8, 222.2], [243.58, 222.42], [281.45, 216.01], [310.25, 215.85], [276.8, 175.45], [243.58, 175.47], [281.45, 175.13], [310.25, 175.12]]}]\n```" - } - ] - }, - { - "id": 3860, - "data_source": "Mono3DRefer", - "scene_name": "007218", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[347.29, 234.58], [297.86, 235.38], [319.15, 227.3], [362.03, 226.7], [347.29, 186.59], [297.86, 186.77], [319.15, 184.97], [362.03, 184.84]]}, {\"category\": \"car\", \"corners_3d\": [[-111.69, 349.82], [-293.93, 358.93], [-72.62, 301.63], [47.81, 297.2], [-111.69, 225.92], [-293.93, 228.66], [-72.62, 211.47], [47.81, 210.14]]}, {\"category\": \"car\", \"corners_3d\": [[117.62, 225.63], [110.91, 227.44], [17.54, 229.7], [27.73, 227.74], [117.62, 201.74], [110.91, 202.73], [17.54, 203.97], [27.73, 202.9]]}, {\"category\": \"car\", \"corners_3d\": [[818.08, 204.87], [790.63, 204.2], [810.1, 202.1], [836.06, 202.69], [818.08, 178.97], [790.63, 178.84], [810.1, 178.44], [836.06, 178.55]]}]\n```" - } - ] - }, - { - "id": 3861, - "data_source": "Mono3DRefer", - "scene_name": "007218", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[347.29, 234.58], [297.86, 235.38], [319.15, 227.3], [362.03, 226.7], [347.29, 186.59], [297.86, 186.77], [319.15, 184.97], [362.03, 184.84]]}, {\"category\": \"van\", \"corners_3d\": [[276.8, 222.2], [243.58, 222.42], [281.45, 216.01], [310.25, 215.85], [276.8, 175.45], [243.58, 175.47], [281.45, 175.13], [310.25, 175.12]]}, {\"category\": \"car\", \"corners_3d\": [[-111.69, 349.82], [-293.93, 358.93], [-72.62, 301.63], [47.81, 297.2], [-111.69, 225.92], [-293.93, 228.66], [-72.62, 211.47], [47.81, 210.14]]}, {\"category\": \"car\", \"corners_3d\": [[117.62, 225.63], [110.91, 227.44], [17.54, 229.7], [27.73, 227.74], [117.62, 201.74], [110.91, 202.73], [17.54, 203.97], [27.73, 202.9]]}, {\"category\": \"car\", \"corners_3d\": [[818.08, 204.87], [790.63, 204.2], [810.1, 202.1], [836.06, 202.69], [818.08, 178.97], [790.63, 178.84], [810.1, 178.44], [836.06, 178.55]]}]\n```" - } - ] - }, - { - "id": 3862, - "data_source": "Mono3DRefer", - "scene_name": "007222", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[613.64, 230.87], [673.29, 230.97], [684.55, 244.39], [611.12, 244.24], [613.64, 178.57], [673.29, 178.58], [684.55, 179.9], [611.12, 179.88]]}, {\"category\": \"car\", \"corners_3d\": [[540.66, 217.43], [589.74, 217.62], [576.46, 226.94], [517.27, 226.67], [540.66, 170.51], [589.74, 170.5], [576.46, 170.01], [517.27, 170.03]]}, {\"category\": \"car\", \"corners_3d\": [[635.79, 208.53], [677.51, 208.69], [679.31, 215.48], [629.73, 215.24], [635.79, 169.53], [677.51, 169.52], [679.31, 168.88], [629.73, 168.91]]}, {\"category\": \"car\", \"corners_3d\": [[1279.37, 197.1], [1359.16, 199.36], [1233.07, 200.97], [1160.24, 198.44], [1279.37, 138.24], [1359.16, 135.01], [1233.07, 132.71], [1160.24, 136.33]]}, {\"category\": \"car\", \"corners_3d\": [[1161.23, 192.73], [1219.08, 194.29], [1113.72, 195.31], [1061.34, 193.61], [1161.23, 140.48], [1219.08, 137.93], [1113.72, 136.27], [1061.34, 139.06]]}]\n```" - } - ] - }, - { - "id": 3863, - "data_source": "Mono3DRefer", - "scene_name": "007246", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[636.3, 186.14], [610.46, 186.03], [624.7, 184.96], [648.53, 185.05], [636.3, 163.51], [610.46, 163.59], [624.7, 164.34], [648.53, 164.28]]}, {\"category\": \"car\", \"corners_3d\": [[221.87, 417.44], [-24.29, 417.51], [263.65, 306.27], [397.86, 306.25], [221.87, 187.67], [-24.29, 187.68], [263.65, 180.94], [397.86, 180.94]]}, {\"category\": \"car\", \"corners_3d\": [[672.57, 197.08], [705.11, 197.29], [701.58, 199.25], [666.49, 199.01], [672.57, 164.56], [705.11, 164.49], [701.58, 163.82], [666.49, 163.9]]}]\n```" - } - ] - }, - { - "id": 3864, - "data_source": "Mono3DRefer", - "scene_name": "007251", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[479.52, 228.38], [434.34, 229.0], [438.9, 223.1], [479.33, 222.6], [479.52, 186.92], [434.34, 187.07], [438.9, 185.58], [479.33, 185.45]]}]\n```" - } - ] - }, - { - "id": 3865, - "data_source": "Mono3DRefer", - "scene_name": "007257", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[985.56, 210.61], [1022.0, 211.44], [1022.88, 215.66], [982.53, 214.64], [985.56, 181.42], [1022.0, 181.61], [1022.88, 182.57], [982.53, 182.34]]}]\n```" - } - ] - }, - { - "id": 3866, - "data_source": "Mono3DRefer", - "scene_name": "007259", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[578.3, 207.08], [621.44, 207.45], [598.44, 214.09], [547.44, 213.55], [578.3, 162.85], [621.44, 162.74], [598.44, 160.8], [547.44, 160.96]]}, {\"category\": \"car\", \"corners_3d\": [[131.9, 266.19], [181.1, 261.26], [237.99, 268.26], [188.1, 274.03], [131.9, 207.38], [181.1, 205.55], [237.99, 208.14], [188.1, 210.28]]}, {\"category\": \"car\", \"corners_3d\": [[-61.59, 289.26], [34.71, 279.21], [98.3, 290.52], [-2.61, 302.95], [-61.59, 212.35], [34.71, 208.94], [98.3, 212.77], [-2.61, 216.99]]}, {\"category\": \"car\", \"corners_3d\": [[1252.06, 238.18], [1341.67, 248.07], [1063.2, 246.17], [1011.37, 236.74], [1252.06, 151.83], [1341.67, 148.65], [1063.2, 149.26], [1011.37, 152.29]]}, {\"category\": \"car\", \"corners_3d\": [[1081.4, 222.73], [1126.33, 228.15], [890.93, 226.78], [869.53, 221.62], [1081.4, 154.18], [1126.33, 152.16], [890.93, 152.67], [869.53, 154.6]]}, {\"category\": \"car\", \"corners_3d\": [[637.65, 190.49], [603.5, 190.38], [613.43, 188.64], [644.21, 188.73], [637.65, 157.67], [603.5, 157.76], [613.43, 159.26], [644.21, 159.18]]}, {\"category\": \"car\", \"corners_3d\": [[670.17, 172.4], [651.37, 172.4], [654.49, 172.43], [672.16, 172.42], [670.17, 156.3], [651.37, 156.36], [654.49, 157.35], [672.16, 157.3]]}]\n```" - } - ] - }, - { - "id": 3867, - "data_source": "Mono3DRefer", - "scene_name": "007261", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[709.34, 246.47], [785.22, 246.16], [863.36, 274.25], [758.69, 274.83], [709.34, 159.33], [785.22, 159.39], [863.36, 154.23], [758.69, 154.12]]}]\n```" - } - ] - }, - { - "id": 3868, - "data_source": "Mono3DRefer", - "scene_name": "007261", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[674.57, 226.91], [720.3, 226.81], [748.0, 238.15], [692.7, 238.3], [674.57, 178.85], [720.3, 178.84], [748.0, 180.1], [692.7, 180.11]]}, {\"category\": \"car\", \"corners_3d\": [[659.51, 216.51], [697.53, 216.39], [716.64, 223.13], [672.78, 223.28], [659.51, 178.46], [697.53, 178.44], [716.64, 179.31], [672.78, 179.32]]}, {\"category\": \"car\", \"corners_3d\": [[506.29, 277.51], [406.08, 277.66], [458.39, 249.65], [531.79, 249.56], [506.29, 185.32], [406.08, 185.34], [458.39, 182.0], [531.79, 181.99]]}]\n```" - } - ] - }, - { - "id": 3869, - "data_source": "Mono3DRefer", - "scene_name": "007261", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[674.57, 226.91], [720.3, 226.81], [748.0, 238.15], [692.7, 238.3], [674.57, 178.85], [720.3, 178.84], [748.0, 180.1], [692.7, 180.11]]}, {\"category\": \"car\", \"corners_3d\": [[659.51, 216.51], [697.53, 216.39], [716.64, 223.13], [672.78, 223.28], [659.51, 178.46], [697.53, 178.44], [716.64, 179.31], [672.78, 179.32]]}, {\"category\": \"car\", \"corners_3d\": [[506.29, 277.51], [406.08, 277.66], [458.39, 249.65], [531.79, 249.56], [506.29, 185.32], [406.08, 185.34], [458.39, 182.0], [531.79, 181.99]]}, {\"category\": \"van\", \"corners_3d\": [[709.34, 246.47], [785.22, 246.16], [863.36, 274.25], [758.69, 274.83], [709.34, 159.33], [785.22, 159.39], [863.36, 154.23], [758.69, 154.12]]}]\n```" - } - ] - }, - { - "id": 3870, - "data_source": "Mono3DRefer", - "scene_name": "007267", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[708.06, 191.42], [729.02, 191.67], [710.15, 193.16], [687.83, 192.86], [708.06, 163.13], [729.02, 163.0], [710.15, 162.22], [687.83, 162.37]]}]\n```" - } - ] - }, - { - "id": 3871, - "data_source": "Mono3DRefer", - "scene_name": "007267", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[697.31, 251.22], [774.05, 251.15], [846.09, 284.1], [737.1, 284.26], [697.31, 177.77], [774.05, 177.76], [846.09, 179.83], [737.1, 179.84]]}, {\"category\": \"car\", \"corners_3d\": [[319.67, 294.08], [436.62, 293.32], [327.64, 384.08], [120.4, 386.42], [319.67, 186.47], [436.62, 186.38], [327.64, 196.57], [120.4, 196.84]]}, {\"category\": \"car\", \"corners_3d\": [[646.55, 209.65], [684.67, 209.76], [688.64, 214.45], [645.67, 214.31], [646.55, 177.71], [684.67, 177.72], [688.64, 178.34], [645.67, 178.32]]}]\n```" - } - ] - }, - { - "id": 3872, - "data_source": "Mono3DRefer", - "scene_name": "007267", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[697.31, 251.22], [774.05, 251.15], [846.09, 284.1], [737.1, 284.26], [697.31, 177.77], [774.05, 177.76], [846.09, 179.83], [737.1, 179.84]]}, {\"category\": \"car\", \"corners_3d\": [[319.67, 294.08], [436.62, 293.32], [327.64, 384.08], [120.4, 386.42], [319.67, 186.47], [436.62, 186.38], [327.64, 196.57], [120.4, 196.84]]}, {\"category\": \"car\", \"corners_3d\": [[646.55, 209.65], [684.67, 209.76], [688.64, 214.45], [645.67, 214.31], [646.55, 177.71], [684.67, 177.72], [688.64, 178.34], [645.67, 178.32]]}, {\"category\": \"van\", \"corners_3d\": [[708.06, 191.42], [729.02, 191.67], [710.15, 193.16], [687.83, 192.86], [708.06, 163.13], [729.02, 163.0], [710.15, 162.22], [687.83, 162.37]]}]\n```" - } - ] - }, - { - "id": 3873, - "data_source": "Mono3DRefer", - "scene_name": "007268", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-10.54, 257.45], [-81.29, 259.65], [-18.48, 247.35], [41.09, 245.72], [-10.54, 206.96], [-81.29, 207.85], [-18.48, 202.89], [41.09, 202.23]]}, {\"category\": \"car\", \"corners_3d\": [[609.01, 220.99], [573.53, 220.31], [597.65, 215.59], [629.87, 216.14], [609.01, 186.57], [573.53, 186.37], [597.65, 185.03], [629.87, 185.19]]}]\n```" - } - ] - }, - { - "id": 3874, - "data_source": "Mono3DRefer", - "scene_name": "007270", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[513.42, 221.79], [469.74, 222.26], [472.57, 216.13], [510.84, 215.77], [513.42, 177.58], [469.74, 177.63], [472.57, 177.03], [510.84, 177.0]]}, {\"category\": \"car\", \"corners_3d\": [[493.29, 207.69], [463.38, 208.11], [456.19, 204.85], [483.45, 204.5], [493.29, 182.21], [463.38, 182.32], [456.19, 181.45], [483.45, 181.35]]}]\n```" - } - ] - }, - { - "id": 3875, - "data_source": "Mono3DRefer", - "scene_name": "007272", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[468.3, 312.79], [455.17, 327.07], [341.33, 327.54], [365.62, 313.17], [468.3, 150.82], [455.17, 147.61], [341.33, 147.51], [365.62, 150.73]]}]\n```" - } - ] - }, - { - "id": 3876, - "data_source": "Mono3DRefer", - "scene_name": "007275", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[482.19, 346.27], [314.43, 345.89], [420.38, 285.14], [529.31, 285.3], [482.19, 201.05], [314.43, 200.99], [420.38, 191.11], [529.31, 191.14]]}]\n```" - } - ] - }, - { - "id": 3877, - "data_source": "Mono3DRefer", - "scene_name": "007277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[561.14, 205.3], [550.72, 205.29], [553.69, 203.82], [563.64, 203.82], [561.14, 174.67], [550.72, 174.67], [553.69, 174.59], [563.64, 174.59]]}, {\"category\": \"cyclist\", \"corners_3d\": [[575.17, 203.32], [564.41, 203.32], [565.9, 202.11], [576.23, 202.11], [575.17, 174.64], [564.41, 174.64], [565.9, 174.57], [576.23, 174.57]]}]\n```" - } - ] - }, - { - "id": 3878, - "data_source": "Mono3DRefer", - "scene_name": "007277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[792.76, 329.08], [941.63, 328.78], [1218.34, 456.16], [948.4, 457.15], [792.76, 187.29], [941.63, 187.26], [1218.34, 199.03], [948.4, 199.12]]}, {\"category\": \"car\", \"corners_3d\": [[315.86, 274.95], [414.13, 274.96], [324.05, 321.88], [180.64, 321.86], [315.86, 181.03], [414.13, 181.03], [324.05, 184.78], [180.64, 184.78]]}, {\"category\": \"car\", \"corners_3d\": [[482.06, 233.03], [430.29, 233.03], [455.55, 224.52], [500.01, 224.52], [482.06, 182.51], [430.29, 182.51], [455.55, 181.15], [500.01, 181.14]]}, {\"category\": \"car\", \"corners_3d\": [[523.21, 210.39], [492.02, 210.39], [505.23, 206.15], [532.9, 206.15], [523.21, 183.12], [492.02, 183.12], [505.23, 181.96], [532.9, 181.96]]}, {\"category\": \"car\", \"corners_3d\": [[634.95, 193.11], [655.7, 193.14], [656.34, 194.53], [634.17, 194.5], [634.95, 176.11], [655.7, 176.12], [656.34, 176.34], [634.17, 176.34]]}, {\"category\": \"car\", \"corners_3d\": [[665.03, 223.99], [716.33, 223.92], [738.35, 233.24], [677.71, 233.34], [665.03, 174.38], [716.33, 174.38], [738.35, 174.66], [677.71, 174.66]]}]\n```" - } - ] - }, - { - "id": 3879, - "data_source": "Mono3DRefer", - "scene_name": "007277", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[792.76, 329.08], [941.63, 328.78], [1218.34, 456.16], [948.4, 457.15], [792.76, 187.29], [941.63, 187.26], [1218.34, 199.03], [948.4, 199.12]]}, {\"category\": \"car\", \"corners_3d\": [[315.86, 274.95], [414.13, 274.96], [324.05, 321.88], [180.64, 321.86], [315.86, 181.03], [414.13, 181.03], [324.05, 184.78], [180.64, 184.78]]}, {\"category\": \"car\", \"corners_3d\": [[482.06, 233.03], [430.29, 233.03], [455.55, 224.52], [500.01, 224.52], [482.06, 182.51], [430.29, 182.51], [455.55, 181.15], [500.01, 181.14]]}, {\"category\": \"cyclist\", \"corners_3d\": [[561.14, 205.3], [550.72, 205.29], [553.69, 203.82], [563.64, 203.82], [561.14, 174.67], [550.72, 174.67], [553.69, 174.59], [563.64, 174.59]]}, {\"category\": \"cyclist\", \"corners_3d\": [[575.17, 203.32], [564.41, 203.32], [565.9, 202.11], [576.23, 202.11], [575.17, 174.64], [564.41, 174.64], [565.9, 174.57], [576.23, 174.57]]}, {\"category\": \"car\", \"corners_3d\": [[523.21, 210.39], [492.02, 210.39], [505.23, 206.15], [532.9, 206.15], [523.21, 183.12], [492.02, 183.12], [505.23, 181.96], [532.9, 181.96]]}, {\"category\": \"car\", \"corners_3d\": [[634.95, 193.11], [655.7, 193.14], [656.34, 194.53], [634.17, 194.5], [634.95, 176.11], [655.7, 176.12], [656.34, 176.34], [634.17, 176.34]]}, {\"category\": \"car\", \"corners_3d\": [[665.03, 223.99], [716.33, 223.92], [738.35, 233.24], [677.71, 233.34], [665.03, 174.38], [716.33, 174.38], [738.35, 174.66], [677.71, 174.66]]}]\n```" - } - ] - }, - { - "id": 3880, - "data_source": "Mono3DRefer", - "scene_name": "007285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[216.25, 256.58], [195.01, 256.48], [229.15, 250.03], [248.78, 250.12], [216.25, 192.16], [195.01, 192.14], [229.15, 190.65], [248.78, 190.67]]}]\n```" - } - ] - }, - { - "id": 3881, - "data_source": "Mono3DRefer", - "scene_name": "007285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[433.44, 299.85], [327.48, 299.49], [409.56, 264.37], [486.24, 264.55], [433.44, 198.65], [327.48, 198.58], [409.56, 191.44], [486.24, 191.48]]}]\n```" - } - ] - }, - { - "id": 3882, - "data_source": "Mono3DRefer", - "scene_name": "007285", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[216.25, 256.58], [195.01, 256.48], [229.15, 250.03], [248.78, 250.12], [216.25, 192.16], [195.01, 192.14], [229.15, 190.65], [248.78, 190.67]]}, {\"category\": \"car\", \"corners_3d\": [[433.44, 299.85], [327.48, 299.49], [409.56, 264.37], [486.24, 264.55], [433.44, 198.65], [327.48, 198.58], [409.56, 191.44], [486.24, 191.48]]}]\n```" - } - ] - }, - { - "id": 3883, - "data_source": "Mono3DRefer", - "scene_name": "007294", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[655.95, 222.05], [701.91, 221.96], [723.93, 231.5], [669.08, 231.63], [655.95, 173.68], [701.91, 173.68], [723.93, 173.84], [669.08, 173.84]]}, {\"category\": \"van\", \"corners_3d\": [[642.83, 205.71], [673.99, 205.67], [684.92, 209.86], [649.78, 209.91], [642.83, 170.18], [673.99, 170.19], [684.92, 169.85], [649.78, 169.84]]}]\n```" - } - ] - }, - { - "id": 3884, - "data_source": "Mono3DRefer", - "scene_name": "007294", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[825.33, 259.58], [843.97, 259.56], [868.01, 268.21], [847.52, 268.23], [825.33, 184.78], [843.97, 184.78], [868.01, 185.97], [847.52, 185.97]]}]\n```" - } - ] - }, - { - "id": 3885, - "data_source": "Mono3DRefer", - "scene_name": "007294", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[495.86, 278.79], [399.27, 279.22], [440.09, 256.13], [515.66, 255.87], [495.86, 191.39], [399.27, 191.47], [440.09, 187.43], [515.66, 187.38]]}, {\"category\": \"car\", \"corners_3d\": [[687.05, 250.62], [763.78, 250.55], [809.59, 272.67], [711.04, 272.8], [687.05, 176.46], [763.78, 176.45], [809.59, 177.48], [711.04, 177.48]]}, {\"category\": \"car\", \"corners_3d\": [[671.42, 231.07], [724.86, 230.98], [750.67, 242.6], [686.58, 242.72], [671.42, 182.43], [724.86, 182.41], [750.67, 184.32], [686.58, 184.34]]}, {\"category\": \"car\", \"corners_3d\": [[565.2, 211.07], [531.54, 211.09], [539.21, 206.91], [569.2, 206.9], [565.2, 180.05], [531.54, 180.06], [539.21, 179.27], [569.2, 179.27]]}]\n```" - } - ] - }, - { - "id": 3886, - "data_source": "Mono3DRefer", - "scene_name": "007294", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[825.33, 259.58], [843.97, 259.56], [868.01, 268.21], [847.52, 268.23], [825.33, 184.78], [843.97, 184.78], [868.01, 185.97], [847.52, 185.97]]}, {\"category\": \"van\", \"corners_3d\": [[655.95, 222.05], [701.91, 221.96], [723.93, 231.5], [669.08, 231.63], [655.95, 173.68], [701.91, 173.68], [723.93, 173.84], [669.08, 173.84]]}, {\"category\": \"van\", \"corners_3d\": [[642.83, 205.71], [673.99, 205.67], [684.92, 209.86], [649.78, 209.91], [642.83, 170.18], [673.99, 170.19], [684.92, 169.85], [649.78, 169.84]]}]\n```" - } - ] - }, - { - "id": 3887, - "data_source": "Mono3DRefer", - "scene_name": "007294", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[495.86, 278.79], [399.27, 279.22], [440.09, 256.13], [515.66, 255.87], [495.86, 191.39], [399.27, 191.47], [440.09, 187.43], [515.66, 187.38]]}, {\"category\": \"car\", \"corners_3d\": [[687.05, 250.62], [763.78, 250.55], [809.59, 272.67], [711.04, 272.8], [687.05, 176.46], [763.78, 176.45], [809.59, 177.48], [711.04, 177.48]]}, {\"category\": \"car\", \"corners_3d\": [[671.42, 231.07], [724.86, 230.98], [750.67, 242.6], [686.58, 242.72], [671.42, 182.43], [724.86, 182.41], [750.67, 184.32], [686.58, 184.34]]}, {\"category\": \"van\", \"corners_3d\": [[655.95, 222.05], [701.91, 221.96], [723.93, 231.5], [669.08, 231.63], [655.95, 173.68], [701.91, 173.68], [723.93, 173.84], [669.08, 173.84]]}, {\"category\": \"car\", \"corners_3d\": [[565.2, 211.07], [531.54, 211.09], [539.21, 206.91], [569.2, 206.9], [565.2, 180.05], [531.54, 180.06], [539.21, 179.27], [569.2, 179.27]]}, {\"category\": \"van\", \"corners_3d\": [[642.83, 205.71], [673.99, 205.67], [684.92, 209.86], [649.78, 209.91], [642.83, 170.18], [673.99, 170.19], [684.92, 169.85], [649.78, 169.84]]}]\n```" - } - ] - }, - { - "id": 3888, - "data_source": "Mono3DRefer", - "scene_name": "007294", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[825.33, 259.58], [843.97, 259.56], [868.01, 268.21], [847.52, 268.23], [825.33, 184.78], [843.97, 184.78], [868.01, 185.97], [847.52, 185.97]]}, {\"category\": \"car\", \"corners_3d\": [[495.86, 278.79], [399.27, 279.22], [440.09, 256.13], [515.66, 255.87], [495.86, 191.39], [399.27, 191.47], [440.09, 187.43], [515.66, 187.38]]}, {\"category\": \"car\", \"corners_3d\": [[687.05, 250.62], [763.78, 250.55], [809.59, 272.67], [711.04, 272.8], [687.05, 176.46], [763.78, 176.45], [809.59, 177.48], [711.04, 177.48]]}, {\"category\": \"car\", \"corners_3d\": [[671.42, 231.07], [724.86, 230.98], [750.67, 242.6], [686.58, 242.72], [671.42, 182.43], [724.86, 182.41], [750.67, 184.32], [686.58, 184.34]]}, {\"category\": \"car\", \"corners_3d\": [[565.2, 211.07], [531.54, 211.09], [539.21, 206.91], [569.2, 206.9], [565.2, 180.05], [531.54, 180.06], [539.21, 179.27], [569.2, 179.27]]}]\n```" - } - ] - }, - { - "id": 3889, - "data_source": "Mono3DRefer", - "scene_name": "007294", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[825.33, 259.58], [843.97, 259.56], [868.01, 268.21], [847.52, 268.23], [825.33, 184.78], [843.97, 184.78], [868.01, 185.97], [847.52, 185.97]]}, {\"category\": \"car\", \"corners_3d\": [[495.86, 278.79], [399.27, 279.22], [440.09, 256.13], [515.66, 255.87], [495.86, 191.39], [399.27, 191.47], [440.09, 187.43], [515.66, 187.38]]}, {\"category\": \"car\", \"corners_3d\": [[687.05, 250.62], [763.78, 250.55], [809.59, 272.67], [711.04, 272.8], [687.05, 176.46], [763.78, 176.45], [809.59, 177.48], [711.04, 177.48]]}, {\"category\": \"car\", \"corners_3d\": [[671.42, 231.07], [724.86, 230.98], [750.67, 242.6], [686.58, 242.72], [671.42, 182.43], [724.86, 182.41], [750.67, 184.32], [686.58, 184.34]]}, {\"category\": \"van\", \"corners_3d\": [[655.95, 222.05], [701.91, 221.96], [723.93, 231.5], [669.08, 231.63], [655.95, 173.68], [701.91, 173.68], [723.93, 173.84], [669.08, 173.84]]}, {\"category\": \"car\", \"corners_3d\": [[565.2, 211.07], [531.54, 211.09], [539.21, 206.91], [569.2, 206.9], [565.2, 180.05], [531.54, 180.06], [539.21, 179.27], [569.2, 179.27]]}, {\"category\": \"van\", \"corners_3d\": [[642.83, 205.71], [673.99, 205.67], [684.92, 209.86], [649.78, 209.91], [642.83, 170.18], [673.99, 170.19], [684.92, 169.85], [649.78, 169.84]]}]\n```" - } - ] - }, - { - "id": 3890, - "data_source": "Mono3DRefer", - "scene_name": "007297", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[457.9, 280.46], [368.69, 280.33], [427.16, 254.94], [495.33, 255.02], [457.9, 181.02], [368.69, 181.01], [427.16, 179.09], [495.33, 179.09]]}, {\"category\": \"car\", \"corners_3d\": [[537.86, 229.02], [490.23, 229.09], [503.48, 221.95], [545.06, 221.89], [537.86, 182.46], [490.23, 182.48], [503.48, 181.25], [545.06, 181.24]]}, {\"category\": \"car\", \"corners_3d\": [[569.35, 205.43], [541.18, 205.09], [562.94, 202.39], [588.95, 202.68], [569.35, 175.53], [541.18, 175.5], [562.94, 175.28], [588.95, 175.3]]}]\n```" - } - ] - }, - { - "id": 3891, - "data_source": "Mono3DRefer", - "scene_name": "007299", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-1587.53, 1721.64], [-3259.92, 1822.74], [-44.85, 443.96], [210.67, 441.09], [-1587.53, 216.06], [-3259.92, 218.88], [-44.85, 180.42], [210.67, 180.34]]}, {\"category\": \"car\", \"corners_3d\": [[748.74, 278.51], [875.07, 277.02], [1100.44, 349.75], [889.26, 354.09], [748.74, 155.56], [875.07, 155.81], [1100.44, 143.9], [889.26, 143.19]]}, {\"category\": \"car\", \"corners_3d\": [[281.07, 391.5], [89.17, 393.83], [267.19, 313.7], [388.77, 312.75], [281.07, 183.93], [89.17, 184.05], [267.19, 179.99], [388.77, 179.94]]}, {\"category\": \"car\", \"corners_3d\": [[420.23, 279.28], [323.77, 279.85], [386.19, 253.8], [459.0, 253.47], [420.23, 199.73], [323.77, 199.88], [386.19, 193.3], [459.0, 193.21]]}, {\"category\": \"car\", \"corners_3d\": [[508.97, 218.8], [471.99, 218.83], [486.57, 213.66], [519.38, 213.64], [508.97, 177.71], [471.99, 177.71], [486.57, 177.16], [519.38, 177.16]]}]\n```" - } - ] - }, - { - "id": 3892, - "data_source": "Mono3DRefer", - "scene_name": "007308", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[379.41, 298.53], [390.63, 292.67], [464.71, 292.68], [457.36, 298.55], [379.41, 157.86], [390.63, 158.98], [464.71, 158.98], [457.36, 157.85]]}]\n```" - } - ] - }, - { - "id": 3893, - "data_source": "Mono3DRefer", - "scene_name": "007317", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[49.57, 236.48], [2.9, 237.42], [64.76, 227.59], [103.64, 226.91], [49.57, 178.69], [2.9, 178.77], [64.76, 177.87], [103.64, 177.81]]}]\n```" - } - ] - }, - { - "id": 3894, - "data_source": "Mono3DRefer", - "scene_name": "007317", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[94.5, 262.04], [12.41, 265.2], [65.02, 249.44], [132.0, 247.25], [94.5, 193.31], [12.41, 194.03], [65.02, 190.42], [132.0, 189.92]]}, {\"category\": \"car\", \"corners_3d\": [[145.5, 245.8], [90.35, 247.25], [126.41, 238.17], [174.33, 237.05], [145.5, 203.46], [90.35, 204.06], [126.41, 200.25], [174.33, 199.79]]}, {\"category\": \"car\", \"corners_3d\": [[677.15, 215.18], [644.93, 214.47], [666.41, 210.81], [696.08, 211.39], [677.15, 184.01], [644.93, 183.82], [666.41, 182.86], [696.08, 183.01]]}]\n```" - } - ] - }, - { - "id": 3895, - "data_source": "Mono3DRefer", - "scene_name": "007317", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[94.5, 262.04], [12.41, 265.2], [65.02, 249.44], [132.0, 247.25], [94.5, 193.31], [12.41, 194.03], [65.02, 190.42], [132.0, 189.92]]}, {\"category\": \"van\", \"corners_3d\": [[49.57, 236.48], [2.9, 237.42], [64.76, 227.59], [103.64, 226.91], [49.57, 178.69], [2.9, 178.77], [64.76, 177.87], [103.64, 177.81]]}, {\"category\": \"car\", \"corners_3d\": [[145.5, 245.8], [90.35, 247.25], [126.41, 238.17], [174.33, 237.05], [145.5, 203.46], [90.35, 204.06], [126.41, 200.25], [174.33, 199.79]]}, {\"category\": \"car\", \"corners_3d\": [[677.15, 215.18], [644.93, 214.47], [666.41, 210.81], [696.08, 211.39], [677.15, 184.01], [644.93, 183.82], [666.41, 182.86], [696.08, 183.01]]}]\n```" - } - ] - }, - { - "id": 3896, - "data_source": "Mono3DRefer", - "scene_name": "007320", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[559.01, 298.47], [561.67, 292.57], [636.25, 292.71], [637.51, 298.62], [559.01, 156.91], [561.67, 158.09], [636.25, 158.06], [637.51, 156.88]]}]\n```" - } - ] - }, - { - "id": 3897, - "data_source": "Mono3DRefer", - "scene_name": "007321", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[430.41, 234.61], [442.32, 232.98], [485.07, 234.04], [474.27, 235.74], [430.41, 156.49], [442.32, 157.22], [485.07, 156.75], [474.27, 155.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[441.36, 232.51], [452.78, 230.95], [496.95, 232.0], [486.67, 233.62], [441.36, 157.65], [452.78, 158.34], [496.95, 157.87], [486.67, 157.16]]}]\n```" - } - ] - }, - { - "id": 3898, - "data_source": "Mono3DRefer", - "scene_name": "007321", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[787.06, 227.01], [740.74, 223.52], [845.81, 219.06], [894.54, 221.83], [787.06, 157.53], [740.74, 159.25], [845.81, 161.46], [894.54, 160.09]]}]\n```" - } - ] - }, - { - "id": 3899, - "data_source": "Mono3DRefer", - "scene_name": "007321", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[430.41, 234.61], [442.32, 232.98], [485.07, 234.04], [474.27, 235.74], [430.41, 156.49], [442.32, 157.22], [485.07, 156.75], [474.27, 155.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[441.36, 232.51], [452.78, 230.95], [496.95, 232.0], [486.67, 233.62], [441.36, 157.65], [452.78, 158.34], [496.95, 157.87], [486.67, 157.16]]}, {\"category\": \"van\", \"corners_3d\": [[787.06, 227.01], [740.74, 223.52], [845.81, 219.06], [894.54, 221.83], [787.06, 157.53], [740.74, 159.25], [845.81, 161.46], [894.54, 160.09]]}]\n```" - } - ] - }, - { - "id": 3900, - "data_source": "Mono3DRefer", - "scene_name": "007325", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[635.49, 210.09], [668.04, 210.04], [678.23, 214.8], [641.53, 214.86], [635.49, 183.66], [668.04, 183.64], [678.23, 185.03], [641.53, 185.04]]}, {\"category\": \"car\", \"corners_3d\": [[500.22, 221.44], [493.82, 224.28], [378.77, 224.28], [391.53, 221.44], [500.22, 181.03], [493.82, 181.51], [378.77, 181.51], [391.53, 181.03]]}, {\"category\": \"car\", \"corners_3d\": [[278.91, 245.74], [304.51, 239.95], [445.62, 239.8], [432.23, 245.56], [278.91, 186.63], [304.51, 185.54], [445.62, 185.51], [432.23, 186.6]]}, {\"category\": \"car\", \"corners_3d\": [[879.28, 235.34], [905.13, 240.75], [756.77, 241.31], [742.63, 235.82], [879.28, 182.07], [905.13, 182.87], [756.77, 182.96], [742.63, 182.14]]}, {\"category\": \"car\", \"corners_3d\": [[141.72, 269.42], [188.76, 258.51], [393.09, 256.84], [372.56, 267.31], [141.72, 172.29], [188.76, 172.35], [393.09, 172.36], [372.56, 172.3]]}, {\"category\": \"car\", \"corners_3d\": [[375.16, 274.72], [335.93, 289.9], [55.43, 288.05], [129.99, 273.32], [375.16, 185.79], [335.93, 187.72], [55.43, 187.48], [129.99, 185.61]]}, {\"category\": \"car\", \"corners_3d\": [[331.89, 302.27], [275.78, 325.42], [-43.33, 321.96], [59.17, 299.78], [331.89, 191.95], [275.78, 195.37], [-43.33, 194.86], [59.17, 191.58]]}, {\"category\": \"car\", \"corners_3d\": [[196.69, 330.31], [79.06, 370.9], [-294.23, 364.46], [-105.14, 326.21], [196.69, 192.39], [79.06, 197.43], [-294.23, 196.63], [-105.14, 191.88]]}, {\"category\": \"car\", \"corners_3d\": [[1157.34, 288.97], [1261.58, 309.2], [992.74, 313.0], [926.87, 291.72], [1157.34, 183.7], [1261.58, 185.59], [992.74, 185.94], [926.87, 183.95]]}, {\"category\": \"car\", \"corners_3d\": [[1310.82, 313.48], [1489.87, 346.34], [1144.6, 354.76], [1026.51, 318.96], [1310.82, 192.67], [1489.87, 197.3], [1144.6, 198.49], [1026.51, 193.44]]}]\n```" - } - ] - }, - { - "id": 3901, - "data_source": "Mono3DRefer", - "scene_name": "007327", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[939.42, 305.81], [979.97, 319.02], [872.62, 320.55], [842.0, 307.07], [939.42, 147.01], [979.97, 143.48], [872.62, 143.07], [842.0, 146.68]]}]\n```" - } - ] - }, - { - "id": 3902, - "data_source": "Mono3DRefer", - "scene_name": "007331", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[279.75, 220.38], [273.31, 221.05], [264.94, 220.94], [271.47, 220.28], [279.75, 183.83], [273.31, 183.99], [264.94, 183.96], [271.47, 183.81]]}]\n```" - } - ] - }, - { - "id": 3903, - "data_source": "Mono3DRefer", - "scene_name": "007331", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[593.92, 200.14], [622.41, 200.16], [622.2, 202.74], [591.03, 202.71], [593.92, 172.0], [622.41, 171.99], [622.2, 171.91], [591.03, 171.91]]}]\n```" - } - ] - }, - { - "id": 3904, - "data_source": "Mono3DRefer", - "scene_name": "007331", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[593.92, 200.14], [622.41, 200.16], [622.2, 202.74], [591.03, 202.71], [593.92, 172.0], [622.41, 171.99], [622.2, 171.91], [591.03, 171.91]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[279.75, 220.38], [273.31, 221.05], [264.94, 220.94], [271.47, 220.28], [279.75, 183.83], [273.31, 183.99], [264.94, 183.96], [271.47, 183.81]]}]\n```" - } - ] - }, - { - "id": 3905, - "data_source": "Mono3DRefer", - "scene_name": "007344", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[93.97, 335.27], [79.59, 326.89], [196.87, 312.21], [215.73, 318.95], [93.97, 166.62], [79.59, 167.37], [196.87, 168.69], [215.73, 168.08]]}]\n```" - } - ] - }, - { - "id": 3906, - "data_source": "Mono3DRefer", - "scene_name": "007358", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-102.63, 934.74], [-645.99, 871.41], [227.91, 411.5], [428.62, 418.47], [-102.63, 274.31], [-645.99, 265.87], [227.91, 204.63], [428.62, 205.56]]}, {\"category\": \"car\", \"corners_3d\": [[350.79, 359.41], [213.84, 360.14], [326.75, 304.6], [422.89, 304.23], [350.79, 198.39], [213.84, 198.49], [326.75, 190.89], [422.89, 190.84]]}, {\"category\": \"car\", \"corners_3d\": [[526.82, 205.14], [504.68, 205.32], [502.37, 203.15], [523.05, 202.99], [526.82, 182.54], [504.68, 182.59], [502.37, 181.94], [523.05, 181.89]]}]\n```" - } - ] - }, - { - "id": 3907, - "data_source": "Mono3DRefer", - "scene_name": "007362", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[148.51, 270.75], [45.05, 288.48], [-320.95, 282.39], [-165.16, 266.45], [148.51, 133.34], [45.05, 124.07], [-320.95, 127.25], [-165.16, 135.58]]}]\n```" - } - ] - }, - { - "id": 3908, - "data_source": "Mono3DRefer", - "scene_name": "007370", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[454.79, 261.59], [364.91, 262.79], [401.97, 242.51], [471.42, 241.79], [454.79, 171.77], [364.91, 171.75], [401.97, 172.0], [471.42, 172.01]]}, {\"category\": \"car\", \"corners_3d\": [[480.62, 215.97], [435.19, 216.67], [432.74, 210.78], [472.19, 210.26], [480.62, 177.78], [435.19, 177.86], [432.74, 177.19], [472.19, 177.13]]}]\n```" - } - ] - }, - { - "id": 3909, - "data_source": "Mono3DRefer", - "scene_name": "007383", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1328.23, 245.58], [1486.43, 260.54], [1109.51, 264.05], [1012.5, 247.98], [1328.23, 87.86], [1486.43, 70.39], [1109.51, 66.28], [1012.5, 85.06]]}]\n```" - } - ] - }, - { - "id": 3910, - "data_source": "Mono3DRefer", - "scene_name": "007383", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[481.05, 265.38], [394.02, 265.5], [440.59, 244.73], [508.09, 244.66], [481.05, 174.41], [394.02, 174.41], [440.59, 174.06], [508.09, 174.06]]}, {\"category\": \"car\", \"corners_3d\": [[683.33, 220.62], [740.62, 220.43], [778.33, 231.2], [708.18, 231.48], [683.33, 166.11], [740.62, 166.13], [778.33, 164.61], [708.18, 164.57]]}, {\"category\": \"car\", \"corners_3d\": [[518.23, 239.09], [460.07, 239.25], [479.02, 229.37], [528.5, 229.25], [518.23, 174.76], [460.07, 174.76], [479.02, 174.48], [528.5, 174.47]]}, {\"category\": \"car\", \"corners_3d\": [[199.63, 449.07], [-77.84, 450.21], [254.44, 314.6], [395.83, 314.3], [199.63, 164.77], [-77.84, 164.73], [254.44, 168.7], [395.83, 168.71]]}, {\"category\": \"car\", \"corners_3d\": [[722.26, 239.26], [793.14, 239.07], [853.16, 258.49], [761.65, 258.81], [722.26, 178.69], [793.14, 178.67], [853.16, 180.38], [761.65, 180.4]]}]\n```" - } - ] - }, - { - "id": 3911, - "data_source": "Mono3DRefer", - "scene_name": "007383", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[481.05, 265.38], [394.02, 265.5], [440.59, 244.73], [508.09, 244.66], [481.05, 174.41], [394.02, 174.41], [440.59, 174.06], [508.09, 174.06]]}, {\"category\": \"car\", \"corners_3d\": [[683.33, 220.62], [740.62, 220.43], [778.33, 231.2], [708.18, 231.48], [683.33, 166.11], [740.62, 166.13], [778.33, 164.61], [708.18, 164.57]]}, {\"category\": \"car\", \"corners_3d\": [[518.23, 239.09], [460.07, 239.25], [479.02, 229.37], [528.5, 229.25], [518.23, 174.76], [460.07, 174.76], [479.02, 174.48], [528.5, 174.47]]}, {\"category\": \"car\", \"corners_3d\": [[199.63, 449.07], [-77.84, 450.21], [254.44, 314.6], [395.83, 314.3], [199.63, 164.77], [-77.84, 164.73], [254.44, 168.7], [395.83, 168.71]]}, {\"category\": \"van\", \"corners_3d\": [[1328.23, 245.58], [1486.43, 260.54], [1109.51, 264.05], [1012.5, 247.98], [1328.23, 87.86], [1486.43, 70.39], [1109.51, 66.28], [1012.5, 85.06]]}, {\"category\": \"car\", \"corners_3d\": [[722.26, 239.26], [793.14, 239.07], [853.16, 258.49], [761.65, 258.81], [722.26, 178.69], [793.14, 178.67], [853.16, 180.38], [761.65, 180.4]]}]\n```" - } - ] - }, - { - "id": 3912, - "data_source": "Mono3DRefer", - "scene_name": "007390", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[0.5, 242.16], [-58.57, 241.87], [72.55, 229.01], [121.03, 229.2], [0.5, 173.23], [-58.57, 173.23], [72.55, 173.16], [121.03, 173.16]]}, {\"category\": \"car\", \"corners_3d\": [[151.41, 228.44], [110.24, 228.17], [181.98, 221.04], [218.13, 221.24], [151.41, 183.84], [110.24, 183.79], [181.98, 182.38], [218.13, 182.42]]}]\n```" - } - ] - }, - { - "id": 3913, - "data_source": "Mono3DRefer", - "scene_name": "007406", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[438.37, 238.79], [376.83, 238.85], [420.02, 226.18], [469.72, 226.14], [438.37, 176.54], [376.83, 176.54], [420.02, 175.84], [469.72, 175.83]]}]\n```" - } - ] - }, - { - "id": 3914, - "data_source": "Mono3DRefer", - "scene_name": "007406", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[962.56, 243.12], [991.98, 245.27], [974.68, 247.91], [944.75, 245.61], [962.56, 170.17], [991.98, 170.09], [974.68, 169.99], [944.75, 170.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1019.2, 237.24], [995.57, 235.75], [1008.22, 234.05], [1031.49, 235.46], [1019.2, 176.51], [995.57, 176.42], [1008.22, 176.33], [1031.49, 176.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1203.61, 235.47], [1186.23, 233.66], [1199.98, 233.65], [1217.76, 235.45], [1203.61, 159.39], [1186.23, 159.78], [1199.98, 159.78], [1217.76, 159.39]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1086.21, 216.64], [1076.32, 216.02], [1080.09, 215.93], [1090.03, 216.55], [1086.21, 171.33], [1076.32, 171.36], [1080.09, 171.36], [1090.03, 171.34]]}]\n```" - } - ] - }, - { - "id": 3915, - "data_source": "Mono3DRefer", - "scene_name": "007406", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[575.78, 247.6], [642.03, 247.47], [653.99, 266.73], [570.63, 266.93], [575.78, 184.9], [642.03, 184.88], [653.99, 187.98], [570.63, 188.01]]}, {\"category\": \"car\", \"corners_3d\": [[26.88, 240.19], [0.2, 243.77], [-102.28, 244.78], [-69.99, 241.11], [26.88, 204.12], [0.2, 205.78], [-102.28, 206.25], [-69.99, 204.55]]}]\n```" - } - ] - }, - { - "id": 3916, - "data_source": "Mono3DRefer", - "scene_name": "007406", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[438.37, 238.79], [376.83, 238.85], [420.02, 226.18], [469.72, 226.14], [438.37, 176.54], [376.83, 176.54], [420.02, 175.84], [469.72, 175.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[962.56, 243.12], [991.98, 245.27], [974.68, 247.91], [944.75, 245.61], [962.56, 170.17], [991.98, 170.09], [974.68, 169.99], [944.75, 170.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1019.2, 237.24], [995.57, 235.75], [1008.22, 234.05], [1031.49, 235.46], [1019.2, 176.51], [995.57, 176.42], [1008.22, 176.33], [1031.49, 176.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1203.61, 235.47], [1186.23, 233.66], [1199.98, 233.65], [1217.76, 235.45], [1203.61, 159.39], [1186.23, 159.78], [1199.98, 159.78], [1217.76, 159.39]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1086.21, 216.64], [1076.32, 216.02], [1080.09, 215.93], [1090.03, 216.55], [1086.21, 171.33], [1076.32, 171.36], [1080.09, 171.36], [1090.03, 171.34]]}]\n```" - } - ] - }, - { - "id": 3917, - "data_source": "Mono3DRefer", - "scene_name": "007406", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[575.78, 247.6], [642.03, 247.47], [653.99, 266.73], [570.63, 266.93], [575.78, 184.9], [642.03, 184.88], [653.99, 187.98], [570.63, 188.01]]}, {\"category\": \"van\", \"corners_3d\": [[438.37, 238.79], [376.83, 238.85], [420.02, 226.18], [469.72, 226.14], [438.37, 176.54], [376.83, 176.54], [420.02, 175.84], [469.72, 175.83]]}, {\"category\": \"car\", \"corners_3d\": [[26.88, 240.19], [0.2, 243.77], [-102.28, 244.78], [-69.99, 241.11], [26.88, 204.12], [0.2, 205.78], [-102.28, 206.25], [-69.99, 204.55]]}]\n```" - } - ] - }, - { - "id": 3918, - "data_source": "Mono3DRefer", - "scene_name": "007406", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[575.78, 247.6], [642.03, 247.47], [653.99, 266.73], [570.63, 266.93], [575.78, 184.9], [642.03, 184.88], [653.99, 187.98], [570.63, 188.01]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[962.56, 243.12], [991.98, 245.27], [974.68, 247.91], [944.75, 245.61], [962.56, 170.17], [991.98, 170.09], [974.68, 169.99], [944.75, 170.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1019.2, 237.24], [995.57, 235.75], [1008.22, 234.05], [1031.49, 235.46], [1019.2, 176.51], [995.57, 176.42], [1008.22, 176.33], [1031.49, 176.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1203.61, 235.47], [1186.23, 233.66], [1199.98, 233.65], [1217.76, 235.45], [1203.61, 159.39], [1186.23, 159.78], [1199.98, 159.78], [1217.76, 159.39]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1086.21, 216.64], [1076.32, 216.02], [1080.09, 215.93], [1090.03, 216.55], [1086.21, 171.33], [1076.32, 171.36], [1080.09, 171.36], [1090.03, 171.34]]}, {\"category\": \"car\", \"corners_3d\": [[26.88, 240.19], [0.2, 243.77], [-102.28, 244.78], [-69.99, 241.11], [26.88, 204.12], [0.2, 205.78], [-102.28, 206.25], [-69.99, 204.55]]}]\n```" - } - ] - }, - { - "id": 3919, - "data_source": "Mono3DRefer", - "scene_name": "007406", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[575.78, 247.6], [642.03, 247.47], [653.99, 266.73], [570.63, 266.93], [575.78, 184.9], [642.03, 184.88], [653.99, 187.98], [570.63, 188.01]]}, {\"category\": \"van\", \"corners_3d\": [[438.37, 238.79], [376.83, 238.85], [420.02, 226.18], [469.72, 226.14], [438.37, 176.54], [376.83, 176.54], [420.02, 175.84], [469.72, 175.83]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[962.56, 243.12], [991.98, 245.27], [974.68, 247.91], [944.75, 245.61], [962.56, 170.17], [991.98, 170.09], [974.68, 169.99], [944.75, 170.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1019.2, 237.24], [995.57, 235.75], [1008.22, 234.05], [1031.49, 235.46], [1019.2, 176.51], [995.57, 176.42], [1008.22, 176.33], [1031.49, 176.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1203.61, 235.47], [1186.23, 233.66], [1199.98, 233.65], [1217.76, 235.45], [1203.61, 159.39], [1186.23, 159.78], [1199.98, 159.78], [1217.76, 159.39]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1086.21, 216.64], [1076.32, 216.02], [1080.09, 215.93], [1090.03, 216.55], [1086.21, 171.33], [1076.32, 171.36], [1080.09, 171.36], [1090.03, 171.34]]}, {\"category\": \"car\", \"corners_3d\": [[26.88, 240.19], [0.2, 243.77], [-102.28, 244.78], [-69.99, 241.11], [26.88, 204.12], [0.2, 205.78], [-102.28, 206.25], [-69.99, 204.55]]}]\n```" - } - ] - }, - { - "id": 3920, - "data_source": "Mono3DRefer", - "scene_name": "007410", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[745.8, 279.85], [807.4, 278.91], [834.87, 288.27], [767.58, 289.41], [745.8, 166.64], [807.4, 166.78], [834.87, 165.37], [767.58, 165.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[641.04, 223.53], [673.44, 223.38], [676.81, 224.46], [643.58, 224.62], [641.04, 162.09], [673.44, 162.16], [676.81, 161.66], [643.58, 161.59]]}]\n```" - } - ] - }, - { - "id": 3921, - "data_source": "Mono3DRefer", - "scene_name": "007410", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[477.51, 228.16], [445.91, 228.28], [455.21, 224.35], [484.14, 224.25], [477.51, 164.91], [445.91, 164.87], [455.21, 166.26], [484.14, 166.3]]}, {\"category\": \"cyclist\", \"corners_3d\": [[425.82, 238.17], [396.8, 238.31], [413.62, 232.33], [439.56, 232.22], [425.82, 163.69], [396.8, 163.65], [413.62, 165.53], [439.56, 165.57]]}]\n```" - } - ] - }, - { - "id": 3922, - "data_source": "Mono3DRefer", - "scene_name": "007410", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[541.17, 195.3], [580.12, 195.24], [586.78, 197.73], [540.77, 197.81], [541.17, 146.85], [580.12, 147.0], [586.78, 140.74], [540.77, 140.53]]}]\n```" - } - ] - }, - { - "id": 3923, - "data_source": "Mono3DRefer", - "scene_name": "007410", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[745.8, 279.85], [807.4, 278.91], [834.87, 288.27], [767.58, 289.41], [745.8, 166.64], [807.4, 166.78], [834.87, 165.37], [767.58, 165.2]]}, {\"category\": \"cyclist\", \"corners_3d\": [[477.51, 228.16], [445.91, 228.28], [455.21, 224.35], [484.14, 224.25], [477.51, 164.91], [445.91, 164.87], [455.21, 166.26], [484.14, 166.3]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[641.04, 223.53], [673.44, 223.38], [676.81, 224.46], [643.58, 224.62], [641.04, 162.09], [673.44, 162.16], [676.81, 161.66], [643.58, 161.59]]}, {\"category\": \"cyclist\", \"corners_3d\": [[425.82, 238.17], [396.8, 238.31], [413.62, 232.33], [439.56, 232.22], [425.82, 163.69], [396.8, 163.65], [413.62, 165.53], [439.56, 165.57]]}]\n```" - } - ] - }, - { - "id": 3924, - "data_source": "Mono3DRefer", - "scene_name": "007410", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[745.8, 279.85], [807.4, 278.91], [834.87, 288.27], [767.58, 289.41], [745.8, 166.64], [807.4, 166.78], [834.87, 165.37], [767.58, 165.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[641.04, 223.53], [673.44, 223.38], [676.81, 224.46], [643.58, 224.62], [641.04, 162.09], [673.44, 162.16], [676.81, 161.66], [643.58, 161.59]]}, {\"category\": \"van\", \"corners_3d\": [[541.17, 195.3], [580.12, 195.24], [586.78, 197.73], [540.77, 197.81], [541.17, 146.85], [580.12, 147.0], [586.78, 140.74], [540.77, 140.53]]}]\n```" - } - ] - }, - { - "id": 3925, - "data_source": "Mono3DRefer", - "scene_name": "007410", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[477.51, 228.16], [445.91, 228.28], [455.21, 224.35], [484.14, 224.25], [477.51, 164.91], [445.91, 164.87], [455.21, 166.26], [484.14, 166.3]]}, {\"category\": \"cyclist\", \"corners_3d\": [[425.82, 238.17], [396.8, 238.31], [413.62, 232.33], [439.56, 232.22], [425.82, 163.69], [396.8, 163.65], [413.62, 165.53], [439.56, 165.57]]}, {\"category\": \"van\", \"corners_3d\": [[541.17, 195.3], [580.12, 195.24], [586.78, 197.73], [540.77, 197.81], [541.17, 146.85], [580.12, 147.0], [586.78, 140.74], [540.77, 140.53]]}]\n```" - } - ] - }, - { - "id": 3926, - "data_source": "Mono3DRefer", - "scene_name": "007410", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[745.8, 279.85], [807.4, 278.91], [834.87, 288.27], [767.58, 289.41], [745.8, 166.64], [807.4, 166.78], [834.87, 165.37], [767.58, 165.2]]}, {\"category\": \"cyclist\", \"corners_3d\": [[477.51, 228.16], [445.91, 228.28], [455.21, 224.35], [484.14, 224.25], [477.51, 164.91], [445.91, 164.87], [455.21, 166.26], [484.14, 166.3]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[641.04, 223.53], [673.44, 223.38], [676.81, 224.46], [643.58, 224.62], [641.04, 162.09], [673.44, 162.16], [676.81, 161.66], [643.58, 161.59]]}, {\"category\": \"cyclist\", \"corners_3d\": [[425.82, 238.17], [396.8, 238.31], [413.62, 232.33], [439.56, 232.22], [425.82, 163.69], [396.8, 163.65], [413.62, 165.53], [439.56, 165.57]]}, {\"category\": \"van\", \"corners_3d\": [[541.17, 195.3], [580.12, 195.24], [586.78, 197.73], [540.77, 197.81], [541.17, 146.85], [580.12, 147.0], [586.78, 140.74], [540.77, 140.53]]}]\n```" - } - ] - }, - { - "id": 3927, - "data_source": "Mono3DRefer", - "scene_name": "007412", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[427.72, 254.33], [436.4, 251.76], [470.73, 252.11], [463.28, 254.71], [427.72, 188.3], [436.4, 188.06], [470.73, 188.09], [463.28, 188.34]]}]\n```" - } - ] - }, - { - "id": 3928, - "data_source": "Mono3DRefer", - "scene_name": "007412", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[536.73, 245.34], [505.81, 245.45], [512.48, 238.92], [540.23, 238.83], [536.73, 166.52], [505.81, 166.49], [512.48, 168.03], [540.23, 168.05]]}]\n```" - } - ] - }, - { - "id": 3929, - "data_source": "Mono3DRefer", - "scene_name": "007412", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[851.62, 278.03], [917.88, 279.8], [940.07, 293.11], [864.57, 290.83], [851.62, 145.92], [917.88, 145.27], [940.07, 140.36], [864.57, 141.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[580.44, 256.04], [622.1, 259.33], [590.11, 264.18], [547.41, 260.47], [580.44, 160.35], [622.1, 159.41], [590.11, 158.04], [547.41, 159.09]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[288.92, 277.73], [291.59, 272.37], [347.55, 270.13], [348.16, 275.22], [288.92, 161.21], [291.59, 162.34], [347.55, 162.81], [348.16, 161.74]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[499.39, 262.48], [454.56, 262.23], [466.02, 257.12], [508.05, 257.34], [499.39, 159.32], [454.56, 159.39], [466.02, 160.79], [508.05, 160.73]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[799.3, 261.1], [751.08, 261.26], [736.9, 254.72], [781.19, 254.58], [799.3, 144.07], [751.08, 143.99], [736.9, 147.07], [781.19, 147.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[535.68, 213.04], [555.59, 212.98], [555.99, 214.0], [535.43, 214.08], [535.68, 163.78], [555.59, 163.82], [555.99, 163.24], [535.43, 163.2]]}]\n```" - } - ] - }, - { - "id": 3930, - "data_source": "Mono3DRefer", - "scene_name": "007412", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[536.73, 245.34], [505.81, 245.45], [512.48, 238.92], [540.23, 238.83], [536.73, 166.52], [505.81, 166.49], [512.48, 168.03], [540.23, 168.05]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[427.72, 254.33], [436.4, 251.76], [470.73, 252.11], [463.28, 254.71], [427.72, 188.3], [436.4, 188.06], [470.73, 188.09], [463.28, 188.34]]}]\n```" - } - ] - }, - { - "id": 3931, - "data_source": "Mono3DRefer", - "scene_name": "007412", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[851.62, 278.03], [917.88, 279.8], [940.07, 293.11], [864.57, 290.83], [851.62, 145.92], [917.88, 145.27], [940.07, 140.36], [864.57, 141.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[580.44, 256.04], [622.1, 259.33], [590.11, 264.18], [547.41, 260.47], [580.44, 160.35], [622.1, 159.41], [590.11, 158.04], [547.41, 159.09]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[288.92, 277.73], [291.59, 272.37], [347.55, 270.13], [348.16, 275.22], [288.92, 161.21], [291.59, 162.34], [347.55, 162.81], [348.16, 161.74]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[499.39, 262.48], [454.56, 262.23], [466.02, 257.12], [508.05, 257.34], [499.39, 159.32], [454.56, 159.39], [466.02, 160.79], [508.05, 160.73]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[799.3, 261.1], [751.08, 261.26], [736.9, 254.72], [781.19, 254.58], [799.3, 144.07], [751.08, 143.99], [736.9, 147.07], [781.19, 147.14]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[427.72, 254.33], [436.4, 251.76], [470.73, 252.11], [463.28, 254.71], [427.72, 188.3], [436.4, 188.06], [470.73, 188.09], [463.28, 188.34]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[535.68, 213.04], [555.59, 212.98], [555.99, 214.0], [535.43, 214.08], [535.68, 163.78], [555.59, 163.82], [555.99, 163.24], [535.43, 163.2]]}]\n```" - } - ] - }, - { - "id": 3932, - "data_source": "Mono3DRefer", - "scene_name": "007412", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[851.62, 278.03], [917.88, 279.8], [940.07, 293.11], [864.57, 290.83], [851.62, 145.92], [917.88, 145.27], [940.07, 140.36], [864.57, 141.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[580.44, 256.04], [622.1, 259.33], [590.11, 264.18], [547.41, 260.47], [580.44, 160.35], [622.1, 159.41], [590.11, 158.04], [547.41, 159.09]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[288.92, 277.73], [291.59, 272.37], [347.55, 270.13], [348.16, 275.22], [288.92, 161.21], [291.59, 162.34], [347.55, 162.81], [348.16, 161.74]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[499.39, 262.48], [454.56, 262.23], [466.02, 257.12], [508.05, 257.34], [499.39, 159.32], [454.56, 159.39], [466.02, 160.79], [508.05, 160.73]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[799.3, 261.1], [751.08, 261.26], [736.9, 254.72], [781.19, 254.58], [799.3, 144.07], [751.08, 143.99], [736.9, 147.07], [781.19, 147.14]]}, {\"category\": \"cyclist\", \"corners_3d\": [[536.73, 245.34], [505.81, 245.45], [512.48, 238.92], [540.23, 238.83], [536.73, 166.52], [505.81, 166.49], [512.48, 168.03], [540.23, 168.05]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[535.68, 213.04], [555.59, 212.98], [555.99, 214.0], [535.43, 214.08], [535.68, 163.78], [555.59, 163.82], [555.99, 163.24], [535.43, 163.2]]}]\n```" - } - ] - }, - { - "id": 3933, - "data_source": "Mono3DRefer", - "scene_name": "007412", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (person_sitting, cyclist, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[851.62, 278.03], [917.88, 279.8], [940.07, 293.11], [864.57, 290.83], [851.62, 145.92], [917.88, 145.27], [940.07, 140.36], [864.57, 141.2]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[580.44, 256.04], [622.1, 259.33], [590.11, 264.18], [547.41, 260.47], [580.44, 160.35], [622.1, 159.41], [590.11, 158.04], [547.41, 159.09]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[288.92, 277.73], [291.59, 272.37], [347.55, 270.13], [348.16, 275.22], [288.92, 161.21], [291.59, 162.34], [347.55, 162.81], [348.16, 161.74]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[499.39, 262.48], [454.56, 262.23], [466.02, 257.12], [508.05, 257.34], [499.39, 159.32], [454.56, 159.39], [466.02, 160.79], [508.05, 160.73]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[799.3, 261.1], [751.08, 261.26], [736.9, 254.72], [781.19, 254.58], [799.3, 144.07], [751.08, 143.99], [736.9, 147.07], [781.19, 147.14]]}, {\"category\": \"cyclist\", \"corners_3d\": [[536.73, 245.34], [505.81, 245.45], [512.48, 238.92], [540.23, 238.83], [536.73, 166.52], [505.81, 166.49], [512.48, 168.03], [540.23, 168.05]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[427.72, 254.33], [436.4, 251.76], [470.73, 252.11], [463.28, 254.71], [427.72, 188.3], [436.4, 188.06], [470.73, 188.09], [463.28, 188.34]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[535.68, 213.04], [555.59, 212.98], [555.99, 214.0], [535.43, 214.08], [535.68, 163.78], [555.59, 163.82], [555.99, 163.24], [535.43, 163.2]]}]\n```" - } - ] - }, - { - "id": 3934, - "data_source": "Mono3DRefer", - "scene_name": "007413", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[939.58, 224.97], [969.77, 229.35], [826.32, 229.81], [807.11, 225.36], [939.58, 167.52], [969.77, 167.07], [826.32, 167.03], [807.11, 167.48]]}, {\"category\": \"car\", \"corners_3d\": [[620.77, 194.59], [646.4, 194.55], [653.05, 196.54], [625.07, 196.59], [620.77, 170.19], [646.4, 170.2], [653.05, 169.95], [625.07, 169.95]]}]\n```" - } - ] - }, - { - "id": 3935, - "data_source": "Mono3DRefer", - "scene_name": "007414", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[292.94, 279.43], [194.47, 281.5], [268.95, 255.01], [342.68, 253.82], [292.94, 184.62], [194.47, 184.85], [268.95, 181.92], [342.68, 181.79]]}, {\"category\": \"van\", \"corners_3d\": [[483.35, 203.73], [513.98, 203.51], [517.07, 206.93], [483.02, 207.2], [483.35, 170.34], [513.98, 170.36], [517.07, 170.08], [483.02, 170.06]]}]\n```" - } - ] - }, - { - "id": 3936, - "data_source": "Mono3DRefer", - "scene_name": "007414", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[830.0, 241.89], [793.44, 242.49], [776.78, 238.63], [811.46, 238.1], [830.0, 160.28], [793.44, 160.17], [776.78, 160.88], [811.46, 160.97]]}]\n```" - } - ] - }, - { - "id": 3937, - "data_source": "Mono3DRefer", - "scene_name": "007414", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[524.75, 258.49], [565.86, 257.63], [577.92, 269.38], [531.19, 270.5], [524.75, 170.82], [565.86, 170.84], [577.92, 170.56], [531.19, 170.54]]}]\n```" - } - ] - }, - { - "id": 3938, - "data_source": "Mono3DRefer", - "scene_name": "007414", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1087.96, 354.76], [1370.61, 383.88], [2138.33, 838.06], [1193.89, 614.98], [1087.96, 172.81], [1370.61, 172.8], [2138.33, 172.7], [1193.89, 172.75]]}, {\"category\": \"car\", \"corners_3d\": [[695.11, 287.97], [805.03, 285.28], [995.95, 355.91], [821.79, 363.13], [695.11, 180.89], [805.03, 180.71], [995.95, 185.64], [821.79, 186.15]]}, {\"category\": \"car\", \"corners_3d\": [[152.39, 337.81], [-5.52, 341.56], [170.45, 286.05], [274.54, 284.35], [152.39, 208.24], [-5.52, 209.04], [170.45, 197.14], [274.54, 196.77]]}, {\"category\": \"car\", \"corners_3d\": [[608.02, 242.26], [680.33, 241.21], [727.92, 259.36], [636.96, 261.04], [608.02, 172.4], [680.33, 172.41], [727.92, 172.29], [636.96, 172.28]]}, {\"category\": \"car\", \"corners_3d\": [[572.18, 228.43], [628.33, 227.56], [662.77, 239.02], [595.3, 240.29], [572.18, 177.83], [628.33, 177.76], [662.77, 178.78], [595.3, 178.9]]}, {\"category\": \"car\", \"corners_3d\": [[352.94, 250.73], [287.31, 252.07], [316.65, 238.82], [371.04, 237.89], [352.94, 195.48], [287.31, 195.87], [316.65, 192.02], [371.04, 191.75]]}, {\"category\": \"car\", \"corners_3d\": [[373.49, 232.13], [326.54, 232.79], [346.35, 224.94], [387.01, 224.43], [373.49, 190.24], [326.54, 190.44], [346.35, 188.13], [387.01, 187.99]]}, {\"category\": \"car\", \"corners_3d\": [[554.66, 217.64], [594.91, 217.2], [611.82, 223.66], [565.83, 224.25], [554.66, 180.67], [594.91, 180.59], [611.82, 181.72], [565.83, 181.82]]}]\n```" - } - ] - }, - { - "id": 3939, - "data_source": "Mono3DRefer", - "scene_name": "007414", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[292.94, 279.43], [194.47, 281.5], [268.95, 255.01], [342.68, 253.82], [292.94, 184.62], [194.47, 184.85], [268.95, 181.92], [342.68, 181.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[830.0, 241.89], [793.44, 242.49], [776.78, 238.63], [811.46, 238.1], [830.0, 160.28], [793.44, 160.17], [776.78, 160.88], [811.46, 160.97]]}, {\"category\": \"van\", \"corners_3d\": [[483.35, 203.73], [513.98, 203.51], [517.07, 206.93], [483.02, 207.2], [483.35, 170.34], [513.98, 170.36], [517.07, 170.08], [483.02, 170.06]]}]\n```" - } - ] - }, - { - "id": 3940, - "data_source": "Mono3DRefer", - "scene_name": "007414", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[524.75, 258.49], [565.86, 257.63], [577.92, 269.38], [531.19, 270.5], [524.75, 170.82], [565.86, 170.84], [577.92, 170.56], [531.19, 170.54]]}, {\"category\": \"van\", \"corners_3d\": [[292.94, 279.43], [194.47, 281.5], [268.95, 255.01], [342.68, 253.82], [292.94, 184.62], [194.47, 184.85], [268.95, 181.92], [342.68, 181.79]]}, {\"category\": \"van\", \"corners_3d\": [[483.35, 203.73], [513.98, 203.51], [517.07, 206.93], [483.02, 207.2], [483.35, 170.34], [513.98, 170.36], [517.07, 170.08], [483.02, 170.06]]}]\n```" - } - ] - }, - { - "id": 3941, - "data_source": "Mono3DRefer", - "scene_name": "007414", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1087.96, 354.76], [1370.61, 383.88], [2138.33, 838.06], [1193.89, 614.98], [1087.96, 172.81], [1370.61, 172.8], [2138.33, 172.7], [1193.89, 172.75]]}, {\"category\": \"car\", \"corners_3d\": [[695.11, 287.97], [805.03, 285.28], [995.95, 355.91], [821.79, 363.13], [695.11, 180.89], [805.03, 180.71], [995.95, 185.64], [821.79, 186.15]]}, {\"category\": \"car\", \"corners_3d\": [[152.39, 337.81], [-5.52, 341.56], [170.45, 286.05], [274.54, 284.35], [152.39, 208.24], [-5.52, 209.04], [170.45, 197.14], [274.54, 196.77]]}, {\"category\": \"van\", \"corners_3d\": [[292.94, 279.43], [194.47, 281.5], [268.95, 255.01], [342.68, 253.82], [292.94, 184.62], [194.47, 184.85], [268.95, 181.92], [342.68, 181.79]]}, {\"category\": \"car\", \"corners_3d\": [[608.02, 242.26], [680.33, 241.21], [727.92, 259.36], [636.96, 261.04], [608.02, 172.4], [680.33, 172.41], [727.92, 172.29], [636.96, 172.28]]}, {\"category\": \"car\", \"corners_3d\": [[572.18, 228.43], [628.33, 227.56], [662.77, 239.02], [595.3, 240.29], [572.18, 177.83], [628.33, 177.76], [662.77, 178.78], [595.3, 178.9]]}, {\"category\": \"car\", \"corners_3d\": [[352.94, 250.73], [287.31, 252.07], [316.65, 238.82], [371.04, 237.89], [352.94, 195.48], [287.31, 195.87], [316.65, 192.02], [371.04, 191.75]]}, {\"category\": \"car\", \"corners_3d\": [[373.49, 232.13], [326.54, 232.79], [346.35, 224.94], [387.01, 224.43], [373.49, 190.24], [326.54, 190.44], [346.35, 188.13], [387.01, 187.99]]}, {\"category\": \"car\", \"corners_3d\": [[554.66, 217.64], [594.91, 217.2], [611.82, 223.66], [565.83, 224.25], [554.66, 180.67], [594.91, 180.59], [611.82, 181.72], [565.83, 181.82]]}, {\"category\": \"van\", \"corners_3d\": [[483.35, 203.73], [513.98, 203.51], [517.07, 206.93], [483.02, 207.2], [483.35, 170.34], [513.98, 170.36], [517.07, 170.08], [483.02, 170.06]]}]\n```" - } - ] - }, - { - "id": 3942, - "data_source": "Mono3DRefer", - "scene_name": "007414", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[524.75, 258.49], [565.86, 257.63], [577.92, 269.38], [531.19, 270.5], [524.75, 170.82], [565.86, 170.84], [577.92, 170.56], [531.19, 170.54]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[830.0, 241.89], [793.44, 242.49], [776.78, 238.63], [811.46, 238.1], [830.0, 160.28], [793.44, 160.17], [776.78, 160.88], [811.46, 160.97]]}]\n```" - } - ] - }, - { - "id": 3943, - "data_source": "Mono3DRefer", - "scene_name": "007414", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1087.96, 354.76], [1370.61, 383.88], [2138.33, 838.06], [1193.89, 614.98], [1087.96, 172.81], [1370.61, 172.8], [2138.33, 172.7], [1193.89, 172.75]]}, {\"category\": \"car\", \"corners_3d\": [[695.11, 287.97], [805.03, 285.28], [995.95, 355.91], [821.79, 363.13], [695.11, 180.89], [805.03, 180.71], [995.95, 185.64], [821.79, 186.15]]}, {\"category\": \"car\", \"corners_3d\": [[152.39, 337.81], [-5.52, 341.56], [170.45, 286.05], [274.54, 284.35], [152.39, 208.24], [-5.52, 209.04], [170.45, 197.14], [274.54, 196.77]]}, {\"category\": \"car\", \"corners_3d\": [[608.02, 242.26], [680.33, 241.21], [727.92, 259.36], [636.96, 261.04], [608.02, 172.4], [680.33, 172.41], [727.92, 172.29], [636.96, 172.28]]}, {\"category\": \"car\", \"corners_3d\": [[572.18, 228.43], [628.33, 227.56], [662.77, 239.02], [595.3, 240.29], [572.18, 177.83], [628.33, 177.76], [662.77, 178.78], [595.3, 178.9]]}, {\"category\": \"car\", \"corners_3d\": [[352.94, 250.73], [287.31, 252.07], [316.65, 238.82], [371.04, 237.89], [352.94, 195.48], [287.31, 195.87], [316.65, 192.02], [371.04, 191.75]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[830.0, 241.89], [793.44, 242.49], [776.78, 238.63], [811.46, 238.1], [830.0, 160.28], [793.44, 160.17], [776.78, 160.88], [811.46, 160.97]]}, {\"category\": \"car\", \"corners_3d\": [[373.49, 232.13], [326.54, 232.79], [346.35, 224.94], [387.01, 224.43], [373.49, 190.24], [326.54, 190.44], [346.35, 188.13], [387.01, 187.99]]}, {\"category\": \"car\", \"corners_3d\": [[554.66, 217.64], [594.91, 217.2], [611.82, 223.66], [565.83, 224.25], [554.66, 180.67], [594.91, 180.59], [611.82, 181.72], [565.83, 181.82]]}]\n```" - } - ] - }, - { - "id": 3944, - "data_source": "Mono3DRefer", - "scene_name": "007414", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[524.75, 258.49], [565.86, 257.63], [577.92, 269.38], [531.19, 270.5], [524.75, 170.82], [565.86, 170.84], [577.92, 170.56], [531.19, 170.54]]}, {\"category\": \"car\", \"corners_3d\": [[1087.96, 354.76], [1370.61, 383.88], [2138.33, 838.06], [1193.89, 614.98], [1087.96, 172.81], [1370.61, 172.8], [2138.33, 172.7], [1193.89, 172.75]]}, {\"category\": \"car\", \"corners_3d\": [[695.11, 287.97], [805.03, 285.28], [995.95, 355.91], [821.79, 363.13], [695.11, 180.89], [805.03, 180.71], [995.95, 185.64], [821.79, 186.15]]}, {\"category\": \"car\", \"corners_3d\": [[152.39, 337.81], [-5.52, 341.56], [170.45, 286.05], [274.54, 284.35], [152.39, 208.24], [-5.52, 209.04], [170.45, 197.14], [274.54, 196.77]]}, {\"category\": \"car\", \"corners_3d\": [[608.02, 242.26], [680.33, 241.21], [727.92, 259.36], [636.96, 261.04], [608.02, 172.4], [680.33, 172.41], [727.92, 172.29], [636.96, 172.28]]}, {\"category\": \"car\", \"corners_3d\": [[572.18, 228.43], [628.33, 227.56], [662.77, 239.02], [595.3, 240.29], [572.18, 177.83], [628.33, 177.76], [662.77, 178.78], [595.3, 178.9]]}, {\"category\": \"car\", \"corners_3d\": [[352.94, 250.73], [287.31, 252.07], [316.65, 238.82], [371.04, 237.89], [352.94, 195.48], [287.31, 195.87], [316.65, 192.02], [371.04, 191.75]]}, {\"category\": \"car\", \"corners_3d\": [[373.49, 232.13], [326.54, 232.79], [346.35, 224.94], [387.01, 224.43], [373.49, 190.24], [326.54, 190.44], [346.35, 188.13], [387.01, 187.99]]}, {\"category\": \"car\", \"corners_3d\": [[554.66, 217.64], [594.91, 217.2], [611.82, 223.66], [565.83, 224.25], [554.66, 180.67], [594.91, 180.59], [611.82, 181.72], [565.83, 181.82]]}]\n```" - } - ] - }, - { - "id": 3945, - "data_source": "Mono3DRefer", - "scene_name": "007414", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[524.75, 258.49], [565.86, 257.63], [577.92, 269.38], [531.19, 270.5], [524.75, 170.82], [565.86, 170.84], [577.92, 170.56], [531.19, 170.54]]}, {\"category\": \"van\", \"corners_3d\": [[292.94, 279.43], [194.47, 281.5], [268.95, 255.01], [342.68, 253.82], [292.94, 184.62], [194.47, 184.85], [268.95, 181.92], [342.68, 181.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[830.0, 241.89], [793.44, 242.49], [776.78, 238.63], [811.46, 238.1], [830.0, 160.28], [793.44, 160.17], [776.78, 160.88], [811.46, 160.97]]}, {\"category\": \"van\", \"corners_3d\": [[483.35, 203.73], [513.98, 203.51], [517.07, 206.93], [483.02, 207.2], [483.35, 170.34], [513.98, 170.36], [517.07, 170.08], [483.02, 170.06]]}]\n```" - } - ] - }, - { - "id": 3946, - "data_source": "Mono3DRefer", - "scene_name": "007414", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1087.96, 354.76], [1370.61, 383.88], [2138.33, 838.06], [1193.89, 614.98], [1087.96, 172.81], [1370.61, 172.8], [2138.33, 172.7], [1193.89, 172.75]]}, {\"category\": \"car\", \"corners_3d\": [[695.11, 287.97], [805.03, 285.28], [995.95, 355.91], [821.79, 363.13], [695.11, 180.89], [805.03, 180.71], [995.95, 185.64], [821.79, 186.15]]}, {\"category\": \"car\", \"corners_3d\": [[152.39, 337.81], [-5.52, 341.56], [170.45, 286.05], [274.54, 284.35], [152.39, 208.24], [-5.52, 209.04], [170.45, 197.14], [274.54, 196.77]]}, {\"category\": \"van\", \"corners_3d\": [[292.94, 279.43], [194.47, 281.5], [268.95, 255.01], [342.68, 253.82], [292.94, 184.62], [194.47, 184.85], [268.95, 181.92], [342.68, 181.79]]}, {\"category\": \"car\", \"corners_3d\": [[608.02, 242.26], [680.33, 241.21], [727.92, 259.36], [636.96, 261.04], [608.02, 172.4], [680.33, 172.41], [727.92, 172.29], [636.96, 172.28]]}, {\"category\": \"car\", \"corners_3d\": [[572.18, 228.43], [628.33, 227.56], [662.77, 239.02], [595.3, 240.29], [572.18, 177.83], [628.33, 177.76], [662.77, 178.78], [595.3, 178.9]]}, {\"category\": \"car\", \"corners_3d\": [[352.94, 250.73], [287.31, 252.07], [316.65, 238.82], [371.04, 237.89], [352.94, 195.48], [287.31, 195.87], [316.65, 192.02], [371.04, 191.75]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[830.0, 241.89], [793.44, 242.49], [776.78, 238.63], [811.46, 238.1], [830.0, 160.28], [793.44, 160.17], [776.78, 160.88], [811.46, 160.97]]}, {\"category\": \"car\", \"corners_3d\": [[373.49, 232.13], [326.54, 232.79], [346.35, 224.94], [387.01, 224.43], [373.49, 190.24], [326.54, 190.44], [346.35, 188.13], [387.01, 187.99]]}, {\"category\": \"car\", \"corners_3d\": [[554.66, 217.64], [594.91, 217.2], [611.82, 223.66], [565.83, 224.25], [554.66, 180.67], [594.91, 180.59], [611.82, 181.72], [565.83, 181.82]]}, {\"category\": \"van\", \"corners_3d\": [[483.35, 203.73], [513.98, 203.51], [517.07, 206.93], [483.02, 207.2], [483.35, 170.34], [513.98, 170.36], [517.07, 170.08], [483.02, 170.06]]}]\n```" - } - ] - }, - { - "id": 3947, - "data_source": "Mono3DRefer", - "scene_name": "007414", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[524.75, 258.49], [565.86, 257.63], [577.92, 269.38], [531.19, 270.5], [524.75, 170.82], [565.86, 170.84], [577.92, 170.56], [531.19, 170.54]]}, {\"category\": \"car\", \"corners_3d\": [[1087.96, 354.76], [1370.61, 383.88], [2138.33, 838.06], [1193.89, 614.98], [1087.96, 172.81], [1370.61, 172.8], [2138.33, 172.7], [1193.89, 172.75]]}, {\"category\": \"car\", \"corners_3d\": [[695.11, 287.97], [805.03, 285.28], [995.95, 355.91], [821.79, 363.13], [695.11, 180.89], [805.03, 180.71], [995.95, 185.64], [821.79, 186.15]]}, {\"category\": \"car\", \"corners_3d\": [[152.39, 337.81], [-5.52, 341.56], [170.45, 286.05], [274.54, 284.35], [152.39, 208.24], [-5.52, 209.04], [170.45, 197.14], [274.54, 196.77]]}, {\"category\": \"van\", \"corners_3d\": [[292.94, 279.43], [194.47, 281.5], [268.95, 255.01], [342.68, 253.82], [292.94, 184.62], [194.47, 184.85], [268.95, 181.92], [342.68, 181.79]]}, {\"category\": \"car\", \"corners_3d\": [[608.02, 242.26], [680.33, 241.21], [727.92, 259.36], [636.96, 261.04], [608.02, 172.4], [680.33, 172.41], [727.92, 172.29], [636.96, 172.28]]}, {\"category\": \"car\", \"corners_3d\": [[572.18, 228.43], [628.33, 227.56], [662.77, 239.02], [595.3, 240.29], [572.18, 177.83], [628.33, 177.76], [662.77, 178.78], [595.3, 178.9]]}, {\"category\": \"car\", \"corners_3d\": [[352.94, 250.73], [287.31, 252.07], [316.65, 238.82], [371.04, 237.89], [352.94, 195.48], [287.31, 195.87], [316.65, 192.02], [371.04, 191.75]]}, {\"category\": \"car\", \"corners_3d\": [[373.49, 232.13], [326.54, 232.79], [346.35, 224.94], [387.01, 224.43], [373.49, 190.24], [326.54, 190.44], [346.35, 188.13], [387.01, 187.99]]}, {\"category\": \"car\", \"corners_3d\": [[554.66, 217.64], [594.91, 217.2], [611.82, 223.66], [565.83, 224.25], [554.66, 180.67], [594.91, 180.59], [611.82, 181.72], [565.83, 181.82]]}, {\"category\": \"van\", \"corners_3d\": [[483.35, 203.73], [513.98, 203.51], [517.07, 206.93], [483.02, 207.2], [483.35, 170.34], [513.98, 170.36], [517.07, 170.08], [483.02, 170.06]]}]\n```" - } - ] - }, - { - "id": 3948, - "data_source": "Mono3DRefer", - "scene_name": "007414", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[524.75, 258.49], [565.86, 257.63], [577.92, 269.38], [531.19, 270.5], [524.75, 170.82], [565.86, 170.84], [577.92, 170.56], [531.19, 170.54]]}, {\"category\": \"car\", \"corners_3d\": [[1087.96, 354.76], [1370.61, 383.88], [2138.33, 838.06], [1193.89, 614.98], [1087.96, 172.81], [1370.61, 172.8], [2138.33, 172.7], [1193.89, 172.75]]}, {\"category\": \"car\", \"corners_3d\": [[695.11, 287.97], [805.03, 285.28], [995.95, 355.91], [821.79, 363.13], [695.11, 180.89], [805.03, 180.71], [995.95, 185.64], [821.79, 186.15]]}, {\"category\": \"car\", \"corners_3d\": [[152.39, 337.81], [-5.52, 341.56], [170.45, 286.05], [274.54, 284.35], [152.39, 208.24], [-5.52, 209.04], [170.45, 197.14], [274.54, 196.77]]}, {\"category\": \"car\", \"corners_3d\": [[608.02, 242.26], [680.33, 241.21], [727.92, 259.36], [636.96, 261.04], [608.02, 172.4], [680.33, 172.41], [727.92, 172.29], [636.96, 172.28]]}, {\"category\": \"car\", \"corners_3d\": [[572.18, 228.43], [628.33, 227.56], [662.77, 239.02], [595.3, 240.29], [572.18, 177.83], [628.33, 177.76], [662.77, 178.78], [595.3, 178.9]]}, {\"category\": \"car\", \"corners_3d\": [[352.94, 250.73], [287.31, 252.07], [316.65, 238.82], [371.04, 237.89], [352.94, 195.48], [287.31, 195.87], [316.65, 192.02], [371.04, 191.75]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[830.0, 241.89], [793.44, 242.49], [776.78, 238.63], [811.46, 238.1], [830.0, 160.28], [793.44, 160.17], [776.78, 160.88], [811.46, 160.97]]}, {\"category\": \"car\", \"corners_3d\": [[373.49, 232.13], [326.54, 232.79], [346.35, 224.94], [387.01, 224.43], [373.49, 190.24], [326.54, 190.44], [346.35, 188.13], [387.01, 187.99]]}, {\"category\": \"car\", \"corners_3d\": [[554.66, 217.64], [594.91, 217.2], [611.82, 223.66], [565.83, 224.25], [554.66, 180.67], [594.91, 180.59], [611.82, 181.72], [565.83, 181.82]]}]\n```" - } - ] - }, - { - "id": 3949, - "data_source": "Mono3DRefer", - "scene_name": "007414", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian, cyclist, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[524.75, 258.49], [565.86, 257.63], [577.92, 269.38], [531.19, 270.5], [524.75, 170.82], [565.86, 170.84], [577.92, 170.56], [531.19, 170.54]]}, {\"category\": \"car\", \"corners_3d\": [[1087.96, 354.76], [1370.61, 383.88], [2138.33, 838.06], [1193.89, 614.98], [1087.96, 172.81], [1370.61, 172.8], [2138.33, 172.7], [1193.89, 172.75]]}, {\"category\": \"car\", \"corners_3d\": [[695.11, 287.97], [805.03, 285.28], [995.95, 355.91], [821.79, 363.13], [695.11, 180.89], [805.03, 180.71], [995.95, 185.64], [821.79, 186.15]]}, {\"category\": \"car\", \"corners_3d\": [[152.39, 337.81], [-5.52, 341.56], [170.45, 286.05], [274.54, 284.35], [152.39, 208.24], [-5.52, 209.04], [170.45, 197.14], [274.54, 196.77]]}, {\"category\": \"van\", \"corners_3d\": [[292.94, 279.43], [194.47, 281.5], [268.95, 255.01], [342.68, 253.82], [292.94, 184.62], [194.47, 184.85], [268.95, 181.92], [342.68, 181.79]]}, {\"category\": \"car\", \"corners_3d\": [[608.02, 242.26], [680.33, 241.21], [727.92, 259.36], [636.96, 261.04], [608.02, 172.4], [680.33, 172.41], [727.92, 172.29], [636.96, 172.28]]}, {\"category\": \"car\", \"corners_3d\": [[572.18, 228.43], [628.33, 227.56], [662.77, 239.02], [595.3, 240.29], [572.18, 177.83], [628.33, 177.76], [662.77, 178.78], [595.3, 178.9]]}, {\"category\": \"car\", \"corners_3d\": [[352.94, 250.73], [287.31, 252.07], [316.65, 238.82], [371.04, 237.89], [352.94, 195.48], [287.31, 195.87], [316.65, 192.02], [371.04, 191.75]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[830.0, 241.89], [793.44, 242.49], [776.78, 238.63], [811.46, 238.1], [830.0, 160.28], [793.44, 160.17], [776.78, 160.88], [811.46, 160.97]]}, {\"category\": \"car\", \"corners_3d\": [[373.49, 232.13], [326.54, 232.79], [346.35, 224.94], [387.01, 224.43], [373.49, 190.24], [326.54, 190.44], [346.35, 188.13], [387.01, 187.99]]}, {\"category\": \"car\", \"corners_3d\": [[554.66, 217.64], [594.91, 217.2], [611.82, 223.66], [565.83, 224.25], [554.66, 180.67], [594.91, 180.59], [611.82, 181.72], [565.83, 181.82]]}, {\"category\": \"van\", \"corners_3d\": [[483.35, 203.73], [513.98, 203.51], [517.07, 206.93], [483.02, 207.2], [483.35, 170.34], [513.98, 170.36], [517.07, 170.08], [483.02, 170.06]]}]\n```" - } - ] - }, - { - "id": 3950, - "data_source": "Mono3DRefer", - "scene_name": "007430", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[836.0, 362.41], [1022.3, 362.45], [1739.85, 692.53], [1229.0, 692.23], [836.0, 194.26], [1022.3, 194.26], [1739.85, 231.53], [1229.0, 231.5]]}, {\"category\": \"car\", \"corners_3d\": [[176.65, 377.95], [356.02, 378.51], [-89.52, 729.12], [-569.25, 725.08], [176.65, 209.28], [356.02, 209.38], [-89.52, 271.65], [-569.25, 270.94]]}, {\"category\": \"car\", \"corners_3d\": [[737.27, 277.43], [834.29, 276.88], [956.89, 327.27], [813.46, 328.5], [737.27, 186.69], [834.29, 186.62], [956.89, 193.29], [813.46, 193.45]]}, {\"category\": \"car\", \"corners_3d\": [[687.99, 246.46], [755.2, 246.12], [806.56, 266.9], [720.49, 267.46], [687.99, 179.81], [755.2, 179.78], [806.56, 181.74], [720.49, 181.79]]}, {\"category\": \"car\", \"corners_3d\": [[473.94, 238.42], [531.81, 238.27], [516.77, 255.63], [443.45, 255.88], [473.94, 182.96], [531.81, 182.93], [516.77, 185.61], [443.45, 185.65]]}, {\"category\": \"car\", \"corners_3d\": [[556.4, 219.89], [510.06, 219.96], [522.04, 213.28], [561.8, 213.23], [556.4, 178.93], [510.06, 178.94], [522.04, 178.08], [561.8, 178.07]]}, {\"category\": \"car\", \"corners_3d\": [[641.32, 202.4], [669.42, 202.4], [675.39, 205.38], [644.47, 205.38], [641.32, 175.77], [669.42, 175.77], [675.39, 176.06], [644.47, 176.06]]}, {\"category\": \"car\", \"corners_3d\": [[580.46, 195.38], [559.0, 195.39], [562.31, 193.65], [582.1, 193.64], [580.46, 174.62], [559.0, 174.62], [562.31, 174.49], [582.1, 174.49]]}]\n```" - } - ] - }, - { - "id": 3951, - "data_source": "Mono3DRefer", - "scene_name": "007430", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[674.11, 232.27], [728.49, 232.14], [758.11, 244.68], [692.27, 244.87], [674.11, 167.75], [728.49, 167.76], [758.11, 166.68], [692.27, 166.67]]}]\n```" - } - ] - }, - { - "id": 3952, - "data_source": "Mono3DRefer", - "scene_name": "007430", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[836.0, 362.41], [1022.3, 362.45], [1739.85, 692.53], [1229.0, 692.23], [836.0, 194.26], [1022.3, 194.26], [1739.85, 231.53], [1229.0, 231.5]]}, {\"category\": \"car\", \"corners_3d\": [[176.65, 377.95], [356.02, 378.51], [-89.52, 729.12], [-569.25, 725.08], [176.65, 209.28], [356.02, 209.38], [-89.52, 271.65], [-569.25, 270.94]]}, {\"category\": \"car\", \"corners_3d\": [[737.27, 277.43], [834.29, 276.88], [956.89, 327.27], [813.46, 328.5], [737.27, 186.69], [834.29, 186.62], [956.89, 193.29], [813.46, 193.45]]}, {\"category\": \"car\", \"corners_3d\": [[687.99, 246.46], [755.2, 246.12], [806.56, 266.9], [720.49, 267.46], [687.99, 179.81], [755.2, 179.78], [806.56, 181.74], [720.49, 181.79]]}, {\"category\": \"car\", \"corners_3d\": [[473.94, 238.42], [531.81, 238.27], [516.77, 255.63], [443.45, 255.88], [473.94, 182.96], [531.81, 182.93], [516.77, 185.61], [443.45, 185.65]]}, {\"category\": \"van\", \"corners_3d\": [[674.11, 232.27], [728.49, 232.14], [758.11, 244.68], [692.27, 244.87], [674.11, 167.75], [728.49, 167.76], [758.11, 166.68], [692.27, 166.67]]}, {\"category\": \"car\", \"corners_3d\": [[556.4, 219.89], [510.06, 219.96], [522.04, 213.28], [561.8, 213.23], [556.4, 178.93], [510.06, 178.94], [522.04, 178.08], [561.8, 178.07]]}, {\"category\": \"car\", \"corners_3d\": [[641.32, 202.4], [669.42, 202.4], [675.39, 205.38], [644.47, 205.38], [641.32, 175.77], [669.42, 175.77], [675.39, 176.06], [644.47, 176.06]]}, {\"category\": \"car\", \"corners_3d\": [[580.46, 195.38], [559.0, 195.39], [562.31, 193.65], [582.1, 193.64], [580.46, 174.62], [559.0, 174.62], [562.31, 174.49], [582.1, 174.49]]}]\n```" - } - ] - }, - { - "id": 3953, - "data_source": "Mono3DRefer", - "scene_name": "007435", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[741.22, 243.24], [802.89, 243.06], [850.12, 258.53], [774.97, 258.79], [741.22, 181.05], [802.89, 181.03], [850.12, 182.83], [774.97, 182.86]]}, {\"category\": \"car\", \"corners_3d\": [[719.34, 228.41], [768.13, 228.6], [790.53, 238.84], [732.71, 238.58], [719.34, 184.02], [768.13, 184.06], [790.53, 186.11], [732.71, 186.06]]}, {\"category\": \"car\", \"corners_3d\": [[703.58, 219.3], [744.32, 219.39], [759.77, 225.77], [713.41, 225.67], [703.58, 182.48], [744.32, 182.5], [759.77, 183.82], [713.41, 183.8]]}, {\"category\": \"car\", \"corners_3d\": [[590.02, 202.66], [568.14, 202.57], [574.21, 200.87], [594.85, 200.94], [590.02, 182.65], [568.14, 182.62], [574.21, 182.06], [594.85, 182.08]]}, {\"category\": \"car\", \"corners_3d\": [[906.4, 336.66], [1078.24, 337.47], [2044.15, 687.94], [1497.35, 680.14], [906.4, 167.07], [1078.24, 167.04], [2044.15, 154.67], [1497.35, 154.94]]}, {\"category\": \"car\", \"corners_3d\": [[517.77, 230.61], [470.57, 230.43], [493.21, 222.96], [534.34, 223.1], [517.77, 188.36], [470.57, 188.31], [493.21, 186.31], [534.34, 186.35]]}, {\"category\": \"car\", \"corners_3d\": [[791.76, 270.29], [884.5, 270.79], [987.63, 311.94], [855.46, 310.92], [791.76, 192.66], [884.5, 192.76], [987.63, 201.12], [855.46, 200.91]]}]\n```" - } - ] - }, - { - "id": 3954, - "data_source": "Mono3DRefer", - "scene_name": "007438", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[598.43, 219.69], [553.75, 219.78], [558.73, 212.82], [596.79, 212.76], [598.43, 180.14], [553.75, 180.15], [558.73, 179.07], [596.79, 179.06]]}]\n```" - } - ] - }, - { - "id": 3955, - "data_source": "Mono3DRefer", - "scene_name": "007469", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[625.6, 190.42], [667.56, 190.44], [676.55, 192.22], [626.21, 192.2], [625.6, 138.4], [667.56, 138.32], [676.55, 129.71], [626.21, 129.82]]}]\n```" - } - ] - }, - { - "id": 3956, - "data_source": "Mono3DRefer", - "scene_name": "007469", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[602.27, 184.4], [573.49, 184.4], [577.73, 184.14], [603.97, 184.15], [602.27, 158.83], [573.49, 158.86], [577.73, 160.87], [603.97, 160.85]]}]\n```" - } - ] - }, - { - "id": 3957, - "data_source": "Mono3DRefer", - "scene_name": "007469", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[706.51, 213.88], [739.16, 213.94], [741.99, 214.78], [708.49, 214.72], [706.51, 153.09], [739.16, 153.03], [741.99, 152.3], [708.49, 152.35]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[805.9, 260.53], [864.45, 260.54], [887.83, 267.55], [824.08, 267.54], [805.9, 156.97], [864.45, 156.97], [887.83, 154.8], [824.08, 154.8]]}]\n```" - } - ] - }, - { - "id": 3958, - "data_source": "Mono3DRefer", - "scene_name": "007469", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[437.92, 298.25], [370.17, 297.47], [417.18, 278.0], [473.76, 278.54], [437.92, 157.69], [370.17, 157.85], [417.18, 161.82], [473.76, 161.71]]}, {\"category\": \"cyclist\", \"corners_3d\": [[292.94, 360.05], [213.99, 359.47], [314.8, 315.43], [374.41, 315.76], [292.94, 150.94], [213.99, 151.04], [314.8, 158.58], [374.41, 158.52]]}]\n```" - } - ] - }, - { - "id": 3959, - "data_source": "Mono3DRefer", - "scene_name": "007469", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[602.27, 184.4], [573.49, 184.4], [577.73, 184.14], [603.97, 184.15], [602.27, 158.83], [573.49, 158.86], [577.73, 160.87], [603.97, 160.85]]}, {\"category\": \"van\", \"corners_3d\": [[625.6, 190.42], [667.56, 190.44], [676.55, 192.22], [626.21, 192.2], [625.6, 138.4], [667.56, 138.32], [676.55, 129.71], [626.21, 129.82]]}]\n```" - } - ] - }, - { - "id": 3960, - "data_source": "Mono3DRefer", - "scene_name": "007469", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[706.51, 213.88], [739.16, 213.94], [741.99, 214.78], [708.49, 214.72], [706.51, 153.09], [739.16, 153.03], [741.99, 152.3], [708.49, 152.35]]}, {\"category\": \"van\", \"corners_3d\": [[625.6, 190.42], [667.56, 190.44], [676.55, 192.22], [626.21, 192.2], [625.6, 138.4], [667.56, 138.32], [676.55, 129.71], [626.21, 129.82]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[805.9, 260.53], [864.45, 260.54], [887.83, 267.55], [824.08, 267.54], [805.9, 156.97], [864.45, 156.97], [887.83, 154.8], [824.08, 154.8]]}]\n```" - } - ] - }, - { - "id": 3961, - "data_source": "Mono3DRefer", - "scene_name": "007469", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[437.92, 298.25], [370.17, 297.47], [417.18, 278.0], [473.76, 278.54], [437.92, 157.69], [370.17, 157.85], [417.18, 161.82], [473.76, 161.71]]}, {\"category\": \"cyclist\", \"corners_3d\": [[292.94, 360.05], [213.99, 359.47], [314.8, 315.43], [374.41, 315.76], [292.94, 150.94], [213.99, 151.04], [314.8, 158.58], [374.41, 158.52]]}, {\"category\": \"van\", \"corners_3d\": [[625.6, 190.42], [667.56, 190.44], [676.55, 192.22], [626.21, 192.2], [625.6, 138.4], [667.56, 138.32], [676.55, 129.71], [626.21, 129.82]]}]\n```" - } - ] - }, - { - "id": 3962, - "data_source": "Mono3DRefer", - "scene_name": "007469", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[602.27, 184.4], [573.49, 184.4], [577.73, 184.14], [603.97, 184.15], [602.27, 158.83], [573.49, 158.86], [577.73, 160.87], [603.97, 160.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[706.51, 213.88], [739.16, 213.94], [741.99, 214.78], [708.49, 214.72], [706.51, 153.09], [739.16, 153.03], [741.99, 152.3], [708.49, 152.35]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[805.9, 260.53], [864.45, 260.54], [887.83, 267.55], [824.08, 267.54], [805.9, 156.97], [864.45, 156.97], [887.83, 154.8], [824.08, 154.8]]}]\n```" - } - ] - }, - { - "id": 3963, - "data_source": "Mono3DRefer", - "scene_name": "007469", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[437.92, 298.25], [370.17, 297.47], [417.18, 278.0], [473.76, 278.54], [437.92, 157.69], [370.17, 157.85], [417.18, 161.82], [473.76, 161.71]]}, {\"category\": \"car\", \"corners_3d\": [[602.27, 184.4], [573.49, 184.4], [577.73, 184.14], [603.97, 184.15], [602.27, 158.83], [573.49, 158.86], [577.73, 160.87], [603.97, 160.85]]}, {\"category\": \"cyclist\", \"corners_3d\": [[292.94, 360.05], [213.99, 359.47], [314.8, 315.43], [374.41, 315.76], [292.94, 150.94], [213.99, 151.04], [314.8, 158.58], [374.41, 158.52]]}]\n```" - } - ] - }, - { - "id": 3964, - "data_source": "Mono3DRefer", - "scene_name": "007469", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[437.92, 298.25], [370.17, 297.47], [417.18, 278.0], [473.76, 278.54], [437.92, 157.69], [370.17, 157.85], [417.18, 161.82], [473.76, 161.71]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[706.51, 213.88], [739.16, 213.94], [741.99, 214.78], [708.49, 214.72], [706.51, 153.09], [739.16, 153.03], [741.99, 152.3], [708.49, 152.35]]}, {\"category\": \"cyclist\", \"corners_3d\": [[292.94, 360.05], [213.99, 359.47], [314.8, 315.43], [374.41, 315.76], [292.94, 150.94], [213.99, 151.04], [314.8, 158.58], [374.41, 158.52]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[805.9, 260.53], [864.45, 260.54], [887.83, 267.55], [824.08, 267.54], [805.9, 156.97], [864.45, 156.97], [887.83, 154.8], [824.08, 154.8]]}]\n```" - } - ] - }, - { - "id": 3965, - "data_source": "Mono3DRefer", - "scene_name": "007469", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, pedestrian). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[602.27, 184.4], [573.49, 184.4], [577.73, 184.14], [603.97, 184.15], [602.27, 158.83], [573.49, 158.86], [577.73, 160.87], [603.97, 160.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[706.51, 213.88], [739.16, 213.94], [741.99, 214.78], [708.49, 214.72], [706.51, 153.09], [739.16, 153.03], [741.99, 152.3], [708.49, 152.35]]}, {\"category\": \"van\", \"corners_3d\": [[625.6, 190.42], [667.56, 190.44], [676.55, 192.22], [626.21, 192.2], [625.6, 138.4], [667.56, 138.32], [676.55, 129.71], [626.21, 129.82]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[805.9, 260.53], [864.45, 260.54], [887.83, 267.55], [824.08, 267.54], [805.9, 156.97], [864.45, 156.97], [887.83, 154.8], [824.08, 154.8]]}]\n```" - } - ] - }, - { - "id": 3966, - "data_source": "Mono3DRefer", - "scene_name": "007469", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[437.92, 298.25], [370.17, 297.47], [417.18, 278.0], [473.76, 278.54], [437.92, 157.69], [370.17, 157.85], [417.18, 161.82], [473.76, 161.71]]}, {\"category\": \"car\", \"corners_3d\": [[602.27, 184.4], [573.49, 184.4], [577.73, 184.14], [603.97, 184.15], [602.27, 158.83], [573.49, 158.86], [577.73, 160.87], [603.97, 160.85]]}, {\"category\": \"cyclist\", \"corners_3d\": [[292.94, 360.05], [213.99, 359.47], [314.8, 315.43], [374.41, 315.76], [292.94, 150.94], [213.99, 151.04], [314.8, 158.58], [374.41, 158.52]]}, {\"category\": \"van\", \"corners_3d\": [[625.6, 190.42], [667.56, 190.44], [676.55, 192.22], [626.21, 192.2], [625.6, 138.4], [667.56, 138.32], [676.55, 129.71], [626.21, 129.82]]}]\n```" - } - ] - }, - { - "id": 3967, - "data_source": "Mono3DRefer", - "scene_name": "007469", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[437.92, 298.25], [370.17, 297.47], [417.18, 278.0], [473.76, 278.54], [437.92, 157.69], [370.17, 157.85], [417.18, 161.82], [473.76, 161.71]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[706.51, 213.88], [739.16, 213.94], [741.99, 214.78], [708.49, 214.72], [706.51, 153.09], [739.16, 153.03], [741.99, 152.3], [708.49, 152.35]]}, {\"category\": \"cyclist\", \"corners_3d\": [[292.94, 360.05], [213.99, 359.47], [314.8, 315.43], [374.41, 315.76], [292.94, 150.94], [213.99, 151.04], [314.8, 158.58], [374.41, 158.52]]}, {\"category\": \"van\", \"corners_3d\": [[625.6, 190.42], [667.56, 190.44], [676.55, 192.22], [626.21, 192.2], [625.6, 138.4], [667.56, 138.32], [676.55, 129.71], [626.21, 129.82]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[805.9, 260.53], [864.45, 260.54], [887.83, 267.55], [824.08, 267.54], [805.9, 156.97], [864.45, 156.97], [887.83, 154.8], [824.08, 154.8]]}]\n```" - } - ] - }, - { - "id": 3968, - "data_source": "Mono3DRefer", - "scene_name": "007469", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car, pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[437.92, 298.25], [370.17, 297.47], [417.18, 278.0], [473.76, 278.54], [437.92, 157.69], [370.17, 157.85], [417.18, 161.82], [473.76, 161.71]]}, {\"category\": \"car\", \"corners_3d\": [[602.27, 184.4], [573.49, 184.4], [577.73, 184.14], [603.97, 184.15], [602.27, 158.83], [573.49, 158.86], [577.73, 160.87], [603.97, 160.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[706.51, 213.88], [739.16, 213.94], [741.99, 214.78], [708.49, 214.72], [706.51, 153.09], [739.16, 153.03], [741.99, 152.3], [708.49, 152.35]]}, {\"category\": \"cyclist\", \"corners_3d\": [[292.94, 360.05], [213.99, 359.47], [314.8, 315.43], [374.41, 315.76], [292.94, 150.94], [213.99, 151.04], [314.8, 158.58], [374.41, 158.52]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[805.9, 260.53], [864.45, 260.54], [887.83, 267.55], [824.08, 267.54], [805.9, 156.97], [864.45, 156.97], [887.83, 154.8], [824.08, 154.8]]}]\n```" - } - ] - }, - { - "id": 3969, - "data_source": "Mono3DRefer", - "scene_name": "007469", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (van, car, pedestrian, cyclist). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[437.92, 298.25], [370.17, 297.47], [417.18, 278.0], [473.76, 278.54], [437.92, 157.69], [370.17, 157.85], [417.18, 161.82], [473.76, 161.71]]}, {\"category\": \"car\", \"corners_3d\": [[602.27, 184.4], [573.49, 184.4], [577.73, 184.14], [603.97, 184.15], [602.27, 158.83], [573.49, 158.86], [577.73, 160.87], [603.97, 160.85]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[706.51, 213.88], [739.16, 213.94], [741.99, 214.78], [708.49, 214.72], [706.51, 153.09], [739.16, 153.03], [741.99, 152.3], [708.49, 152.35]]}, {\"category\": \"cyclist\", \"corners_3d\": [[292.94, 360.05], [213.99, 359.47], [314.8, 315.43], [374.41, 315.76], [292.94, 150.94], [213.99, 151.04], [314.8, 158.58], [374.41, 158.52]]}, {\"category\": \"van\", \"corners_3d\": [[625.6, 190.42], [667.56, 190.44], [676.55, 192.22], [626.21, 192.2], [625.6, 138.4], [667.56, 138.32], [676.55, 129.71], [626.21, 129.82]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[805.9, 260.53], [864.45, 260.54], [887.83, 267.55], [824.08, 267.54], [805.9, 156.97], [864.45, 156.97], [887.83, 154.8], [824.08, 154.8]]}]\n```" - } - ] - }, - { - "id": 3970, - "data_source": "Mono3DRefer", - "scene_name": "007471", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[559.46, 201.89], [532.06, 201.94], [536.05, 199.52], [561.17, 199.49], [559.46, 174.91], [532.06, 174.91], [536.05, 174.74], [561.17, 174.74]]}]\n```" - } - ] - }, - { - "id": 3971, - "data_source": "Mono3DRefer", - "scene_name": "007475", - "question_type": "mobj_3dBbox_json", - "conversations": [ - { - "from": "human", - "value": "Please identify all target objects in this scene from the following categories: (car). Ignore objects that are too small or too far away to be recognized. Given the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). Based on the provided information, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the response JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```" - }, - { - "from": "gpt", - "value": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[685.64, 227.0], [737.22, 227.57], [735.62, 235.24], [674.87, 234.45], [685.64, 180.86], [737.22, 180.8], [735.62, 180.0], [674.87, 180.08]]}, {\"category\": \"car\", \"corners_3d\": [[156.77, 289.25], [257.35, 287.89], [128.34, 335.48], [-22.31, 338.42], [156.77, 197.57], [257.35, 197.41], [128.34, 203.06], [-22.31, 203.41]]}, {\"category\": \"car\", \"corners_3d\": [[220.2, 243.59], [271.12, 243.12], [218.63, 255.28], [156.4, 255.96], [220.2, 197.17], [271.12, 197.07], [218.63, 199.56], [156.4, 199.7]]}, {\"category\": \"car\", \"corners_3d\": [[321.06, 241.73], [375.85, 241.36], [339.01, 253.75], [271.73, 254.3], [321.06, 194.43], [375.85, 194.37], [339.01, 196.39], [271.73, 196.48]]}, {\"category\": \"car\", \"corners_3d\": [[21.65, 265.73], [81.24, 265.11], [-14.31, 281.92], [-87.45, 282.82], [21.65, 212.73], [81.24, 212.52], [-14.31, 218.27], [-87.45, 218.57]]}, {\"category\": \"car\", \"corners_3d\": [[776.71, 202.96], [798.34, 203.21], [786.22, 204.34], [763.43, 204.05], [776.71, 183.82], [798.34, 183.8], [786.22, 183.72], [763.43, 183.74]]}, {\"category\": \"car\", \"corners_3d\": [[845.1, 196.11], [862.05, 196.24], [852.1, 196.75], [834.51, 196.6], [845.1, 182.13], [862.05, 182.09], [852.1, 181.95], [834.51, 181.99]]}]\n```" - } - ] - } -] \ No newline at end of file