diff --git "a/val/qa_mobj_3dBbox_json.json" "b/val/qa_mobj_3dBbox_json.json" deleted file mode 100644--- "a/val/qa_mobj_3dBbox_json.json" +++ /dev/null @@ -1,6959 +0,0 @@ -[ - { - "dataset": "Mono3DRefer", - "scene_name": "000000", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[808.69, 300.53], [820.29, 307.59], [716.27, 307.4], [710.44, 300.37], [808.69, 146.03], [820.29, 144.0], [716.27, 144.06], [710.44, 146.08]]}]\n```", - "options": null, - "id": 0 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000003", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[727.9, 286.51], [615.61, 285.64], [623.58, 255.16], [705.39, 255.62], [727.9, 184.52], [615.61, 184.43], [623.58, 181.3], [705.39, 181.35]]}]\n```", - "options": null, - "id": 1 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000011", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[942.47, 257.0], [935.61, 259.3], [872.31, 254.81], [880.43, 252.74], [942.47, 145.17], [935.61, 144.41], [872.31, 145.89], [880.43, 146.57]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[933.21, 253.91], [925.7, 256.36], [874.97, 252.91], [883.6, 250.65], [933.21, 152.84], [925.7, 152.23], [874.97, 153.08], [883.6, 153.64]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[664.44, 205.69], [668.9, 206.18], [650.19, 206.48], [645.97, 205.99], [664.44, 168.45], [668.9, 168.39], [650.19, 168.35], [645.97, 168.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[262.49, 261.72], [236.36, 261.73], [245.82, 259.47], [271.28, 259.47], [262.49, 191.16], [236.36, 191.17], [245.82, 190.7], [271.28, 190.7]]}]\n```", - "options": null, - "id": 2 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000011", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[491.01, 226.21], [445.02, 226.28], [465.08, 219.11], [504.89, 219.06], [491.01, 171.97], [445.02, 171.97], [465.08, 172.09], [504.89, 172.09]]}, {\"category\": \"car\", \"corners_3d\": [[-748.08, 755.46], [-1220.52, 759.49], [-92.09, 396.22], [85.6, 395.64], [-748.08, 282.98], [-1220.52, 283.75], [-92.09, 215.08], [85.6, 214.97]]}]\n```", - "options": null, - "id": 3 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000011", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[942.47, 257.0], [935.61, 259.3], [872.31, 254.81], [880.43, 252.74], [942.47, 145.17], [935.61, 144.41], [872.31, 145.89], [880.43, 146.57]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[933.21, 253.91], [925.7, 256.36], [874.97, 252.91], [883.6, 250.65], [933.21, 152.84], [925.7, 152.23], [874.97, 153.08], [883.6, 153.64]]}, {\"category\": \"car\", \"corners_3d\": [[491.01, 226.21], [445.02, 226.28], [465.08, 219.11], [504.89, 219.06], [491.01, 171.97], [445.02, 171.97], [465.08, 172.09], [504.89, 172.09]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[664.44, 205.69], [668.9, 206.18], [650.19, 206.48], [645.97, 205.99], [664.44, 168.45], [668.9, 168.39], [650.19, 168.35], [645.97, 168.41]]}, {\"category\": \"car\", \"corners_3d\": [[-748.08, 755.46], [-1220.52, 759.49], [-92.09, 396.22], [85.6, 395.64], [-748.08, 282.98], [-1220.52, 283.75], [-92.09, 215.08], [85.6, 214.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[262.49, 261.72], [236.36, 261.73], [245.82, 259.47], [271.28, 259.47], [262.49, 191.16], [236.36, 191.17], [245.82, 190.7], [271.28, 190.7]]}]\n```", - "options": null, - "id": 4 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000012", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[474.36, 206.6], [448.5, 206.68], [457.84, 203.81], [481.48, 203.74], [474.36, 177.6], [448.5, 177.62], [457.84, 177.21], [481.48, 177.2]]}]\n```", - "options": null, - "id": 5 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000012", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[662.41, 203.41], [678.11, 203.28], [690.13, 205.12], [673.54, 205.26], [662.41, 185.74], [678.11, 185.69], [690.13, 186.46], [673.54, 186.52]]}]\n```", - "options": null, - "id": 6 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000012", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[662.41, 203.41], [678.11, 203.28], [690.13, 205.12], [673.54, 205.26], [662.41, 185.74], [678.11, 185.69], [690.13, 186.46], [673.54, 186.52]]}, {\"category\": \"van\", \"corners_3d\": [[474.36, 206.6], [448.5, 206.68], [457.84, 203.81], [481.48, 203.74], [474.36, 177.6], [448.5, 177.62], [457.84, 177.21], [481.48, 177.2]]}]\n```", - "options": null, - "id": 7 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000014", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[647.44, 194.43], [626.68, 194.31], [635.58, 192.82], [654.95, 192.93], [647.44, 176.36], [626.68, 176.34], [635.58, 176.1], [654.95, 176.12]]}]\n```", - "options": null, - "id": 8 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000027", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[155.03, 315.94], [16.09, 316.51], [228.03, 263.87], [315.66, 263.65], [155.03, 120.33], [16.09, 120.12], [228.03, 139.44], [315.66, 139.53]]}]\n```", - "options": null, - "id": 9 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000027", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[591.1, 193.57], [606.45, 193.56], [606.98, 194.59], [590.88, 194.6], [591.1, 181.48], [606.45, 181.48], [606.98, 181.91], [590.88, 181.91]]}]\n```", - "options": null, - "id": 10 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000027", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[155.03, 315.94], [16.09, 316.51], [228.03, 263.87], [315.66, 263.65], [155.03, 120.33], [16.09, 120.12], [228.03, 139.44], [315.66, 139.53]]}, {\"category\": \"car\", \"corners_3d\": [[591.1, 193.57], [606.45, 193.56], [606.98, 194.59], [590.88, 194.6], [591.1, 181.48], [606.45, 181.48], [606.98, 181.91], [590.88, 181.91]]}]\n```", - "options": null, - "id": 11 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000034", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[328.21, 271.78], [293.94, 284.48], [46.28, 285.24], [109.12, 272.38], [328.21, 196.85], [293.94, 199.93], [46.28, 200.12], [109.12, 197.0]]}, {\"category\": \"car\", \"corners_3d\": [[467.21, 206.94], [471.3, 205.89], [522.1, 205.87], [519.61, 206.91], [467.21, 185.41], [471.3, 185.02], [522.1, 185.01], [519.61, 185.4]]}]\n```", - "options": null, - "id": 12 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000038", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[685.11, 196.37], [707.59, 196.99], [663.68, 198.87], [640.68, 198.16], [685.11, 168.38], [707.59, 168.27], [663.68, 167.91], [640.68, 168.04]]}]\n```", - "options": null, - "id": 13 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000038", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[593.7, 214.05], [633.21, 214.03], [637.44, 219.77], [592.41, 219.8], [593.7, 176.68], [633.21, 176.68], [637.44, 177.21], [592.41, 177.21]]}, {\"category\": \"car\", \"corners_3d\": [[1064.34, 302.12], [1242.36, 323.53], [1131.03, 404.7], [899.2, 357.62], [1064.34, 165.27], [1242.36, 164.01], [1131.03, 159.25], [899.2, 162.01]]}, {\"category\": \"car\", \"corners_3d\": [[930.68, 266.27], [1037.95, 276.78], [937.34, 303.3], [818.53, 287.16], [930.68, 177.59], [1037.95, 178.13], [937.34, 179.47], [818.53, 178.65]]}, {\"category\": \"car\", \"corners_3d\": [[869.12, 241.59], [941.19, 247.3], [826.8, 263.52], [750.99, 255.19], [869.12, 177.81], [941.19, 178.23], [826.8, 179.4], [750.99, 178.79]]}, {\"category\": \"car\", \"corners_3d\": [[513.45, 201.16], [489.31, 201.15], [498.71, 199.05], [521.07, 199.06], [513.45, 180.96], [489.31, 180.96], [498.71, 180.36], [521.07, 180.36]]}, {\"category\": \"car\", \"corners_3d\": [[365.8, 251.19], [302.84, 251.19], [356.97, 237.34], [408.8, 237.34], [365.8, 192.1], [302.84, 192.11], [356.97, 188.7], [408.8, 188.7]]}, {\"category\": \"car\", \"corners_3d\": [[760.52, 214.22], [796.76, 216.2], [719.49, 221.2], [683.19, 218.76], [760.52, 173.93], [796.76, 173.98], [719.49, 174.11], [683.19, 174.04]]}, {\"category\": \"car\", \"corners_3d\": [[715.68, 205.16], [745.68, 206.28], [690.8, 209.62], [659.93, 208.26], [715.68, 175.0], [745.68, 175.08], [690.8, 175.3], [659.93, 175.21]]}, {\"category\": \"car\", \"corners_3d\": [[458.77, 200.1], [434.54, 200.08], [450.16, 197.83], [472.39, 197.84], [458.77, 175.89], [434.54, 175.89], [450.16, 175.64], [472.39, 175.64]]}, {\"category\": \"car\", \"corners_3d\": [[476.44, 198.14], [456.4, 198.12], [468.44, 196.37], [487.11, 196.39], [476.44, 177.32], [456.4, 177.32], [468.44, 177.01], [487.11, 177.01]]}, {\"category\": \"car\", \"corners_3d\": [[428.86, 195.92], [412.9, 195.92], [421.97, 194.81], [437.16, 194.81], [428.86, 177.78], [412.9, 177.78], [421.97, 177.55], [437.16, 177.55]]}]\n```", - "options": null, - "id": 14 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000038", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[593.7, 214.05], [633.21, 214.03], [637.44, 219.77], [592.41, 219.8], [593.7, 176.68], [633.21, 176.68], [637.44, 177.21], [592.41, 177.21]]}, {\"category\": \"car\", \"corners_3d\": [[1064.34, 302.12], [1242.36, 323.53], [1131.03, 404.7], [899.2, 357.62], [1064.34, 165.27], [1242.36, 164.01], [1131.03, 159.25], [899.2, 162.01]]}, {\"category\": \"car\", \"corners_3d\": [[930.68, 266.27], [1037.95, 276.78], [937.34, 303.3], [818.53, 287.16], [930.68, 177.59], [1037.95, 178.13], [937.34, 179.47], [818.53, 178.65]]}, {\"category\": \"car\", \"corners_3d\": [[869.12, 241.59], [941.19, 247.3], [826.8, 263.52], [750.99, 255.19], [869.12, 177.81], [941.19, 178.23], [826.8, 179.4], [750.99, 178.79]]}, {\"category\": \"car\", \"corners_3d\": [[513.45, 201.16], [489.31, 201.15], [498.71, 199.05], [521.07, 199.06], [513.45, 180.96], [489.31, 180.96], [498.71, 180.36], [521.07, 180.36]]}, {\"category\": \"car\", \"corners_3d\": [[365.8, 251.19], [302.84, 251.19], [356.97, 237.34], [408.8, 237.34], [365.8, 192.1], [302.84, 192.11], [356.97, 188.7], [408.8, 188.7]]}, {\"category\": \"car\", \"corners_3d\": [[760.52, 214.22], [796.76, 216.2], [719.49, 221.2], [683.19, 218.76], [760.52, 173.93], [796.76, 173.98], [719.49, 174.11], [683.19, 174.04]]}, {\"category\": \"car\", \"corners_3d\": [[715.68, 205.16], [745.68, 206.28], [690.8, 209.62], [659.93, 208.26], [715.68, 175.0], [745.68, 175.08], [690.8, 175.3], [659.93, 175.21]]}, {\"category\": \"van\", \"corners_3d\": [[685.11, 196.37], [707.59, 196.99], [663.68, 198.87], [640.68, 198.16], [685.11, 168.38], [707.59, 168.27], [663.68, 167.91], [640.68, 168.04]]}, {\"category\": \"car\", \"corners_3d\": [[458.77, 200.1], [434.54, 200.08], [450.16, 197.83], [472.39, 197.84], [458.77, 175.89], [434.54, 175.89], [450.16, 175.64], [472.39, 175.64]]}, {\"category\": \"car\", \"corners_3d\": [[476.44, 198.14], [456.4, 198.12], [468.44, 196.37], [487.11, 196.39], [476.44, 177.32], [456.4, 177.32], [468.44, 177.01], [487.11, 177.01]]}, {\"category\": \"car\", \"corners_3d\": [[428.86, 195.92], [412.9, 195.92], [421.97, 194.81], [437.16, 194.81], [428.86, 177.78], [412.9, 177.78], [421.97, 177.55], [437.16, 177.55]]}]\n```", - "options": null, - "id": 15 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000044", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[93.64, 542.05], [-248.35, 540.42], [200.72, 349.5], [365.65, 349.87], [93.64, 234.68], [-248.35, 234.41], [200.72, 202.43], [365.65, 202.5]]}, {\"category\": \"car\", \"corners_3d\": [[475.99, 277.24], [379.31, 277.26], [432.9, 252.9], [507.02, 252.89], [475.99, 187.15], [379.31, 187.15], [432.9, 183.82], [507.02, 183.82]]}]\n```", - "options": null, - "id": 16 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000063", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[405.77, 290.97], [440.86, 287.62], [455.89, 292.77], [419.63, 296.43], [405.77, 172.12], [440.86, 172.14], [455.89, 172.11], [419.63, 172.09]]}]\n```", - "options": null, - "id": 17 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000063", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[876.25, 236.77], [932.78, 237.81], [976.01, 255.22], [903.76, 253.56], [876.25, 184.81], [932.78, 185.0], [976.01, 188.26], [903.76, 187.95]]}, {\"category\": \"car\", \"corners_3d\": [[287.49, 277.34], [348.73, 265.8], [514.52, 273.58], [469.99, 287.28], [287.49, 174.13], [348.73, 173.99], [514.52, 174.08], [469.99, 174.25]]}]\n```", - "options": null, - "id": 18 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000063", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[842.93, 210.38], [879.93, 210.85], [890.12, 215.82], [848.2, 215.22], [842.93, 170.12], [879.93, 170.09], [890.12, 169.73], [848.2, 169.77]]}]\n```", - "options": null, - "id": 19 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000063", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[876.25, 236.77], [932.78, 237.81], [976.01, 255.22], [903.76, 253.56], [876.25, 184.81], [932.78, 185.0], [976.01, 188.26], [903.76, 187.95]]}, {\"category\": \"car\", \"corners_3d\": [[287.49, 277.34], [348.73, 265.8], [514.52, 273.58], [469.99, 287.28], [287.49, 174.13], [348.73, 173.99], [514.52, 174.08], [469.99, 174.25]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[405.77, 290.97], [440.86, 287.62], [455.89, 292.77], [419.63, 296.43], [405.77, 172.12], [440.86, 172.14], [455.89, 172.11], [419.63, 172.09]]}]\n```", - "options": null, - "id": 20 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000063", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[405.77, 290.97], [440.86, 287.62], [455.89, 292.77], [419.63, 296.43], [405.77, 172.12], [440.86, 172.14], [455.89, 172.11], [419.63, 172.09]]}, {\"category\": \"van\", \"corners_3d\": [[842.93, 210.38], [879.93, 210.85], [890.12, 215.82], [848.2, 215.22], [842.93, 170.12], [879.93, 170.09], [890.12, 169.73], [848.2, 169.77]]}]\n```", - "options": null, - "id": 21 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000063", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[876.25, 236.77], [932.78, 237.81], [976.01, 255.22], [903.76, 253.56], [876.25, 184.81], [932.78, 185.0], [976.01, 188.26], [903.76, 187.95]]}, {\"category\": \"car\", \"corners_3d\": [[287.49, 277.34], [348.73, 265.8], [514.52, 273.58], [469.99, 287.28], [287.49, 174.13], [348.73, 173.99], [514.52, 174.08], [469.99, 174.25]]}, {\"category\": \"van\", \"corners_3d\": [[842.93, 210.38], [879.93, 210.85], [890.12, 215.82], [848.2, 215.22], [842.93, 170.12], [879.93, 170.09], [890.12, 169.73], [848.2, 169.77]]}]\n```", - "options": null, - "id": 22 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000063", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[876.25, 236.77], [932.78, 237.81], [976.01, 255.22], [903.76, 253.56], [876.25, 184.81], [932.78, 185.0], [976.01, 188.26], [903.76, 187.95]]}, {\"category\": \"car\", \"corners_3d\": [[287.49, 277.34], [348.73, 265.8], [514.52, 273.58], [469.99, 287.28], [287.49, 174.13], [348.73, 173.99], [514.52, 174.08], [469.99, 174.25]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[405.77, 290.97], [440.86, 287.62], [455.89, 292.77], [419.63, 296.43], [405.77, 172.12], [440.86, 172.14], [455.89, 172.11], [419.63, 172.09]]}, {\"category\": \"van\", \"corners_3d\": [[842.93, 210.38], [879.93, 210.85], [890.12, 215.82], [848.2, 215.22], [842.93, 170.12], [879.93, 170.09], [890.12, 169.73], [848.2, 169.77]]}]\n```", - "options": null, - "id": 23 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000065", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (person_sitting). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[984.75, 365.92], [1020.37, 356.41], [1139.82, 379.62], [1106.6, 391.91], [984.75, 217.74], [1020.37, 215.87], [1139.82, 220.43], [1106.6, 222.85]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1001.62, 351.52], [922.36, 343.53], [929.8, 328.3], [1001.6, 334.82], [1001.62, 208.57], [922.36, 207.3], [929.8, 204.87], [1001.6, 205.91]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[764.39, 280.33], [809.07, 280.39], [823.2, 287.34], [775.37, 287.27], [764.39, 199.1], [809.07, 199.11], [823.2, 200.35], [775.37, 200.34]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[800.78, 281.3], [820.57, 277.75], [862.76, 282.26], [843.64, 286.15], [800.78, 198.28], [820.57, 197.69], [862.76, 198.45], [843.64, 199.1]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[796.32, 275.81], [760.48, 277.22], [738.77, 271.44], [772.78, 270.19], [796.32, 203.44], [760.48, 203.77], [738.77, 202.43], [772.78, 202.14]]}]\n```", - "options": null, - "id": 24 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000065", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[599.17, 220.26], [587.95, 219.07], [613.84, 218.39], [625.66, 219.54], [599.17, 172.95], [587.95, 173.21], [613.84, 173.36], [625.66, 173.11]]}]\n```", - "options": null, - "id": 25 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000065", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[442.74, 208.34], [451.83, 210.35], [302.52, 212.51], [303.96, 210.19], [442.74, 137.75], [451.83, 134.46], [302.52, 130.95], [303.96, 134.73]]}, {\"category\": \"tram\", \"corners_3d\": [[503.74, 204.56], [516.94, 206.24], [453.88, 207.04], [444.68, 205.26], [503.74, 136.63], [516.94, 133.37], [453.88, 131.81], [444.68, 135.27]]}, {\"category\": \"tram\", \"corners_3d\": [[713.07, 197.3], [738.16, 198.31], [519.08, 200.46], [505.77, 199.18], [713.07, 136.05], [738.16, 133.14], [519.08, 126.95], [505.77, 130.63]]}]\n```", - "options": null, - "id": 26 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000065", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (person_sitting, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[599.17, 220.26], [587.95, 219.07], [613.84, 218.39], [625.66, 219.54], [599.17, 172.95], [587.95, 173.21], [613.84, 173.36], [625.66, 173.11]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[984.75, 365.92], [1020.37, 356.41], [1139.82, 379.62], [1106.6, 391.91], [984.75, 217.74], [1020.37, 215.87], [1139.82, 220.43], [1106.6, 222.85]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1001.62, 351.52], [922.36, 343.53], [929.8, 328.3], [1001.6, 334.82], [1001.62, 208.57], [922.36, 207.3], [929.8, 204.87], [1001.6, 205.91]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[764.39, 280.33], [809.07, 280.39], [823.2, 287.34], [775.37, 287.27], [764.39, 199.1], [809.07, 199.11], [823.2, 200.35], [775.37, 200.34]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[800.78, 281.3], [820.57, 277.75], [862.76, 282.26], [843.64, 286.15], [800.78, 198.28], [820.57, 197.69], [862.76, 198.45], [843.64, 199.1]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[796.32, 275.81], [760.48, 277.22], [738.77, 271.44], [772.78, 270.19], [796.32, 203.44], [760.48, 203.77], [738.77, 202.43], [772.78, 202.14]]}]\n```", - "options": null, - "id": 27 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000065", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (person_sitting, tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[442.74, 208.34], [451.83, 210.35], [302.52, 212.51], [303.96, 210.19], [442.74, 137.75], [451.83, 134.46], [302.52, 130.95], [303.96, 134.73]]}, {\"category\": \"tram\", \"corners_3d\": [[503.74, 204.56], [516.94, 206.24], [453.88, 207.04], [444.68, 205.26], [503.74, 136.63], [516.94, 133.37], [453.88, 131.81], [444.68, 135.27]]}, {\"category\": \"tram\", \"corners_3d\": [[713.07, 197.3], [738.16, 198.31], [519.08, 200.46], [505.77, 199.18], [713.07, 136.05], [738.16, 133.14], [519.08, 126.95], [505.77, 130.63]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[984.75, 365.92], [1020.37, 356.41], [1139.82, 379.62], [1106.6, 391.91], [984.75, 217.74], [1020.37, 215.87], [1139.82, 220.43], [1106.6, 222.85]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1001.62, 351.52], [922.36, 343.53], [929.8, 328.3], [1001.6, 334.82], [1001.62, 208.57], [922.36, 207.3], [929.8, 204.87], [1001.6, 205.91]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[764.39, 280.33], [809.07, 280.39], [823.2, 287.34], [775.37, 287.27], [764.39, 199.1], [809.07, 199.11], [823.2, 200.35], [775.37, 200.34]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[800.78, 281.3], [820.57, 277.75], [862.76, 282.26], [843.64, 286.15], [800.78, 198.28], [820.57, 197.69], [862.76, 198.45], [843.64, 199.1]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[796.32, 275.81], [760.48, 277.22], [738.77, 271.44], [772.78, 270.19], [796.32, 203.44], [760.48, 203.77], [738.77, 202.43], [772.78, 202.14]]}]\n```", - "options": null, - "id": 28 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000065", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[599.17, 220.26], [587.95, 219.07], [613.84, 218.39], [625.66, 219.54], [599.17, 172.95], [587.95, 173.21], [613.84, 173.36], [625.66, 173.11]]}, {\"category\": \"tram\", \"corners_3d\": [[442.74, 208.34], [451.83, 210.35], [302.52, 212.51], [303.96, 210.19], [442.74, 137.75], [451.83, 134.46], [302.52, 130.95], [303.96, 134.73]]}, {\"category\": \"tram\", \"corners_3d\": [[503.74, 204.56], [516.94, 206.24], [453.88, 207.04], [444.68, 205.26], [503.74, 136.63], [516.94, 133.37], [453.88, 131.81], [444.68, 135.27]]}, {\"category\": \"tram\", \"corners_3d\": [[713.07, 197.3], [738.16, 198.31], [519.08, 200.46], [505.77, 199.18], [713.07, 136.05], [738.16, 133.14], [519.08, 126.95], [505.77, 130.63]]}]\n```", - "options": null, - "id": 29 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000065", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (person_sitting, pedestrian, tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[599.17, 220.26], [587.95, 219.07], [613.84, 218.39], [625.66, 219.54], [599.17, 172.95], [587.95, 173.21], [613.84, 173.36], [625.66, 173.11]]}, {\"category\": \"tram\", \"corners_3d\": [[442.74, 208.34], [451.83, 210.35], [302.52, 212.51], [303.96, 210.19], [442.74, 137.75], [451.83, 134.46], [302.52, 130.95], [303.96, 134.73]]}, {\"category\": \"tram\", \"corners_3d\": [[503.74, 204.56], [516.94, 206.24], [453.88, 207.04], [444.68, 205.26], [503.74, 136.63], [516.94, 133.37], [453.88, 131.81], [444.68, 135.27]]}, {\"category\": \"tram\", \"corners_3d\": [[713.07, 197.3], [738.16, 198.31], [519.08, 200.46], [505.77, 199.18], [713.07, 136.05], [738.16, 133.14], [519.08, 126.95], [505.77, 130.63]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[984.75, 365.92], [1020.37, 356.41], [1139.82, 379.62], [1106.6, 391.91], [984.75, 217.74], [1020.37, 215.87], [1139.82, 220.43], [1106.6, 222.85]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[1001.62, 351.52], [922.36, 343.53], [929.8, 328.3], [1001.6, 334.82], [1001.62, 208.57], [922.36, 207.3], [929.8, 204.87], [1001.6, 205.91]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[764.39, 280.33], [809.07, 280.39], [823.2, 287.34], [775.37, 287.27], [764.39, 199.1], [809.07, 199.11], [823.2, 200.35], [775.37, 200.34]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[800.78, 281.3], [820.57, 277.75], [862.76, 282.26], [843.64, 286.15], [800.78, 198.28], [820.57, 197.69], [862.76, 198.45], [843.64, 199.1]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[796.32, 275.81], [760.48, 277.22], [738.77, 271.44], [772.78, 270.19], [796.32, 203.44], [760.48, 203.77], [738.77, 202.43], [772.78, 202.14]]}]\n```", - "options": null, - "id": 30 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000067", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[681.25, 193.63], [702.68, 193.77], [695.39, 195.77], [671.99, 195.6], [681.25, 174.58], [702.68, 174.59], [695.39, 174.76], [671.99, 174.74]]}, {\"category\": \"car\", \"corners_3d\": [[716.02, 197.43], [744.16, 197.69], [736.01, 200.45], [704.89, 200.12], [716.02, 172.14], [744.16, 172.13], [736.01, 172.05], [704.89, 172.06]]}, {\"category\": \"car\", \"corners_3d\": [[523.65, 253.61], [600.24, 256.47], [524.62, 282.7], [428.45, 277.81], [523.65, 179.95], [600.24, 180.2], [524.62, 182.5], [428.45, 182.07]]}, {\"category\": \"car\", \"corners_3d\": [[312.05, 262.01], [384.65, 265.6], [168.47, 308.02], [76.48, 300.53], [312.05, 181.38], [384.65, 181.72], [168.47, 185.78], [76.48, 185.06]]}, {\"category\": \"car\", \"corners_3d\": [[501.78, 207.78], [479.09, 207.65], [494.52, 205.18], [515.64, 205.29], [501.78, 182.88], [479.09, 182.84], [494.52, 182.13], [515.64, 182.16]]}]\n```", - "options": null, - "id": 31 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000067", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[80.97, 274.75], [-19.7, 270.96], [389.89, 220.1], [444.87, 220.96], [80.97, 76.83], [-19.7, 80.4], [389.89, 128.33], [444.87, 127.52]]}]\n```", - "options": null, - "id": 32 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000067", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[651.28, 189.31], [630.43, 189.48], [591.76, 187.08], [609.96, 186.96], [651.28, 159.34], [630.43, 159.2], [591.76, 161.17], [609.96, 161.27]]}]\n```", - "options": null, - "id": 33 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000067", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[681.25, 193.63], [702.68, 193.77], [695.39, 195.77], [671.99, 195.6], [681.25, 174.58], [702.68, 174.59], [695.39, 174.76], [671.99, 174.74]]}, {\"category\": \"truck\", \"corners_3d\": [[80.97, 274.75], [-19.7, 270.96], [389.89, 220.1], [444.87, 220.96], [80.97, 76.83], [-19.7, 80.4], [389.89, 128.33], [444.87, 127.52]]}, {\"category\": \"car\", \"corners_3d\": [[716.02, 197.43], [744.16, 197.69], [736.01, 200.45], [704.89, 200.12], [716.02, 172.14], [744.16, 172.13], [736.01, 172.05], [704.89, 172.06]]}, {\"category\": \"car\", \"corners_3d\": [[523.65, 253.61], [600.24, 256.47], [524.62, 282.7], [428.45, 277.81], [523.65, 179.95], [600.24, 180.2], [524.62, 182.5], [428.45, 182.07]]}, {\"category\": \"car\", \"corners_3d\": [[312.05, 262.01], [384.65, 265.6], [168.47, 308.02], [76.48, 300.53], [312.05, 181.38], [384.65, 181.72], [168.47, 185.78], [76.48, 185.06]]}, {\"category\": \"car\", \"corners_3d\": [[501.78, 207.78], [479.09, 207.65], [494.52, 205.18], [515.64, 205.29], [501.78, 182.88], [479.09, 182.84], [494.52, 182.13], [515.64, 182.16]]}]\n```", - "options": null, - "id": 34 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000067", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[681.25, 193.63], [702.68, 193.77], [695.39, 195.77], [671.99, 195.6], [681.25, 174.58], [702.68, 174.59], [695.39, 174.76], [671.99, 174.74]]}, {\"category\": \"car\", \"corners_3d\": [[716.02, 197.43], [744.16, 197.69], [736.01, 200.45], [704.89, 200.12], [716.02, 172.14], [744.16, 172.13], [736.01, 172.05], [704.89, 172.06]]}, {\"category\": \"car\", \"corners_3d\": [[523.65, 253.61], [600.24, 256.47], [524.62, 282.7], [428.45, 277.81], [523.65, 179.95], [600.24, 180.2], [524.62, 182.5], [428.45, 182.07]]}, {\"category\": \"car\", \"corners_3d\": [[312.05, 262.01], [384.65, 265.6], [168.47, 308.02], [76.48, 300.53], [312.05, 181.38], [384.65, 181.72], [168.47, 185.78], [76.48, 185.06]]}, {\"category\": \"car\", \"corners_3d\": [[501.78, 207.78], [479.09, 207.65], [494.52, 205.18], [515.64, 205.29], [501.78, 182.88], [479.09, 182.84], [494.52, 182.13], [515.64, 182.16]]}, {\"category\": \"tram\", \"corners_3d\": [[651.28, 189.31], [630.43, 189.48], [591.76, 187.08], [609.96, 186.96], [651.28, 159.34], [630.43, 159.2], [591.76, 161.17], [609.96, 161.27]]}]\n```", - "options": null, - "id": 35 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000067", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck, tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[80.97, 274.75], [-19.7, 270.96], [389.89, 220.1], [444.87, 220.96], [80.97, 76.83], [-19.7, 80.4], [389.89, 128.33], [444.87, 127.52]]}, {\"category\": \"tram\", \"corners_3d\": [[651.28, 189.31], [630.43, 189.48], [591.76, 187.08], [609.96, 186.96], [651.28, 159.34], [630.43, 159.2], [591.76, 161.17], [609.96, 161.27]]}]\n```", - "options": null, - "id": 36 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000067", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, truck, tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[681.25, 193.63], [702.68, 193.77], [695.39, 195.77], [671.99, 195.6], [681.25, 174.58], [702.68, 174.59], [695.39, 174.76], [671.99, 174.74]]}, {\"category\": \"truck\", \"corners_3d\": [[80.97, 274.75], [-19.7, 270.96], [389.89, 220.1], [444.87, 220.96], [80.97, 76.83], [-19.7, 80.4], [389.89, 128.33], [444.87, 127.52]]}, {\"category\": \"car\", \"corners_3d\": [[716.02, 197.43], [744.16, 197.69], [736.01, 200.45], [704.89, 200.12], [716.02, 172.14], [744.16, 172.13], [736.01, 172.05], [704.89, 172.06]]}, {\"category\": \"car\", \"corners_3d\": [[523.65, 253.61], [600.24, 256.47], [524.62, 282.7], [428.45, 277.81], [523.65, 179.95], [600.24, 180.2], [524.62, 182.5], [428.45, 182.07]]}, {\"category\": \"car\", \"corners_3d\": [[312.05, 262.01], [384.65, 265.6], [168.47, 308.02], [76.48, 300.53], [312.05, 181.38], [384.65, 181.72], [168.47, 185.78], [76.48, 185.06]]}, {\"category\": \"car\", \"corners_3d\": [[501.78, 207.78], [479.09, 207.65], [494.52, 205.18], [515.64, 205.29], [501.78, 182.88], [479.09, 182.84], [494.52, 182.13], [515.64, 182.16]]}, {\"category\": \"tram\", \"corners_3d\": [[651.28, 189.31], [630.43, 189.48], [591.76, 187.08], [609.96, 186.96], [651.28, 159.34], [630.43, 159.2], [591.76, 161.17], [609.96, 161.27]]}]\n```", - "options": null, - "id": 37 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000072", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-2.02, 229.44], [-38.81, 229.33], [87.85, 218.76], [117.95, 218.83], [-2.02, 179.52], [-38.81, 179.51], [87.85, 178.26], [117.95, 178.27]]}, {\"category\": \"car\", \"corners_3d\": [[407.45, 288.94], [290.79, 289.14], [382.13, 254.99], [464.46, 254.89], [407.45, 192.07], [290.79, 192.1], [382.13, 186.45], [464.46, 186.43]]}, {\"category\": \"car\", \"corners_3d\": [[593.76, 208.25], [627.5, 208.25], [630.08, 213.5], [591.33, 213.49], [593.76, 174.71], [627.5, 174.71], [630.08, 174.99], [591.33, 174.98]]}]\n```", - "options": null, - "id": 38 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000073", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[223.01, 364.63], [267.59, 345.75], [322.55, 347.02], [283.66, 366.19], [223.01, 175.21], [267.59, 174.98], [322.55, 174.99], [283.66, 175.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[265.18, 341.57], [211.03, 357.45], [165.47, 352.91], [222.22, 337.78], [265.18, 160.0], [211.03, 158.79], [165.47, 159.13], [222.22, 160.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[795.47, 283.69], [777.47, 289.19], [744.37, 285.74], [763.38, 280.55], [795.47, 164.67], [777.47, 164.26], [744.37, 164.52], [763.38, 164.9]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[542.03, 214.01], [554.93, 213.56], [563.51, 214.4], [550.44, 214.87], [542.03, 177.18], [554.93, 177.13], [563.51, 177.22], [550.44, 177.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[530.64, 213.55], [542.43, 213.16], [551.91, 214.18], [539.91, 214.59], [530.64, 174.54], [542.43, 174.53], [551.91, 174.57], [539.91, 174.58]]}]\n```", - "options": null, - "id": 39 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000073", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[469.93, 242.34], [445.22, 242.32], [463.59, 234.86], [485.65, 234.87], [469.93, 171.1], [445.22, 171.1], [463.59, 171.29], [485.65, 171.29]]}, {\"category\": \"cyclist\", \"corners_3d\": [[513.25, 224.01], [494.61, 224.09], [499.81, 220.34], [517.08, 220.27], [513.25, 168.53], [494.61, 168.52], [499.81, 168.84], [517.08, 168.84]]}]\n```", - "options": null, - "id": 40 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000073", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[223.01, 364.63], [267.59, 345.75], [322.55, 347.02], [283.66, 366.19], [223.01, 175.21], [267.59, 174.98], [322.55, 174.99], [283.66, 175.23]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[265.18, 341.57], [211.03, 357.45], [165.47, 352.91], [222.22, 337.78], [265.18, 160.0], [211.03, 158.79], [165.47, 159.13], [222.22, 160.29]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[795.47, 283.69], [777.47, 289.19], [744.37, 285.74], [763.38, 280.55], [795.47, 164.67], [777.47, 164.26], [744.37, 164.52], [763.38, 164.9]]}, {\"category\": \"cyclist\", \"corners_3d\": [[469.93, 242.34], [445.22, 242.32], [463.59, 234.86], [485.65, 234.87], [469.93, 171.1], [445.22, 171.1], [463.59, 171.29], [485.65, 171.29]]}, {\"category\": \"cyclist\", \"corners_3d\": [[513.25, 224.01], [494.61, 224.09], [499.81, 220.34], [517.08, 220.27], [513.25, 168.53], [494.61, 168.52], [499.81, 168.84], [517.08, 168.84]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[542.03, 214.01], [554.93, 213.56], [563.51, 214.4], [550.44, 214.87], [542.03, 177.18], [554.93, 177.13], [563.51, 177.22], [550.44, 177.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[530.64, 213.55], [542.43, 213.16], [551.91, 214.18], [539.91, 214.59], [530.64, 174.54], [542.43, 174.53], [551.91, 174.57], [539.91, 174.58]]}]\n```", - "options": null, - "id": 41 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000077", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[606.69, 208.51], [574.69, 208.27], [587.53, 205.1], [616.72, 205.3], [606.69, 179.75], [574.69, 179.7], [587.53, 179.09], [616.72, 179.13]]}]\n```", - "options": null, - "id": 42 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000092", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[664.9, 199.53], [688.66, 199.76], [679.77, 201.86], [654.25, 201.6], [664.9, 176.95], [688.66, 176.99], [679.77, 177.31], [654.25, 177.27]]}, {\"category\": \"car\", \"corners_3d\": [[409.65, 241.62], [360.41, 240.78], [413.5, 231.1], [456.21, 231.72], [409.65, 189.54], [360.41, 189.33], [413.5, 186.98], [456.21, 187.13]]}, {\"category\": \"car\", \"corners_3d\": [[462.55, 229.58], [420.75, 229.07], [459.69, 221.7], [496.28, 222.09], [462.55, 176.17], [420.75, 176.14], [459.69, 175.71], [496.28, 175.73]]}, {\"category\": \"car\", \"corners_3d\": [[687.4, 207.46], [723.32, 207.54], [733.22, 212.0], [692.67, 211.89], [687.4, 173.7], [723.32, 173.7], [733.22, 173.81], [692.67, 173.81]]}, {\"category\": \"car\", \"corners_3d\": [[696.25, 203.77], [724.72, 203.96], [724.51, 207.37], [692.95, 207.13], [696.25, 165.35], [724.72, 165.3], [724.51, 164.48], [692.95, 164.53]]}, {\"category\": \"car\", \"corners_3d\": [[1217.73, 196.98], [1182.91, 195.38], [1304.91, 195.78], [1349.16, 197.43], [1217.73, 149.73], [1182.91, 151.25], [1304.91, 150.88], [1349.16, 149.3]]}]\n```", - "options": null, - "id": 43 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000094", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[660.86, 237.17], [714.01, 236.99], [748.0, 253.41], [681.31, 253.7], [660.86, 189.43], [714.01, 189.38], [748.0, 193.62], [681.31, 193.69]]}]\n```", - "options": null, - "id": 44 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000105", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[978.98, 361.2], [1037.51, 360.27], [1227.72, 436.0], [1146.71, 437.85], [978.98, 120.55], [1037.51, 120.81], [1227.72, 99.77], [1146.71, 99.26]]}]\n```", - "options": null, - "id": 45 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000105", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[553.2, 195.42], [532.27, 195.46], [534.75, 194.08], [554.4, 194.05], [553.2, 178.03], [532.27, 178.04], [534.75, 177.72], [554.4, 177.72]]}]\n```", - "options": null, - "id": 46 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000105", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[978.98, 361.2], [1037.51, 360.27], [1227.72, 436.0], [1146.71, 437.85], [978.98, 120.55], [1037.51, 120.81], [1227.72, 99.77], [1146.71, 99.26]]}, {\"category\": \"car\", \"corners_3d\": [[553.2, 195.42], [532.27, 195.46], [534.75, 194.08], [554.4, 194.05], [553.2, 178.03], [532.27, 178.04], [534.75, 177.72], [554.4, 177.72]]}]\n```", - "options": null, - "id": 47 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000120", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[248.63, 386.72], [56.82, 385.03], [275.08, 304.33], [394.64, 304.97], [248.63, 219.82], [56.82, 219.45], [275.08, 201.73], [394.64, 201.87]]}, {\"category\": \"car\", \"corners_3d\": [[-12593.07, 7576.98], [-19868.23, 7635.57], [-145.16, 447.7], [119.07, 447.62], [-12593.07, 1128.27], [-19868.23, 1135.83], [-145.16, 208.32], [119.07, 208.31]]}, {\"category\": \"car\", \"corners_3d\": [[441.49, 279.19], [347.66, 278.64], [412.73, 254.92], [485.69, 255.25], [441.49, 189.2], [347.66, 189.11], [412.73, 185.47], [485.69, 185.52]]}, {\"category\": \"car\", \"corners_3d\": [[511.73, 244.22], [455.62, 244.0], [485.28, 232.42], [532.32, 232.57], [511.73, 189.13], [455.62, 189.08], [485.28, 186.44], [532.32, 186.48]]}, {\"category\": \"car\", \"corners_3d\": [[538.56, 232.46], [487.28, 232.46], [505.79, 223.4], [549.27, 223.39], [538.56, 188.29], [487.28, 188.29], [505.79, 185.94], [549.27, 185.94]]}, {\"category\": \"car\", \"corners_3d\": [[675.58, 220.5], [721.89, 221.05], [718.75, 228.79], [665.13, 228.05], [675.58, 178.2], [721.89, 178.26], [718.75, 179.13], [665.13, 179.05]]}, {\"category\": \"car\", \"corners_3d\": [[958.64, 225.59], [985.34, 230.55], [796.93, 229.25], [786.64, 224.5], [958.64, 166.49], [985.34, 165.9], [796.93, 166.05], [786.64, 166.63]]}, {\"category\": \"car\", \"corners_3d\": [[709.28, 215.1], [744.99, 216.12], [719.19, 220.75], [680.43, 219.51], [709.28, 180.04], [744.99, 180.22], [719.19, 181.0], [680.43, 180.79]]}, {\"category\": \"car\", \"corners_3d\": [[753.72, 208.8], [783.5, 209.97], [747.28, 212.66], [716.63, 211.32], [753.72, 175.34], [783.5, 175.42], [747.28, 175.61], [716.63, 175.51]]}]\n```", - "options": null, - "id": 48 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000133", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[535.32, 202.09], [530.74, 203.05], [468.46, 202.81], [474.97, 201.86], [535.32, 183.42], [530.74, 183.77], [468.46, 183.68], [474.97, 183.34]]}]\n```", - "options": null, - "id": 49 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000149", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[762.72, 228.53], [788.35, 228.54], [794.81, 230.56], [768.26, 230.55], [762.72, 160.96], [788.35, 160.96], [794.81, 160.53], [768.26, 160.53]]}]\n```", - "options": null, - "id": 50 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000149", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[594.81, 197.53], [617.03, 197.52], [618.08, 199.36], [594.2, 199.37], [594.81, 176.53], [617.03, 176.53], [618.08, 176.8], [594.2, 176.8]]}, {\"category\": \"car\", \"corners_3d\": [[551.3, 218.9], [509.46, 218.9], [520.98, 213.57], [557.98, 213.57], [551.3, 180.74], [509.46, 180.74], [520.98, 179.83], [557.98, 179.83]]}, {\"category\": \"car\", \"corners_3d\": [[747.13, 271.98], [834.91, 271.74], [941.04, 315.62], [814.54, 316.11], [747.13, 183.05], [834.91, 183.02], [941.04, 187.54], [814.54, 187.59]]}, {\"category\": \"car\", \"corners_3d\": [[693.81, 247.41], [764.44, 247.19], [817.88, 269.77], [725.92, 270.14], [693.81, 176.31], [764.44, 176.3], [817.88, 177.34], [725.92, 177.36]]}, {\"category\": \"car\", \"corners_3d\": [[680.38, 235.24], [729.99, 234.9], [757.03, 244.34], [700.0, 244.8], [680.38, 173.93], [729.99, 173.92], [757.03, 174.09], [700.0, 174.09]]}, {\"category\": \"car\", \"corners_3d\": [[653.42, 214.04], [691.3, 213.96], [707.09, 219.85], [663.79, 219.96], [653.42, 178.11], [691.3, 178.1], [707.09, 178.86], [663.79, 178.87]]}, {\"category\": \"car\", \"corners_3d\": [[637.2, 200.07], [662.07, 200.08], [666.45, 202.75], [639.15, 202.73], [637.2, 176.49], [662.07, 176.49], [666.45, 176.85], [639.15, 176.84]]}, {\"category\": \"car\", \"corners_3d\": [[647.58, 208.67], [680.1, 208.65], [689.28, 212.91], [652.9, 212.93], [647.58, 176.53], [680.1, 176.53], [689.28, 176.97], [652.9, 176.97]]}, {\"category\": \"car\", \"corners_3d\": [[456.1, 205.52], [452.61, 206.55], [391.82, 206.72], [397.2, 205.67], [456.1, 183.15], [452.61, 183.48], [391.82, 183.53], [397.2, 183.2]]}, {\"category\": \"car\", \"corners_3d\": [[377.1, 206.54], [369.19, 207.65], [304.74, 207.62], [314.7, 206.51], [377.1, 183.19], [369.19, 183.53], [304.74, 183.52], [314.7, 183.18]]}]\n```", - "options": null, - "id": 51 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000149", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[594.81, 197.53], [617.03, 197.52], [618.08, 199.36], [594.2, 199.37], [594.81, 176.53], [617.03, 176.53], [618.08, 176.8], [594.2, 176.8]]}, {\"category\": \"car\", \"corners_3d\": [[551.3, 218.9], [509.46, 218.9], [520.98, 213.57], [557.98, 213.57], [551.3, 180.74], [509.46, 180.74], [520.98, 179.83], [557.98, 179.83]]}, {\"category\": \"car\", \"corners_3d\": [[747.13, 271.98], [834.91, 271.74], [941.04, 315.62], [814.54, 316.11], [747.13, 183.05], [834.91, 183.02], [941.04, 187.54], [814.54, 187.59]]}, {\"category\": \"car\", \"corners_3d\": [[693.81, 247.41], [764.44, 247.19], [817.88, 269.77], [725.92, 270.14], [693.81, 176.31], [764.44, 176.3], [817.88, 177.34], [725.92, 177.36]]}, {\"category\": \"car\", \"corners_3d\": [[680.38, 235.24], [729.99, 234.9], [757.03, 244.34], [700.0, 244.8], [680.38, 173.93], [729.99, 173.92], [757.03, 174.09], [700.0, 174.09]]}, {\"category\": \"car\", \"corners_3d\": [[653.42, 214.04], [691.3, 213.96], [707.09, 219.85], [663.79, 219.96], [653.42, 178.11], [691.3, 178.1], [707.09, 178.86], [663.79, 178.87]]}, {\"category\": \"car\", \"corners_3d\": [[637.2, 200.07], [662.07, 200.08], [666.45, 202.75], [639.15, 202.73], [637.2, 176.49], [662.07, 176.49], [666.45, 176.85], [639.15, 176.84]]}, {\"category\": \"car\", \"corners_3d\": [[647.58, 208.67], [680.1, 208.65], [689.28, 212.91], [652.9, 212.93], [647.58, 176.53], [680.1, 176.53], [689.28, 176.97], [652.9, 176.97]]}, {\"category\": \"car\", \"corners_3d\": [[456.1, 205.52], [452.61, 206.55], [391.82, 206.72], [397.2, 205.67], [456.1, 183.15], [452.61, 183.48], [391.82, 183.53], [397.2, 183.2]]}, {\"category\": \"car\", \"corners_3d\": [[377.1, 206.54], [369.19, 207.65], [304.74, 207.62], [314.7, 206.51], [377.1, 183.19], [369.19, 183.53], [304.74, 183.52], [314.7, 183.18]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[762.72, 228.53], [788.35, 228.54], [794.81, 230.56], [768.26, 230.55], [762.72, 160.96], [788.35, 160.96], [794.81, 160.53], [768.26, 160.53]]}]\n```", - "options": null, - "id": 52 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000158", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1047.93, 215.13], [1057.06, 213.49], [1179.78, 217.58], [1175.15, 219.57], [1047.93, 177.7], [1057.06, 177.51], [1179.78, 177.98], [1175.15, 178.21]]}]\n```", - "options": null, - "id": 53 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000164", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[78.55, 199.17], [103.18, 199.58], [36.25, 201.96], [10.59, 201.47], [78.55, 165.63], [103.18, 165.52], [36.25, 164.87], [10.59, 165.0]]}, {\"category\": \"car\", \"corners_3d\": [[561.68, 222.84], [611.35, 222.74], [616.7, 234.42], [555.39, 234.57], [561.68, 173.45], [611.35, 173.45], [616.7, 173.59], [555.39, 173.59]]}, {\"category\": \"car\", \"corners_3d\": [[181.08, 195.35], [154.19, 195.37], [180.67, 194.01], [205.92, 194.0], [181.08, 167.83], [154.19, 167.83], [180.67, 168.13], [205.92, 168.13]]}]\n```", - "options": null, - "id": 54 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000164", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[764.58, 205.55], [770.24, 205.54], [776.59, 206.78], [770.71, 206.79], [764.58, 178.19], [770.24, 178.19], [776.59, 178.39], [770.71, 178.39]]}]\n```", - "options": null, - "id": 55 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000164", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[78.55, 199.17], [103.18, 199.58], [36.25, 201.96], [10.59, 201.47], [78.55, 165.63], [103.18, 165.52], [36.25, 164.87], [10.59, 165.0]]}, {\"category\": \"car\", \"corners_3d\": [[561.68, 222.84], [611.35, 222.74], [616.7, 234.42], [555.39, 234.57], [561.68, 173.45], [611.35, 173.45], [616.7, 173.59], [555.39, 173.59]]}, {\"category\": \"car\", \"corners_3d\": [[181.08, 195.35], [154.19, 195.37], [180.67, 194.01], [205.92, 194.0], [181.08, 167.83], [154.19, 167.83], [180.67, 168.13], [205.92, 168.13]]}, {\"category\": \"cyclist\", \"corners_3d\": [[764.58, 205.55], [770.24, 205.54], [776.59, 206.78], [770.71, 206.79], [764.58, 178.19], [770.24, 178.19], [776.59, 178.39], [770.71, 178.39]]}]\n```", - "options": null, - "id": 56 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000169", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[786.33, 317.85], [913.04, 317.36], [1125.1, 413.92], [914.44, 415.29], [786.33, 178.0], [913.04, 177.99], [1125.1, 181.41], [914.44, 181.46]]}, {\"category\": \"car\", \"corners_3d\": [[138.02, 376.37], [-59.99, 453.55], [-638.47, 436.67], [-297.23, 367.35], [138.02, 195.04], [-59.99, 203.45], [-638.47, 201.61], [-297.23, 194.06]]}, {\"category\": \"car\", \"corners_3d\": [[273.1, 310.33], [197.32, 341.3], [-176.52, 341.3], [-32.03, 310.33], [273.1, 180.96], [197.32, 182.79], [-176.52, 182.79], [-32.03, 180.96]]}, {\"category\": \"car\", \"corners_3d\": [[-16.58, 307.15], [75.91, 287.28], [290.25, 287.22], [235.02, 307.08], [-16.58, 160.82], [75.91, 162.6], [290.25, 162.61], [235.02, 160.83]]}, {\"category\": \"car\", \"corners_3d\": [[112.22, 271.85], [166.35, 260.73], [342.97, 260.27], [311.38, 271.27], [112.22, 188.16], [166.35, 186.44], [342.97, 186.37], [311.38, 188.07]]}, {\"category\": \"car\", \"corners_3d\": [[435.48, 249.57], [417.55, 257.16], [247.85, 256.96], [280.99, 249.41], [435.48, 187.23], [417.55, 188.65], [247.85, 188.61], [280.99, 187.19]]}, {\"category\": \"car\", \"corners_3d\": [[468.51, 240.79], [454.04, 247.1], [303.47, 246.79], [330.65, 240.52], [468.51, 182.05], [454.04, 182.91], [303.47, 182.87], [330.65, 182.02]]}, {\"category\": \"car\", \"corners_3d\": [[462.59, 236.79], [447.41, 241.89], [335.99, 241.24], [359.2, 236.23], [462.59, 180.32], [447.41, 180.92], [335.99, 180.84], [359.2, 180.26]]}, {\"category\": \"car\", \"corners_3d\": [[490.75, 226.02], [479.23, 230.17], [338.51, 229.71], [360.06, 225.63], [490.75, 175.29], [479.23, 175.48], [338.51, 175.46], [360.06, 175.27]]}, {\"category\": \"car\", \"corners_3d\": [[394.86, 219.12], [405.19, 216.43], [500.32, 216.14], [495.9, 218.79], [394.86, 178.49], [405.19, 178.17], [500.32, 178.13], [495.9, 178.45]]}, {\"category\": \"car\", \"corners_3d\": [[526.63, 209.94], [523.26, 211.77], [436.52, 211.87], [443.98, 210.03], [526.63, 170.78], [523.26, 170.67], [436.52, 170.67], [443.98, 170.77]]}, {\"category\": \"car\", \"corners_3d\": [[678.25, 228.7], [727.0, 228.74], [746.83, 238.84], [689.25, 238.78], [678.25, 183.62], [727.0, 183.63], [746.83, 185.58], [689.25, 185.57]]}, {\"category\": \"car\", \"corners_3d\": [[841.07, 218.85], [853.29, 221.83], [736.69, 221.37], [731.63, 218.44], [841.07, 172.84], [853.29, 172.84], [736.69, 172.84], [731.63, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[750.82, 203.23], [757.38, 204.51], [688.82, 204.57], [685.02, 203.29], [750.82, 172.68], [757.38, 172.67], [688.82, 172.67], [685.02, 172.67]]}, {\"category\": \"car\", \"corners_3d\": [[633.39, 197.06], [653.31, 197.03], [658.92, 198.92], [637.46, 198.95], [633.39, 179.0], [653.31, 178.99], [658.92, 179.47], [637.46, 179.48]]}]\n```", - "options": null, - "id": 57 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000169", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[581.72, 198.56], [574.11, 198.56], [574.72, 197.8], [582.1, 197.79], [581.72, 177.7], [574.11, 177.7], [574.72, 177.56], [582.1, 177.56]]}]\n```", - "options": null, - "id": 58 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000169", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[786.33, 317.85], [913.04, 317.36], [1125.1, 413.92], [914.44, 415.29], [786.33, 178.0], [913.04, 177.99], [1125.1, 181.41], [914.44, 181.46]]}, {\"category\": \"car\", \"corners_3d\": [[138.02, 376.37], [-59.99, 453.55], [-638.47, 436.67], [-297.23, 367.35], [138.02, 195.04], [-59.99, 203.45], [-638.47, 201.61], [-297.23, 194.06]]}, {\"category\": \"car\", \"corners_3d\": [[273.1, 310.33], [197.32, 341.3], [-176.52, 341.3], [-32.03, 310.33], [273.1, 180.96], [197.32, 182.79], [-176.52, 182.79], [-32.03, 180.96]]}, {\"category\": \"car\", \"corners_3d\": [[-16.58, 307.15], [75.91, 287.28], [290.25, 287.22], [235.02, 307.08], [-16.58, 160.82], [75.91, 162.6], [290.25, 162.61], [235.02, 160.83]]}, {\"category\": \"car\", \"corners_3d\": [[112.22, 271.85], [166.35, 260.73], [342.97, 260.27], [311.38, 271.27], [112.22, 188.16], [166.35, 186.44], [342.97, 186.37], [311.38, 188.07]]}, {\"category\": \"car\", \"corners_3d\": [[435.48, 249.57], [417.55, 257.16], [247.85, 256.96], [280.99, 249.41], [435.48, 187.23], [417.55, 188.65], [247.85, 188.61], [280.99, 187.19]]}, {\"category\": \"car\", \"corners_3d\": [[468.51, 240.79], [454.04, 247.1], [303.47, 246.79], [330.65, 240.52], [468.51, 182.05], [454.04, 182.91], [303.47, 182.87], [330.65, 182.02]]}, {\"category\": \"car\", \"corners_3d\": [[462.59, 236.79], [447.41, 241.89], [335.99, 241.24], [359.2, 236.23], [462.59, 180.32], [447.41, 180.92], [335.99, 180.84], [359.2, 180.26]]}, {\"category\": \"car\", \"corners_3d\": [[490.75, 226.02], [479.23, 230.17], [338.51, 229.71], [360.06, 225.63], [490.75, 175.29], [479.23, 175.48], [338.51, 175.46], [360.06, 175.27]]}, {\"category\": \"car\", \"corners_3d\": [[394.86, 219.12], [405.19, 216.43], [500.32, 216.14], [495.9, 218.79], [394.86, 178.49], [405.19, 178.17], [500.32, 178.13], [495.9, 178.45]]}, {\"category\": \"car\", \"corners_3d\": [[526.63, 209.94], [523.26, 211.77], [436.52, 211.87], [443.98, 210.03], [526.63, 170.78], [523.26, 170.67], [436.52, 170.67], [443.98, 170.77]]}, {\"category\": \"cyclist\", \"corners_3d\": [[581.72, 198.56], [574.11, 198.56], [574.72, 197.8], [582.1, 197.79], [581.72, 177.7], [574.11, 177.7], [574.72, 177.56], [582.1, 177.56]]}, {\"category\": \"car\", \"corners_3d\": [[678.25, 228.7], [727.0, 228.74], [746.83, 238.84], [689.25, 238.78], [678.25, 183.62], [727.0, 183.63], [746.83, 185.58], [689.25, 185.57]]}, {\"category\": \"car\", \"corners_3d\": [[841.07, 218.85], [853.29, 221.83], [736.69, 221.37], [731.63, 218.44], [841.07, 172.84], [853.29, 172.84], [736.69, 172.84], [731.63, 172.84]]}, {\"category\": \"car\", \"corners_3d\": [[750.82, 203.23], [757.38, 204.51], [688.82, 204.57], [685.02, 203.29], [750.82, 172.68], [757.38, 172.67], [688.82, 172.67], [685.02, 172.67]]}, {\"category\": \"car\", \"corners_3d\": [[633.39, 197.06], [653.31, 197.03], [658.92, 198.92], [637.46, 198.95], [633.39, 179.0], [653.31, 178.99], [658.92, 179.47], [637.46, 179.48]]}]\n```", - "options": null, - "id": 59 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000171", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[578.15, 240.22], [638.83, 240.22], [644.76, 254.15], [571.54, 254.14], [578.15, 182.52], [638.83, 182.52], [644.76, 184.52], [571.54, 184.52]]}, {\"category\": \"car\", \"corners_3d\": [[129.47, 271.18], [58.27, 271.09], [176.71, 250.23], [232.86, 250.29], [129.47, 213.81], [58.27, 213.78], [176.71, 205.09], [232.86, 205.11]]}, {\"category\": \"car\", \"corners_3d\": [[48.65, 260.46], [-8.1, 260.54], [84.61, 247.21], [132.68, 247.15], [48.65, 215.1], [-8.1, 215.13], [84.61, 208.71], [132.68, 208.68]]}, {\"category\": \"car\", \"corners_3d\": [[342.79, 227.06], [307.75, 227.22], [329.4, 222.65], [361.44, 222.52], [342.79, 197.09], [307.75, 197.16], [329.4, 195.12], [361.44, 195.06]]}]\n```", - "options": null, - "id": 60 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000182", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[433.25, 283.01], [330.08, 283.18], [400.21, 254.7], [476.7, 254.61], [433.25, 170.97], [330.08, 170.97], [400.21, 171.46], [476.7, 171.46]]}]\n```", - "options": null, - "id": 61 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000182", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[263.76, 389.23], [78.57, 391.45], [274.21, 306.18], [386.39, 305.35], [263.76, 220.23], [78.57, 220.71], [274.21, 202.05], [386.39, 201.86]]}, {\"category\": \"car\", \"corners_3d\": [[547.28, 219.67], [503.22, 219.68], [516.82, 213.65], [555.22, 213.65], [547.28, 182.92], [503.22, 182.92], [516.82, 181.62], [555.22, 181.62]]}]\n```", - "options": null, - "id": 62 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000182", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[263.76, 389.23], [78.57, 391.45], [274.21, 306.18], [386.39, 305.35], [263.76, 220.23], [78.57, 220.71], [274.21, 202.05], [386.39, 201.86]]}, {\"category\": \"van\", \"corners_3d\": [[433.25, 283.01], [330.08, 283.18], [400.21, 254.7], [476.7, 254.61], [433.25, 170.97], [330.08, 170.97], [400.21, 171.46], [476.7, 171.46]]}, {\"category\": \"car\", \"corners_3d\": [[547.28, 219.67], [503.22, 219.68], [516.82, 213.65], [555.22, 213.65], [547.28, 182.92], [503.22, 182.92], [516.82, 181.62], [555.22, 181.62]]}]\n```", - "options": null, - "id": 63 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000194", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[833.54, 203.51], [824.66, 203.61], [818.63, 203.14], [827.4, 203.04], [833.54, 166.97], [824.66, 166.95], [818.63, 167.04], [827.4, 167.06]]}]\n```", - "options": null, - "id": 64 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000194", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[554.25, 237.04], [493.9, 236.67], [521.37, 226.08], [571.8, 226.34], [554.25, 182.28], [493.9, 182.23], [521.37, 180.67], [571.8, 180.71]]}, {\"category\": \"car\", \"corners_3d\": [[617.48, 204.97], [583.87, 204.88], [590.26, 201.89], [620.74, 201.96], [617.48, 175.9], [583.87, 175.89], [590.26, 175.6], [620.74, 175.61]]}]\n```", - "options": null, - "id": 65 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000194", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[554.25, 237.04], [493.9, 236.67], [521.37, 226.08], [571.8, 226.34], [554.25, 182.28], [493.9, 182.23], [521.37, 180.67], [571.8, 180.71]]}, {\"category\": \"car\", \"corners_3d\": [[617.48, 204.97], [583.87, 204.88], [590.26, 201.89], [620.74, 201.96], [617.48, 175.9], [583.87, 175.89], [590.26, 175.6], [620.74, 175.61]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[833.54, 203.51], [824.66, 203.61], [818.63, 203.14], [827.4, 203.04], [833.54, 166.97], [824.66, 166.95], [818.63, 167.04], [827.4, 167.06]]}]\n```", - "options": null, - "id": 66 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000197", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[266.36, 298.08], [165.24, 298.26], [270.28, 268.09], [346.97, 267.98], [266.36, 205.45], [165.24, 205.5], [270.28, 197.64], [346.97, 197.61]]}, {\"category\": \"car\", \"corners_3d\": [[693.96, 235.66], [763.11, 235.66], [803.04, 252.05], [715.84, 252.05], [693.96, 176.0], [763.11, 176.0], [803.04, 176.82], [715.84, 176.82]]}, {\"category\": \"car\", \"corners_3d\": [[381.56, 256.69], [315.52, 256.62], [374.57, 240.17], [427.67, 240.21], [381.56, 196.98], [315.52, 196.96], [374.57, 192.23], [427.67, 192.24]]}, {\"category\": \"car\", \"corners_3d\": [[461.8, 230.56], [413.45, 230.56], [441.41, 222.31], [482.85, 222.31], [461.8, 179.58], [413.45, 179.58], [441.41, 178.62], [482.85, 178.62]]}, {\"category\": \"car\", \"corners_3d\": [[488.06, 219.8], [450.47, 219.75], [470.02, 214.45], [503.37, 214.49], [488.06, 182.74], [450.47, 182.73], [470.02, 181.62], [503.37, 181.62]]}, {\"category\": \"car\", \"corners_3d\": [[636.53, 192.77], [654.23, 192.77], [656.63, 194.08], [637.79, 194.07], [636.53, 176.26], [654.23, 176.26], [656.63, 176.48], [637.79, 176.48]]}]\n```", - "options": null, - "id": 67 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000198", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1402.52, 369.85], [1214.68, 321.8], [1718.1, 325.0], [2078.34, 375.48], [1402.52, 94.0], [1214.68, 113.23], [1718.1, 111.95], [2078.34, 91.74]]}]\n```", - "options": null, - "id": 68 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000198", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[289.43, 305.41], [145.01, 308.46], [268.46, 264.05], [364.42, 262.66], [289.43, 194.79], [145.01, 195.3], [268.46, 187.95], [364.42, 187.72]]}, {\"category\": \"car\", \"corners_3d\": [[406.58, 241.08], [335.26, 241.49], [379.67, 228.26], [437.1, 227.99], [406.58, 176.43], [335.26, 176.45], [379.67, 175.76], [437.1, 175.75]]}]\n```", - "options": null, - "id": 69 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000198", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[289.43, 305.41], [145.01, 308.46], [268.46, 264.05], [364.42, 262.66], [289.43, 194.79], [145.01, 195.3], [268.46, 187.95], [364.42, 187.72]]}, {\"category\": \"car\", \"corners_3d\": [[406.58, 241.08], [335.26, 241.49], [379.67, 228.26], [437.1, 227.99], [406.58, 176.43], [335.26, 176.45], [379.67, 175.76], [437.1, 175.75]]}, {\"category\": \"van\", \"corners_3d\": [[1402.52, 369.85], [1214.68, 321.8], [1718.1, 325.0], [2078.34, 375.48], [1402.52, 94.0], [1214.68, 113.23], [1718.1, 111.95], [2078.34, 91.74]]}]\n```", - "options": null, - "id": 70 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000204", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[481.75, 199.29], [461.83, 199.27], [471.57, 197.68], [490.3, 197.69], [481.75, 178.86], [461.83, 178.86], [471.57, 178.5], [490.3, 178.5]]}]\n```", - "options": null, - "id": 71 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000204", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[600.23, 203.76], [626.96, 203.76], [628.14, 205.92], [599.53, 205.92], [600.23, 178.1], [626.96, 178.1], [628.14, 178.47], [599.53, 178.47]]}, {\"category\": \"car\", \"corners_3d\": [[242.88, 239.48], [192.63, 239.44], [246.99, 230.89], [290.82, 230.92], [242.88, 194.85], [192.63, 194.83], [246.99, 192.01], [290.82, 192.02]]}]\n```", - "options": null, - "id": 72 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000204", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (bus). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[635.92, 192.32], [659.53, 192.27], [679.05, 195.8], [651.19, 195.87], [635.92, 166.51], [659.53, 166.53], [679.05, 165.38], [651.19, 165.35]]}]\n```", - "options": null, - "id": 73 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000204", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[600.23, 203.76], [626.96, 203.76], [628.14, 205.92], [599.53, 205.92], [600.23, 178.1], [626.96, 178.1], [628.14, 178.47], [599.53, 178.47]]}, {\"category\": \"car\", \"corners_3d\": [[242.88, 239.48], [192.63, 239.44], [246.99, 230.89], [290.82, 230.92], [242.88, 194.85], [192.63, 194.83], [246.99, 192.01], [290.82, 192.02]]}, {\"category\": \"van\", \"corners_3d\": [[481.75, 199.29], [461.83, 199.27], [471.57, 197.68], [490.3, 197.69], [481.75, 178.86], [461.83, 178.86], [471.57, 178.5], [490.3, 178.5]]}]\n```", - "options": null, - "id": 74 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000204", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, bus). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[635.92, 192.32], [659.53, 192.27], [679.05, 195.8], [651.19, 195.87], [635.92, 166.51], [659.53, 166.53], [679.05, 165.38], [651.19, 165.35]]}, {\"category\": \"van\", \"corners_3d\": [[481.75, 199.29], [461.83, 199.27], [471.57, 197.68], [490.3, 197.69], [481.75, 178.86], [461.83, 178.86], [471.57, 178.5], [490.3, 178.5]]}]\n```", - "options": null, - "id": 75 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000204", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, bus). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[600.23, 203.76], [626.96, 203.76], [628.14, 205.92], [599.53, 205.92], [600.23, 178.1], [626.96, 178.1], [628.14, 178.47], [599.53, 178.47]]}, {\"category\": \"bus\", \"corners_3d\": [[635.92, 192.32], [659.53, 192.27], [679.05, 195.8], [651.19, 195.87], [635.92, 166.51], [659.53, 166.53], [679.05, 165.38], [651.19, 165.35]]}, {\"category\": \"car\", \"corners_3d\": [[242.88, 239.48], [192.63, 239.44], [246.99, 230.89], [290.82, 230.92], [242.88, 194.85], [192.63, 194.83], [246.99, 192.01], [290.82, 192.02]]}]\n```", - "options": null, - "id": 76 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000204", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car, bus). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[600.23, 203.76], [626.96, 203.76], [628.14, 205.92], [599.53, 205.92], [600.23, 178.1], [626.96, 178.1], [628.14, 178.47], [599.53, 178.47]]}, {\"category\": \"bus\", \"corners_3d\": [[635.92, 192.32], [659.53, 192.27], [679.05, 195.8], [651.19, 195.87], [635.92, 166.51], [659.53, 166.53], [679.05, 165.38], [651.19, 165.35]]}, {\"category\": \"car\", \"corners_3d\": [[242.88, 239.48], [192.63, 239.44], [246.99, 230.89], [290.82, 230.92], [242.88, 194.85], [192.63, 194.83], [246.99, 192.01], [290.82, 192.02]]}, {\"category\": \"van\", \"corners_3d\": [[481.75, 199.29], [461.83, 199.27], [471.57, 197.68], [490.3, 197.69], [481.75, 178.86], [461.83, 178.86], [471.57, 178.5], [490.3, 178.5]]}]\n```", - "options": null, - "id": 77 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000216", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[571.18, 217.01], [608.05, 216.94], [610.49, 222.46], [569.02, 222.54], [571.18, 180.57], [608.05, 180.56], [610.49, 181.53], [569.02, 181.54]]}, {\"category\": \"car\", \"corners_3d\": [[176.99, 280.69], [81.81, 280.84], [203.34, 255.6], [276.17, 255.51], [176.99, 196.15], [81.81, 196.19], [203.34, 190.73], [276.17, 190.71]]}, {\"category\": \"car\", \"corners_3d\": [[347.82, 243.55], [298.0, 243.65], [346.64, 232.03], [388.25, 231.96], [347.82, 191.79], [298.0, 191.81], [346.64, 188.7], [388.25, 188.68]]}, {\"category\": \"car\", \"corners_3d\": [[430.66, 221.07], [395.05, 221.16], [412.71, 216.55], [444.89, 216.47], [430.66, 185.41], [395.05, 185.43], [412.71, 184.23], [444.89, 184.21]]}, {\"category\": \"car\", \"corners_3d\": [[469.16, 211.57], [441.59, 211.57], [456.27, 208.18], [481.42, 208.17], [469.16, 188.74], [441.59, 188.74], [456.27, 187.35], [481.42, 187.35]]}]\n```", - "options": null, - "id": 78 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000220", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[549.07, 197.47], [528.79, 197.33], [544.97, 195.58], [563.88, 195.71], [549.07, 178.17], [528.79, 178.14], [544.97, 177.76], [563.88, 177.79]]}, {\"category\": \"car\", \"corners_3d\": [[718.45, 193.96], [738.75, 194.06], [739.12, 195.72], [717.24, 195.61], [718.45, 175.99], [738.75, 176.01], [739.12, 176.25], [717.24, 176.24]]}]\n```", - "options": null, - "id": 79 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000231", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[571.51, 195.94], [594.03, 195.89], [597.81, 198.41], [572.82, 198.47], [571.51, 164.88], [594.03, 164.89], [597.81, 164.02], [572.82, 164.0]]}]\n```", - "options": null, - "id": 80 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000231", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[563.24, 220.32], [608.69, 220.2], [613.15, 227.9], [560.3, 228.06], [563.24, 184.5], [608.69, 184.47], [613.15, 186.36], [560.3, 186.4]]}, {\"category\": \"car\", \"corners_3d\": [[514.07, 209.64], [547.41, 209.62], [541.17, 213.76], [504.08, 213.78], [514.07, 181.64], [547.41, 181.63], [541.17, 182.62], [504.08, 182.62]]}, {\"category\": \"car\", \"corners_3d\": [[199.15, 235.99], [133.49, 236.54], [188.17, 227.9], [244.58, 227.49], [199.15, 183.42], [133.49, 183.51], [188.17, 182.07], [244.58, 182.0]]}, {\"category\": \"car\", \"corners_3d\": [[340.93, 210.57], [311.32, 210.71], [328.17, 207.96], [355.58, 207.85], [340.93, 184.82], [311.32, 184.86], [328.17, 183.99], [355.58, 183.95]]}]\n```", - "options": null, - "id": 81 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000231", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[563.24, 220.32], [608.69, 220.2], [613.15, 227.9], [560.3, 228.06], [563.24, 184.5], [608.69, 184.47], [613.15, 186.36], [560.3, 186.4]]}, {\"category\": \"car\", \"corners_3d\": [[514.07, 209.64], [547.41, 209.62], [541.17, 213.76], [504.08, 213.78], [514.07, 181.64], [547.41, 181.63], [541.17, 182.62], [504.08, 182.62]]}, {\"category\": \"car\", \"corners_3d\": [[199.15, 235.99], [133.49, 236.54], [188.17, 227.9], [244.58, 227.49], [199.15, 183.42], [133.49, 183.51], [188.17, 182.07], [244.58, 182.0]]}, {\"category\": \"car\", \"corners_3d\": [[340.93, 210.57], [311.32, 210.71], [328.17, 207.96], [355.58, 207.85], [340.93, 184.82], [311.32, 184.86], [328.17, 183.99], [355.58, 183.95]]}, {\"category\": \"van\", \"corners_3d\": [[571.51, 195.94], [594.03, 195.89], [597.81, 198.41], [572.82, 198.47], [571.51, 164.88], [594.03, 164.89], [597.81, 164.02], [572.82, 164.0]]}]\n```", - "options": null, - "id": 82 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000232", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[272.28, 253.29], [207.39, 253.23], [288.26, 237.33], [340.36, 237.37], [272.28, 190.41], [207.39, 190.4], [288.26, 186.93], [340.36, 186.94]]}]\n```", - "options": null, - "id": 83 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000232", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[937.65, 210.6], [951.27, 210.6], [978.1, 213.57], [963.41, 213.57], [937.65, 155.51], [951.27, 155.51], [978.1, 154.15], [963.41, 154.15]]}]\n```", - "options": null, - "id": 84 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000232", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[625.01, 186.31], [645.12, 186.26], [696.89, 191.34], [669.4, 191.44], [625.01, 157.58], [645.12, 157.64], [696.89, 151.87], [669.4, 151.76]]}]\n```", - "options": null, - "id": 85 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000232", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[272.28, 253.29], [207.39, 253.23], [288.26, 237.33], [340.36, 237.37], [272.28, 190.41], [207.39, 190.4], [288.26, 186.93], [340.36, 186.94]]}, {\"category\": \"cyclist\", \"corners_3d\": [[937.65, 210.6], [951.27, 210.6], [978.1, 213.57], [963.41, 213.57], [937.65, 155.51], [951.27, 155.51], [978.1, 154.15], [963.41, 154.15]]}]\n```", - "options": null, - "id": 86 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000232", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[625.01, 186.31], [645.12, 186.26], [696.89, 191.34], [669.4, 191.44], [625.01, 157.58], [645.12, 157.64], [696.89, 151.87], [669.4, 151.76]]}, {\"category\": \"car\", \"corners_3d\": [[272.28, 253.29], [207.39, 253.23], [288.26, 237.33], [340.36, 237.37], [272.28, 190.41], [207.39, 190.4], [288.26, 186.93], [340.36, 186.94]]}]\n```", - "options": null, - "id": 87 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000232", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[625.01, 186.31], [645.12, 186.26], [696.89, 191.34], [669.4, 191.44], [625.01, 157.58], [645.12, 157.64], [696.89, 151.87], [669.4, 151.76]]}, {\"category\": \"cyclist\", \"corners_3d\": [[937.65, 210.6], [951.27, 210.6], [978.1, 213.57], [963.41, 213.57], [937.65, 155.51], [951.27, 155.51], [978.1, 154.15], [963.41, 154.15]]}]\n```", - "options": null, - "id": 88 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000232", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist, tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[625.01, 186.31], [645.12, 186.26], [696.89, 191.34], [669.4, 191.44], [625.01, 157.58], [645.12, 157.64], [696.89, 151.87], [669.4, 151.76]]}, {\"category\": \"car\", \"corners_3d\": [[272.28, 253.29], [207.39, 253.23], [288.26, 237.33], [340.36, 237.37], [272.28, 190.41], [207.39, 190.4], [288.26, 186.93], [340.36, 186.94]]}, {\"category\": \"cyclist\", \"corners_3d\": [[937.65, 210.6], [951.27, 210.6], [978.1, 213.57], [963.41, 213.57], [937.65, 155.51], [951.27, 155.51], [978.1, 154.15], [963.41, 154.15]]}]\n```", - "options": null, - "id": 89 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000233", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[527.59, 210.76], [562.21, 210.41], [574.03, 214.92], [535.34, 215.37], [527.59, 177.23], [562.21, 177.19], [574.03, 177.71], [535.34, 177.76]]}, {\"category\": \"car\", \"corners_3d\": [[460.47, 201.17], [432.74, 201.3], [439.91, 199.01], [465.4, 198.9], [460.47, 173.53], [432.74, 173.53], [439.91, 173.48], [465.4, 173.47]]}]\n```", - "options": null, - "id": 90 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000235", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.18, 200.22], [614.88, 200.2], [616.42, 202.41], [589.73, 202.43], [590.18, 175.81], [614.88, 175.81], [616.42, 176.04], [589.73, 176.05]]}, {\"category\": \"car\", \"corners_3d\": [[156.29, 287.02], [61.0, 287.62], [186.7, 259.83], [258.51, 259.49], [156.29, 204.81], [61.0, 204.97], [186.7, 197.2], [258.51, 197.1]]}, {\"category\": \"car\", \"corners_3d\": [[439.2, 214.7], [405.81, 214.83], [420.37, 210.83], [450.53, 210.72], [439.2, 181.69], [405.81, 181.72], [420.37, 180.87], [450.53, 180.85]]}, {\"category\": \"car\", \"corners_3d\": [[555.6, 196.39], [572.96, 196.38], [571.2, 198.2], [552.51, 198.21], [555.6, 177.9], [572.96, 177.9], [571.2, 178.29], [552.51, 178.29]]}]\n```", - "options": null, - "id": 91 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000246", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[302.02, 238.13], [369.82, 237.94], [272.66, 262.42], [172.69, 262.84], [302.02, 147.65], [369.82, 147.79], [272.66, 130.41], [172.69, 130.11]]}]\n```", - "options": null, - "id": 92 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000246", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[572.34, 209.08], [606.5, 209.03], [609.67, 211.78], [571.57, 211.84], [572.34, 178.01], [606.5, 178.02], [609.67, 177.19], [571.57, 177.18]]}, {\"category\": \"car\", \"corners_3d\": [[109.33, 364.89], [276.51, 364.51], [-79.96, 561.74], [-433.47, 563.43], [109.33, 210.62], [276.51, 210.56], [-79.96, 238.45], [-433.47, 238.69]]}, {\"category\": \"car\", \"corners_3d\": [[-3.95, 291.77], [97.89, 291.93], [-160.61, 345.28], [-312.64, 344.91], [-3.95, 195.14], [97.89, 195.15], [-160.61, 200.12], [-312.64, 200.08]]}, {\"category\": \"car\", \"corners_3d\": [[44.04, 250.05], [120.85, 249.92], [-10.99, 267.98], [-109.63, 268.19], [44.04, 188.96], [120.85, 188.95], [-10.99, 190.0], [-109.63, 190.01]]}, {\"category\": \"car\", \"corners_3d\": [[363.64, 223.54], [408.05, 223.52], [375.81, 229.93], [323.94, 229.96], [363.64, 183.56], [408.05, 183.57], [375.81, 183.29], [323.94, 183.29]]}]\n```", - "options": null, - "id": 93 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000246", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[537.59, 194.75], [563.29, 194.74], [558.49, 197.65], [524.9, 197.67], [537.59, 153.18], [563.29, 153.23], [558.49, 143.44], [524.9, 143.37]]}]\n```", - "options": null, - "id": 94 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000246", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[572.34, 209.08], [606.5, 209.03], [609.67, 211.78], [571.57, 211.84], [572.34, 178.01], [606.5, 178.02], [609.67, 177.19], [571.57, 177.18]]}, {\"category\": \"car\", \"corners_3d\": [[109.33, 364.89], [276.51, 364.51], [-79.96, 561.74], [-433.47, 563.43], [109.33, 210.62], [276.51, 210.56], [-79.96, 238.45], [-433.47, 238.69]]}, {\"category\": \"car\", \"corners_3d\": [[-3.95, 291.77], [97.89, 291.93], [-160.61, 345.28], [-312.64, 344.91], [-3.95, 195.14], [97.89, 195.15], [-160.61, 200.12], [-312.64, 200.08]]}, {\"category\": \"van\", \"corners_3d\": [[302.02, 238.13], [369.82, 237.94], [272.66, 262.42], [172.69, 262.84], [302.02, 147.65], [369.82, 147.79], [272.66, 130.41], [172.69, 130.11]]}, {\"category\": \"car\", \"corners_3d\": [[44.04, 250.05], [120.85, 249.92], [-10.99, 267.98], [-109.63, 268.19], [44.04, 188.96], [120.85, 188.95], [-10.99, 190.0], [-109.63, 190.01]]}, {\"category\": \"car\", \"corners_3d\": [[363.64, 223.54], [408.05, 223.52], [375.81, 229.93], [323.94, 229.96], [363.64, 183.56], [408.05, 183.57], [375.81, 183.29], [323.94, 183.29]]}]\n```", - "options": null, - "id": 95 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000246", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[302.02, 238.13], [369.82, 237.94], [272.66, 262.42], [172.69, 262.84], [302.02, 147.65], [369.82, 147.79], [272.66, 130.41], [172.69, 130.11]]}, {\"category\": \"truck\", \"corners_3d\": [[537.59, 194.75], [563.29, 194.74], [558.49, 197.65], [524.9, 197.67], [537.59, 153.18], [563.29, 153.23], [558.49, 143.44], [524.9, 143.37]]}]\n```", - "options": null, - "id": 96 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000246", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[572.34, 209.08], [606.5, 209.03], [609.67, 211.78], [571.57, 211.84], [572.34, 178.01], [606.5, 178.02], [609.67, 177.19], [571.57, 177.18]]}, {\"category\": \"car\", \"corners_3d\": [[109.33, 364.89], [276.51, 364.51], [-79.96, 561.74], [-433.47, 563.43], [109.33, 210.62], [276.51, 210.56], [-79.96, 238.45], [-433.47, 238.69]]}, {\"category\": \"car\", \"corners_3d\": [[-3.95, 291.77], [97.89, 291.93], [-160.61, 345.28], [-312.64, 344.91], [-3.95, 195.14], [97.89, 195.15], [-160.61, 200.12], [-312.64, 200.08]]}, {\"category\": \"car\", \"corners_3d\": [[44.04, 250.05], [120.85, 249.92], [-10.99, 267.98], [-109.63, 268.19], [44.04, 188.96], [120.85, 188.95], [-10.99, 190.0], [-109.63, 190.01]]}, {\"category\": \"car\", \"corners_3d\": [[363.64, 223.54], [408.05, 223.52], [375.81, 229.93], [323.94, 229.96], [363.64, 183.56], [408.05, 183.57], [375.81, 183.29], [323.94, 183.29]]}, {\"category\": \"truck\", \"corners_3d\": [[537.59, 194.75], [563.29, 194.74], [558.49, 197.65], [524.9, 197.67], [537.59, 153.18], [563.29, 153.23], [558.49, 143.44], [524.9, 143.37]]}]\n```", - "options": null, - "id": 97 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000246", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[572.34, 209.08], [606.5, 209.03], [609.67, 211.78], [571.57, 211.84], [572.34, 178.01], [606.5, 178.02], [609.67, 177.19], [571.57, 177.18]]}, {\"category\": \"car\", \"corners_3d\": [[109.33, 364.89], [276.51, 364.51], [-79.96, 561.74], [-433.47, 563.43], [109.33, 210.62], [276.51, 210.56], [-79.96, 238.45], [-433.47, 238.69]]}, {\"category\": \"car\", \"corners_3d\": [[-3.95, 291.77], [97.89, 291.93], [-160.61, 345.28], [-312.64, 344.91], [-3.95, 195.14], [97.89, 195.15], [-160.61, 200.12], [-312.64, 200.08]]}, {\"category\": \"van\", \"corners_3d\": [[302.02, 238.13], [369.82, 237.94], [272.66, 262.42], [172.69, 262.84], [302.02, 147.65], [369.82, 147.79], [272.66, 130.41], [172.69, 130.11]]}, {\"category\": \"car\", \"corners_3d\": [[44.04, 250.05], [120.85, 249.92], [-10.99, 267.98], [-109.63, 268.19], [44.04, 188.96], [120.85, 188.95], [-10.99, 190.0], [-109.63, 190.01]]}, {\"category\": \"car\", \"corners_3d\": [[363.64, 223.54], [408.05, 223.52], [375.81, 229.93], [323.94, 229.96], [363.64, 183.56], [408.05, 183.57], [375.81, 183.29], [323.94, 183.29]]}, {\"category\": \"truck\", \"corners_3d\": [[537.59, 194.75], [563.29, 194.74], [558.49, 197.65], [524.9, 197.67], [537.59, 153.18], [563.29, 153.23], [558.49, 143.44], [524.9, 143.37]]}]\n```", - "options": null, - "id": 98 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000255", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[353.46, 349.02], [189.84, 348.25], [346.45, 284.9], [451.24, 285.21], [353.46, 203.0], [189.84, 202.87], [346.45, 192.03], [451.24, 192.08]]}, {\"category\": \"car\", \"corners_3d\": [[568.26, 208.77], [535.36, 208.76], [543.23, 205.26], [572.92, 205.27], [568.26, 182.71], [535.36, 182.7], [543.23, 181.75], [572.92, 181.75]]}]\n```", - "options": null, - "id": 99 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000260", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[581.94, 190.41], [603.06, 190.39], [604.08, 191.62], [581.49, 191.63], [581.94, 173.77], [603.06, 173.77], [604.08, 173.84], [581.49, 173.84]]}, {\"category\": \"car\", \"corners_3d\": [[556.67, 188.4], [574.66, 188.38], [573.86, 189.28], [554.84, 189.29], [556.67, 173.29], [574.66, 173.29], [573.86, 173.32], [554.84, 173.32]]}]\n```", - "options": null, - "id": 100 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000261", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[305.68, 206.9], [278.42, 206.84], [309.09, 204.0], [334.12, 204.05], [305.68, 180.0], [278.42, 179.99], [309.09, 179.39], [334.12, 179.4]]}, {\"category\": \"car\", \"corners_3d\": [[342.24, 202.0], [315.36, 201.96], [341.14, 199.63], [365.91, 199.67], [342.24, 180.4], [315.36, 180.39], [341.14, 179.79], [365.91, 179.8]]}, {\"category\": \"car\", \"corners_3d\": [[382.31, 196.74], [361.07, 196.69], [382.58, 194.97], [402.32, 195.02], [382.31, 177.55], [361.07, 177.54], [382.58, 177.2], [402.32, 177.21]]}]\n```", - "options": null, - "id": 101 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000261", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[642.42, 197.29], [685.39, 197.52], [673.69, 204.52], [618.82, 204.14], [642.42, 145.15], [685.39, 144.89], [673.69, 136.95], [618.82, 137.39]]}]\n```", - "options": null, - "id": 102 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000261", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[642.42, 197.29], [685.39, 197.52], [673.69, 204.52], [618.82, 204.14], [642.42, 145.15], [685.39, 144.89], [673.69, 136.95], [618.82, 137.39]]}, {\"category\": \"car\", \"corners_3d\": [[305.68, 206.9], [278.42, 206.84], [309.09, 204.0], [334.12, 204.05], [305.68, 180.0], [278.42, 179.99], [309.09, 179.39], [334.12, 179.4]]}, {\"category\": \"car\", \"corners_3d\": [[342.24, 202.0], [315.36, 201.96], [341.14, 199.63], [365.91, 199.67], [342.24, 180.4], [315.36, 180.39], [341.14, 179.79], [365.91, 179.8]]}, {\"category\": \"car\", \"corners_3d\": [[382.31, 196.74], [361.07, 196.69], [382.58, 194.97], [402.32, 195.02], [382.31, 177.55], [361.07, 177.54], [382.58, 177.2], [402.32, 177.21]]}]\n```", - "options": null, - "id": 103 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000269", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[77.59, 345.79], [-181.01, 374.42], [-72.42, 297.73], [86.85, 286.12], [77.59, 155.81], [-181.01, 152.99], [-72.42, 160.54], [86.85, 161.69]]}, {\"category\": \"car\", \"corners_3d\": [[62.58, 236.65], [118.38, 234.46], [117.19, 242.63], [53.65, 245.45], [62.58, 191.07], [118.38, 190.45], [117.19, 192.78], [53.65, 193.59]]}, {\"category\": \"car\", \"corners_3d\": [[273.0, 230.0], [317.05, 228.22], [337.1, 234.07], [288.94, 236.25], [273.0, 188.03], [317.05, 187.56], [337.1, 189.11], [288.94, 189.69]]}, {\"category\": \"car\", \"corners_3d\": [[243.96, 225.63], [285.06, 224.11], [302.64, 229.46], [257.68, 231.33], [243.96, 188.75], [285.06, 188.29], [302.64, 189.91], [257.68, 190.47]]}, {\"category\": \"car\", \"corners_3d\": [[223.67, 213.98], [254.74, 213.25], [257.21, 216.65], [223.52, 217.52], [223.67, 188.54], [254.74, 188.26], [257.21, 189.55], [223.52, 189.88]]}, {\"category\": \"car\", \"corners_3d\": [[300.98, 240.39], [361.74, 237.49], [396.61, 247.55], [327.73, 251.45], [300.98, 186.28], [361.74, 185.71], [396.61, 187.71], [327.73, 188.48]]}]\n```", - "options": null, - "id": 104 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000269", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[123.06, 223.46], [63.53, 225.33], [74.42, 218.43], [126.02, 217.01], [123.06, 163.37], [63.53, 163.02], [74.42, 164.31], [126.02, 164.58]]}]\n```", - "options": null, - "id": 105 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000269", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[77.59, 345.79], [-181.01, 374.42], [-72.42, 297.73], [86.85, 286.12], [77.59, 155.81], [-181.01, 152.99], [-72.42, 160.54], [86.85, 161.69]]}, {\"category\": \"car\", \"corners_3d\": [[62.58, 236.65], [118.38, 234.46], [117.19, 242.63], [53.65, 245.45], [62.58, 191.07], [118.38, 190.45], [117.19, 192.78], [53.65, 193.59]]}, {\"category\": \"car\", \"corners_3d\": [[273.0, 230.0], [317.05, 228.22], [337.1, 234.07], [288.94, 236.25], [273.0, 188.03], [317.05, 187.56], [337.1, 189.11], [288.94, 189.69]]}, {\"category\": \"car\", \"corners_3d\": [[243.96, 225.63], [285.06, 224.11], [302.64, 229.46], [257.68, 231.33], [243.96, 188.75], [285.06, 188.29], [302.64, 189.91], [257.68, 190.47]]}, {\"category\": \"truck\", \"corners_3d\": [[123.06, 223.46], [63.53, 225.33], [74.42, 218.43], [126.02, 217.01], [123.06, 163.37], [63.53, 163.02], [74.42, 164.31], [126.02, 164.58]]}, {\"category\": \"car\", \"corners_3d\": [[223.67, 213.98], [254.74, 213.25], [257.21, 216.65], [223.52, 217.52], [223.67, 188.54], [254.74, 188.26], [257.21, 189.55], [223.52, 189.88]]}, {\"category\": \"car\", \"corners_3d\": [[300.98, 240.39], [361.74, 237.49], [396.61, 247.55], [327.73, 251.45], [300.98, 186.28], [361.74, 185.71], [396.61, 187.71], [327.73, 188.48]]}]\n```", - "options": null, - "id": 106 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000282", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[277.63, 425.24], [37.09, 431.81], [281.09, 308.89], [405.71, 307.05], [277.63, 224.15], [37.09, 225.48], [281.09, 200.5], [405.71, 200.13]]}, {\"category\": \"car\", \"corners_3d\": [[595.38, 200.78], [617.75, 200.67], [624.3, 202.46], [600.52, 202.58], [595.38, 178.86], [617.75, 178.84], [624.3, 179.22], [600.52, 179.25]]}]\n```", - "options": null, - "id": 107 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000299", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[613.06, 205.4], [593.81, 205.41], [594.37, 203.13], [612.27, 203.13], [613.06, 188.34], [593.81, 188.34], [594.37, 187.26], [612.27, 187.26]]}]\n```", - "options": null, - "id": 108 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000318", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[283.07, 450.02], [18.1, 456.05], [278.75, 321.24], [416.07, 319.56], [283.07, 159.09], [18.1, 158.79], [278.75, 165.49], [416.07, 165.57]]}, {\"category\": \"car\", \"corners_3d\": [[477.34, 302.97], [362.6, 303.4], [429.02, 266.02], [510.82, 265.8], [477.34, 179.57], [362.6, 179.59], [429.02, 177.66], [510.82, 177.65]]}, {\"category\": \"car\", \"corners_3d\": [[575.52, 233.72], [532.78, 232.95], [562.42, 225.2], [599.91, 225.79], [575.52, 192.31], [532.78, 192.07], [562.42, 189.59], [599.91, 189.78]]}, {\"category\": \"car\", \"corners_3d\": [[609.93, 224.87], [579.87, 224.22], [602.92, 219.38], [630.36, 219.91], [609.93, 190.26], [579.87, 190.04], [602.92, 188.42], [630.36, 188.6]]}, {\"category\": \"car\", \"corners_3d\": [[654.12, 218.73], [626.98, 218.11], [646.26, 214.56], [671.47, 215.08], [654.12, 190.86], [626.98, 190.62], [646.26, 189.22], [671.47, 189.43]]}]\n```", - "options": null, - "id": 109 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000325", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[987.63, 242.47], [998.86, 242.42], [1032.92, 247.99], [1020.82, 248.06], [987.63, 186.77], [998.86, 186.76], [1032.92, 187.87], [1020.82, 187.88]]}]\n```", - "options": null, - "id": 110 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000325", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[406.43, 279.58], [317.13, 279.59], [379.02, 256.96], [449.38, 256.95], [406.43, 187.81], [317.13, 187.82], [379.02, 184.64], [449.38, 184.64]]}, {\"category\": \"car\", \"corners_3d\": [[23.09, 230.52], [-18.57, 230.46], [72.53, 222.28], [108.36, 222.33], [23.09, 180.2], [-18.57, 180.19], [72.53, 179.15], [108.36, 179.16]]}, {\"category\": \"car\", \"corners_3d\": [[147.29, 218.4], [108.85, 218.4], [163.91, 213.38], [198.12, 213.38], [147.29, 183.2], [108.85, 183.2], [163.91, 182.06], [198.12, 182.06]]}, {\"category\": \"car\", \"corners_3d\": [[225.4, 214.56], [196.5, 214.54], [238.4, 210.38], [264.43, 210.39], [225.4, 184.31], [196.5, 184.31], [238.4, 183.16], [264.43, 183.17]]}, {\"category\": \"car\", \"corners_3d\": [[540.81, 212.71], [508.79, 212.7], [522.63, 207.56], [550.53, 207.57], [540.81, 182.16], [508.79, 182.16], [522.63, 180.96], [550.53, 180.96]]}, {\"category\": \"car\", \"corners_3d\": [[571.26, 194.7], [554.67, 194.71], [557.3, 193.49], [572.96, 193.49], [571.26, 178.34], [554.67, 178.34], [557.3, 178.04], [572.96, 178.03]]}, {\"category\": \"car\", \"corners_3d\": [[586.16, 217.75], [626.6, 217.75], [629.61, 225.97], [581.77, 225.97], [586.16, 177.56], [626.6, 177.56], [629.61, 178.42], [581.77, 178.42]]}]\n```", - "options": null, - "id": 111 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000325", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[406.43, 279.58], [317.13, 279.59], [379.02, 256.96], [449.38, 256.95], [406.43, 187.81], [317.13, 187.82], [379.02, 184.64], [449.38, 184.64]]}, {\"category\": \"car\", \"corners_3d\": [[23.09, 230.52], [-18.57, 230.46], [72.53, 222.28], [108.36, 222.33], [23.09, 180.2], [-18.57, 180.19], [72.53, 179.15], [108.36, 179.16]]}, {\"category\": \"car\", \"corners_3d\": [[147.29, 218.4], [108.85, 218.4], [163.91, 213.38], [198.12, 213.38], [147.29, 183.2], [108.85, 183.2], [163.91, 182.06], [198.12, 182.06]]}, {\"category\": \"car\", \"corners_3d\": [[225.4, 214.56], [196.5, 214.54], [238.4, 210.38], [264.43, 210.39], [225.4, 184.31], [196.5, 184.31], [238.4, 183.16], [264.43, 183.17]]}, {\"category\": \"car\", \"corners_3d\": [[540.81, 212.71], [508.79, 212.7], [522.63, 207.56], [550.53, 207.57], [540.81, 182.16], [508.79, 182.16], [522.63, 180.96], [550.53, 180.96]]}, {\"category\": \"car\", \"corners_3d\": [[571.26, 194.7], [554.67, 194.71], [557.3, 193.49], [572.96, 193.49], [571.26, 178.34], [554.67, 178.34], [557.3, 178.04], [572.96, 178.03]]}, {\"category\": \"car\", \"corners_3d\": [[586.16, 217.75], [626.6, 217.75], [629.61, 225.97], [581.77, 225.97], [586.16, 177.56], [626.6, 177.56], [629.61, 178.42], [581.77, 178.42]]}, {\"category\": \"cyclist\", \"corners_3d\": [[987.63, 242.47], [998.86, 242.42], [1032.92, 247.99], [1020.82, 248.06], [987.63, 186.77], [998.86, 186.76], [1032.92, 187.87], [1020.82, 187.88]]}]\n```", - "options": null, - "id": 112 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000331", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[604.22, 222.24], [621.28, 222.18], [625.36, 226.44], [606.83, 226.51], [604.22, 175.83], [621.28, 175.83], [625.36, 176.08], [606.83, 176.09]]}]\n```", - "options": null, - "id": 113 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000331", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-40.43, 318.64], [45.76, 297.64], [277.76, 294.91], [231.82, 314.94], [-40.43, 211.91], [45.76, 206.28], [277.76, 205.55], [231.82, 210.92]]}, {\"category\": \"car\", \"corners_3d\": [[315.89, 238.68], [333.43, 234.3], [446.7, 233.91], [437.3, 238.23], [315.89, 191.56], [333.43, 190.31], [446.7, 190.2], [437.3, 191.43]]}]\n```", - "options": null, - "id": 114 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000331", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[604.22, 222.24], [621.28, 222.18], [625.36, 226.44], [606.83, 226.51], [604.22, 175.83], [621.28, 175.83], [625.36, 176.08], [606.83, 176.09]]}, {\"category\": \"car\", \"corners_3d\": [[-40.43, 318.64], [45.76, 297.64], [277.76, 294.91], [231.82, 314.94], [-40.43, 211.91], [45.76, 206.28], [277.76, 205.55], [231.82, 210.92]]}, {\"category\": \"car\", \"corners_3d\": [[315.89, 238.68], [333.43, 234.3], [446.7, 233.91], [437.3, 238.23], [315.89, 191.56], [333.43, 190.31], [446.7, 190.2], [437.3, 191.43]]}]\n```", - "options": null, - "id": 115 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000333", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[992.57, 213.46], [1025.85, 214.1], [1029.66, 217.01], [993.47, 216.26], [992.57, 150.36], [1025.85, 149.77], [1029.66, 147.11], [993.47, 147.8]]}]\n```", - "options": null, - "id": 116 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000333", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[66.35, 407.85], [306.36, 408.69], [-439.54, 966.08], [-1249.13, 956.21], [66.35, 204.65], [306.36, 204.74], [-439.54, 263.94], [-1249.13, 262.89]]}, {\"category\": \"car\", \"corners_3d\": [[592.08, 205.86], [561.33, 206.94], [503.73, 204.62], [534.03, 203.72], [592.08, 170.0], [561.33, 169.56], [503.73, 170.52], [534.03, 170.89]]}, {\"category\": \"car\", \"corners_3d\": [[-19.25, 276.13], [-126.25, 277.48], [46.13, 252.11], [123.64, 251.37], [-19.25, 186.97], [-126.25, 187.07], [46.13, 185.35], [123.64, 185.3]]}]\n```", - "options": null, - "id": 117 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000333", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[66.35, 407.85], [306.36, 408.69], [-439.54, 966.08], [-1249.13, 956.21], [66.35, 204.65], [306.36, 204.74], [-439.54, 263.94], [-1249.13, 262.89]]}, {\"category\": \"cyclist\", \"corners_3d\": [[992.57, 213.46], [1025.85, 214.1], [1029.66, 217.01], [993.47, 216.26], [992.57, 150.36], [1025.85, 149.77], [1029.66, 147.11], [993.47, 147.8]]}, {\"category\": \"car\", \"corners_3d\": [[592.08, 205.86], [561.33, 206.94], [503.73, 204.62], [534.03, 203.72], [592.08, 170.0], [561.33, 169.56], [503.73, 170.52], [534.03, 170.89]]}, {\"category\": \"car\", \"corners_3d\": [[-19.25, 276.13], [-126.25, 277.48], [46.13, 252.11], [123.64, 251.37], [-19.25, 186.97], [-126.25, 187.07], [46.13, 185.35], [123.64, 185.3]]}]\n```", - "options": null, - "id": 118 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000338", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[823.84, 296.88], [949.17, 296.48], [1279.59, 403.32], [1040.23, 404.79], [823.84, 179.6], [949.17, 179.6], [1279.59, 178.77], [1040.23, 178.76]]}, {\"category\": \"car\", \"corners_3d\": [[113.45, 247.94], [47.41, 247.34], [168.85, 234.25], [222.74, 234.63], [113.45, 169.39], [47.41, 169.49], [168.85, 171.65], [222.74, 171.59]]}, {\"category\": \"car\", \"corners_3d\": [[476.72, 198.11], [454.12, 198.03], [473.07, 196.73], [494.08, 196.8], [476.72, 175.85], [454.12, 175.87], [473.07, 176.21], [494.08, 176.19]]}, {\"category\": \"car\", \"corners_3d\": [[701.69, 227.01], [756.09, 227.47], [768.13, 237.06], [702.61, 236.39], [701.69, 174.67], [756.09, 174.61], [768.13, 173.41], [702.61, 173.49]]}, {\"category\": \"car\", \"corners_3d\": [[684.81, 217.89], [729.83, 217.94], [747.78, 223.99], [695.46, 223.93], [684.81, 176.6], [729.83, 176.6], [747.78, 175.97], [695.46, 175.97]]}, {\"category\": \"car\", \"corners_3d\": [[670.57, 209.62], [706.2, 209.71], [713.86, 213.49], [673.6, 213.38], [670.57, 176.91], [706.2, 176.9], [713.86, 176.43], [673.6, 176.44]]}, {\"category\": \"car\", \"corners_3d\": [[663.2, 204.65], [696.5, 204.71], [703.5, 207.71], [666.06, 207.63], [663.2, 176.29], [696.5, 176.27], [703.5, 175.75], [666.06, 175.76]]}, {\"category\": \"car\", \"corners_3d\": [[652.11, 196.8], [675.41, 196.82], [678.52, 198.01], [653.52, 197.99], [652.11, 175.63], [675.41, 175.62], [678.52, 175.27], [653.52, 175.27]]}]\n```", - "options": null, - "id": 119 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000340", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1129.79, 166.76], [1179.25, 166.57], [1190.81, 165.82], [1135.22, 166.05], [1129.79, 118.66], [1179.25, 117.0], [1190.81, 110.28], [1135.22, 112.35]]}]\n```", - "options": null, - "id": 120 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000340", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[157.74, 326.67], [-2.53, 338.62], [81.34, 290.54], [193.17, 284.39], [157.74, 210.68], [-2.53, 213.62], [81.34, 201.8], [193.17, 200.29]]}, {\"category\": \"car\", \"corners_3d\": [[299.75, 235.1], [263.83, 236.11], [270.53, 230.34], [303.13, 229.51], [299.75, 205.37], [263.83, 205.89], [270.53, 202.88], [303.13, 202.44]]}, {\"category\": \"car\", \"corners_3d\": [[249.21, 274.51], [163.6, 278.14], [196.9, 259.7], [266.98, 257.21], [249.21, 206.1], [163.6, 207.29], [196.9, 201.26], [266.98, 200.45]]}]\n```", - "options": null, - "id": 121 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000340", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[157.74, 326.67], [-2.53, 338.62], [81.34, 290.54], [193.17, 284.39], [157.74, 210.68], [-2.53, 213.62], [81.34, 201.8], [193.17, 200.29]]}, {\"category\": \"car\", \"corners_3d\": [[299.75, 235.1], [263.83, 236.11], [270.53, 230.34], [303.13, 229.51], [299.75, 205.37], [263.83, 205.89], [270.53, 202.88], [303.13, 202.44]]}, {\"category\": \"car\", \"corners_3d\": [[249.21, 274.51], [163.6, 278.14], [196.9, 259.7], [266.98, 257.21], [249.21, 206.1], [163.6, 207.29], [196.9, 201.26], [266.98, 200.45]]}, {\"category\": \"van\", \"corners_3d\": [[1129.79, 166.76], [1179.25, 166.57], [1190.81, 165.82], [1135.22, 166.05], [1129.79, 118.66], [1179.25, 117.0], [1190.81, 110.28], [1135.22, 112.35]]}]\n```", - "options": null, - "id": 122 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000347", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[119.87, 238.12], [106.07, 238.07], [126.33, 235.59], [139.61, 235.64], [119.87, 195.79], [106.07, 195.77], [126.33, 194.9], [139.61, 194.91]]}, {\"category\": \"cyclist\", \"corners_3d\": [[722.78, 217.79], [739.57, 217.77], [751.84, 221.6], [733.62, 221.62], [722.78, 165.54], [739.57, 165.54], [751.84, 164.92], [733.62, 164.92]]}]\n```", - "options": null, - "id": 123 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000347", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[304.31, 232.56], [261.23, 232.6], [302.5, 225.36], [340.34, 225.33], [304.31, 193.31], [261.23, 193.32], [302.5, 190.84], [340.34, 190.83]]}]\n```", - "options": null, - "id": 124 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000347", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (bus). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[592.09, 196.03], [618.58, 196.01], [623.34, 200.83], [591.34, 200.85], [592.09, 167.31], [618.58, 167.31], [623.34, 166.16], [591.34, 166.15]]}]\n```", - "options": null, - "id": 125 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000347", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[119.87, 238.12], [106.07, 238.07], [126.33, 235.59], [139.61, 235.64], [119.87, 195.79], [106.07, 195.77], [126.33, 194.9], [139.61, 194.91]]}, {\"category\": \"car\", \"corners_3d\": [[304.31, 232.56], [261.23, 232.6], [302.5, 225.36], [340.34, 225.33], [304.31, 193.31], [261.23, 193.32], [302.5, 190.84], [340.34, 190.83]]}, {\"category\": \"cyclist\", \"corners_3d\": [[722.78, 217.79], [739.57, 217.77], [751.84, 221.6], [733.62, 221.62], [722.78, 165.54], [739.57, 165.54], [751.84, 164.92], [733.62, 164.92]]}]\n```", - "options": null, - "id": 126 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000347", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, bus). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[592.09, 196.03], [618.58, 196.01], [623.34, 200.83], [591.34, 200.85], [592.09, 167.31], [618.58, 167.31], [623.34, 166.16], [591.34, 166.15]]}, {\"category\": \"cyclist\", \"corners_3d\": [[119.87, 238.12], [106.07, 238.07], [126.33, 235.59], [139.61, 235.64], [119.87, 195.79], [106.07, 195.77], [126.33, 194.9], [139.61, 194.91]]}, {\"category\": \"cyclist\", \"corners_3d\": [[722.78, 217.79], [739.57, 217.77], [751.84, 221.6], [733.62, 221.62], [722.78, 165.54], [739.57, 165.54], [751.84, 164.92], [733.62, 164.92]]}]\n```", - "options": null, - "id": 127 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000347", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, bus). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[592.09, 196.03], [618.58, 196.01], [623.34, 200.83], [591.34, 200.85], [592.09, 167.31], [618.58, 167.31], [623.34, 166.16], [591.34, 166.15]]}, {\"category\": \"car\", \"corners_3d\": [[304.31, 232.56], [261.23, 232.6], [302.5, 225.36], [340.34, 225.33], [304.31, 193.31], [261.23, 193.32], [302.5, 190.84], [340.34, 190.83]]}]\n```", - "options": null, - "id": 128 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000347", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car, bus). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[592.09, 196.03], [618.58, 196.01], [623.34, 200.83], [591.34, 200.85], [592.09, 167.31], [618.58, 167.31], [623.34, 166.16], [591.34, 166.15]]}, {\"category\": \"cyclist\", \"corners_3d\": [[119.87, 238.12], [106.07, 238.07], [126.33, 235.59], [139.61, 235.64], [119.87, 195.79], [106.07, 195.77], [126.33, 194.9], [139.61, 194.91]]}, {\"category\": \"car\", \"corners_3d\": [[304.31, 232.56], [261.23, 232.6], [302.5, 225.36], [340.34, 225.33], [304.31, 193.31], [261.23, 193.32], [302.5, 190.84], [340.34, 190.83]]}, {\"category\": \"cyclist\", \"corners_3d\": [[722.78, 217.79], [739.57, 217.77], [751.84, 221.6], [733.62, 221.62], [722.78, 165.54], [739.57, 165.54], [751.84, 164.92], [733.62, 164.92]]}]\n```", - "options": null, - "id": 129 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000351", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[507.72, 203.19], [541.44, 203.18], [532.78, 206.17], [493.43, 206.19], [507.72, 161.2], [541.44, 161.22], [532.78, 157.22], [493.43, 157.19]]}]\n```", - "options": null, - "id": 130 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000351", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[576.51, 215.19], [619.78, 215.17], [622.68, 219.7], [572.88, 219.72], [576.51, 175.84], [619.78, 175.85], [622.68, 174.43], [572.88, 174.43]]}, {\"category\": \"car\", \"corners_3d\": [[208.79, 254.05], [275.9, 254.05], [186.15, 272.67], [100.9, 272.66], [208.79, 189.96], [275.9, 189.96], [186.15, 191.25], [100.9, 191.25]]}, {\"category\": \"car\", \"corners_3d\": [[385.62, 245.13], [457.96, 245.13], [410.59, 264.08], [315.39, 264.07], [385.62, 173.98], [457.96, 173.98], [410.59, 170.42], [315.39, 170.43]]}, {\"category\": \"car\", \"corners_3d\": [[376.97, 218.02], [416.82, 218.01], [385.0, 223.69], [338.23, 223.71], [376.97, 182.23], [416.82, 182.23], [385.0, 181.71], [338.23, 181.71]]}, {\"category\": \"car\", \"corners_3d\": [[468.99, 216.07], [508.41, 216.05], [494.98, 220.55], [449.81, 220.57], [468.99, 181.89], [508.41, 181.89], [494.98, 181.41], [449.81, 181.4]]}, {\"category\": \"car\", \"corners_3d\": [[531.76, 199.36], [554.31, 199.35], [551.22, 200.48], [526.87, 200.49], [531.76, 177.38], [554.31, 177.38], [551.22, 176.76], [526.87, 176.76]]}, {\"category\": \"car\", \"corners_3d\": [[288.75, 218.35], [326.65, 218.33], [286.88, 223.14], [243.45, 223.16], [288.75, 181.97], [326.65, 181.97], [286.88, 181.5], [243.45, 181.49]]}]\n```", - "options": null, - "id": 131 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000351", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[576.51, 215.19], [619.78, 215.17], [622.68, 219.7], [572.88, 219.72], [576.51, 175.84], [619.78, 175.85], [622.68, 174.43], [572.88, 174.43]]}, {\"category\": \"car\", \"corners_3d\": [[208.79, 254.05], [275.9, 254.05], [186.15, 272.67], [100.9, 272.66], [208.79, 189.96], [275.9, 189.96], [186.15, 191.25], [100.9, 191.25]]}, {\"category\": \"car\", \"corners_3d\": [[385.62, 245.13], [457.96, 245.13], [410.59, 264.08], [315.39, 264.07], [385.62, 173.98], [457.96, 173.98], [410.59, 170.42], [315.39, 170.43]]}, {\"category\": \"car\", \"corners_3d\": [[376.97, 218.02], [416.82, 218.01], [385.0, 223.69], [338.23, 223.71], [376.97, 182.23], [416.82, 182.23], [385.0, 181.71], [338.23, 181.71]]}, {\"category\": \"car\", \"corners_3d\": [[468.99, 216.07], [508.41, 216.05], [494.98, 220.55], [449.81, 220.57], [468.99, 181.89], [508.41, 181.89], [494.98, 181.41], [449.81, 181.4]]}, {\"category\": \"van\", \"corners_3d\": [[507.72, 203.19], [541.44, 203.18], [532.78, 206.17], [493.43, 206.19], [507.72, 161.2], [541.44, 161.22], [532.78, 157.22], [493.43, 157.19]]}, {\"category\": \"car\", \"corners_3d\": [[531.76, 199.36], [554.31, 199.35], [551.22, 200.48], [526.87, 200.49], [531.76, 177.38], [554.31, 177.38], [551.22, 176.76], [526.87, 176.76]]}, {\"category\": \"car\", \"corners_3d\": [[288.75, 218.35], [326.65, 218.33], [286.88, 223.14], [243.45, 223.16], [288.75, 181.97], [326.65, 181.97], [286.88, 181.5], [243.45, 181.49]]}]\n```", - "options": null, - "id": 132 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000358", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[315.44, 206.19], [289.13, 206.12], [318.63, 203.42], [342.86, 203.48], [315.44, 180.11], [289.13, 180.1], [318.63, 179.51], [342.86, 179.52]]}, {\"category\": \"car\", \"corners_3d\": [[350.27, 201.5], [324.11, 201.46], [348.56, 199.23], [372.72, 199.26], [350.27, 180.49], [324.11, 180.48], [348.56, 179.88], [372.72, 179.89]]}, {\"category\": \"car\", \"corners_3d\": [[388.29, 196.45], [367.54, 196.4], [388.13, 194.74], [407.46, 194.78], [388.29, 177.7], [367.54, 177.69], [388.13, 177.35], [407.46, 177.35]]}]\n```", - "options": null, - "id": 133 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000358", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[640.64, 197.76], [683.39, 198.0], [671.2, 205.09], [616.69, 204.7], [640.64, 145.87], [683.39, 145.62], [671.2, 137.93], [616.69, 138.35]]}]\n```", - "options": null, - "id": 134 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000358", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[640.64, 197.76], [683.39, 198.0], [671.2, 205.09], [616.69, 204.7], [640.64, 145.87], [683.39, 145.62], [671.2, 137.93], [616.69, 138.35]]}, {\"category\": \"car\", \"corners_3d\": [[315.44, 206.19], [289.13, 206.12], [318.63, 203.42], [342.86, 203.48], [315.44, 180.11], [289.13, 180.1], [318.63, 179.51], [342.86, 179.52]]}, {\"category\": \"car\", \"corners_3d\": [[350.27, 201.5], [324.11, 201.46], [348.56, 199.23], [372.72, 199.26], [350.27, 180.49], [324.11, 180.48], [348.56, 179.88], [372.72, 179.89]]}, {\"category\": \"car\", \"corners_3d\": [[388.29, 196.45], [367.54, 196.4], [388.13, 194.74], [407.46, 194.78], [388.29, 177.7], [367.54, 177.69], [388.13, 177.35], [407.46, 177.35]]}]\n```", - "options": null, - "id": 135 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000359", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (bus). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[624.81, 188.81], [648.4, 188.78], [663.21, 191.66], [635.37, 191.71], [624.81, 163.12], [648.4, 163.13], [663.21, 161.37], [635.37, 161.34]]}]\n```", - "options": null, - "id": 136 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000359", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[343.21, 219.16], [301.81, 219.13], [338.8, 213.69], [375.35, 213.7], [343.21, 176.72], [301.81, 176.72], [338.8, 176.27], [375.35, 176.27]]}]\n```", - "options": null, - "id": 137 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000359", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[596.78, 199.04], [622.37, 199.04], [623.19, 200.78], [595.89, 200.78], [596.78, 174.48], [622.37, 174.48], [623.19, 174.58], [595.89, 174.58]]}]\n```", - "options": null, - "id": 138 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000359", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (bus, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[624.81, 188.81], [648.4, 188.78], [663.21, 191.66], [635.37, 191.71], [624.81, 163.12], [648.4, 163.13], [663.21, 161.37], [635.37, 161.34]]}, {\"category\": \"van\", \"corners_3d\": [[343.21, 219.16], [301.81, 219.13], [338.8, 213.69], [375.35, 213.7], [343.21, 176.72], [301.81, 176.72], [338.8, 176.27], [375.35, 176.27]]}]\n```", - "options": null, - "id": 139 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000359", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (bus, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[596.78, 199.04], [622.37, 199.04], [623.19, 200.78], [595.89, 200.78], [596.78, 174.48], [622.37, 174.48], [623.19, 174.58], [595.89, 174.58]]}, {\"category\": \"bus\", \"corners_3d\": [[624.81, 188.81], [648.4, 188.78], [663.21, 191.66], [635.37, 191.71], [624.81, 163.12], [648.4, 163.13], [663.21, 161.37], [635.37, 161.34]]}]\n```", - "options": null, - "id": 140 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000359", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[596.78, 199.04], [622.37, 199.04], [623.19, 200.78], [595.89, 200.78], [596.78, 174.48], [622.37, 174.48], [623.19, 174.58], [595.89, 174.58]]}, {\"category\": \"van\", \"corners_3d\": [[343.21, 219.16], [301.81, 219.13], [338.8, 213.69], [375.35, 213.7], [343.21, 176.72], [301.81, 176.72], [338.8, 176.27], [375.35, 176.27]]}]\n```", - "options": null, - "id": 141 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000359", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (bus, van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[596.78, 199.04], [622.37, 199.04], [623.19, 200.78], [595.89, 200.78], [596.78, 174.48], [622.37, 174.48], [623.19, 174.58], [595.89, 174.58]]}, {\"category\": \"bus\", \"corners_3d\": [[624.81, 188.81], [648.4, 188.78], [663.21, 191.66], [635.37, 191.71], [624.81, 163.12], [648.4, 163.13], [663.21, 161.37], [635.37, 161.34]]}, {\"category\": \"van\", \"corners_3d\": [[343.21, 219.16], [301.81, 219.13], [338.8, 213.69], [375.35, 213.7], [343.21, 176.72], [301.81, 176.72], [338.8, 176.27], [375.35, 176.27]]}]\n```", - "options": null, - "id": 142 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000402", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[508.46, 226.56], [457.3, 226.53], [480.1, 218.83], [523.93, 218.85], [508.46, 173.46], [457.3, 173.46], [480.1, 173.37], [523.93, 173.37]]}, {\"category\": \"van\", \"corners_3d\": [[563.09, 194.1], [537.47, 194.12], [541.43, 192.43], [565.02, 192.41], [563.09, 166.23], [537.47, 166.22], [541.43, 166.75], [565.02, 166.75]]}]\n```", - "options": null, - "id": 143 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000402", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[467.07, 262.39], [383.46, 261.98], [441.91, 241.5], [506.45, 241.75], [467.07, 193.17], [383.46, 193.08], [441.91, 188.43], [506.45, 188.48]]}, {\"category\": \"car\", \"corners_3d\": [[553.87, 199.17], [527.49, 199.19], [533.18, 196.96], [557.32, 196.94], [553.87, 174.45], [527.49, 174.45], [533.18, 174.31], [557.32, 174.31]]}]\n```", - "options": null, - "id": 144 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000402", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[467.07, 262.39], [383.46, 261.98], [441.91, 241.5], [506.45, 241.75], [467.07, 193.17], [383.46, 193.08], [441.91, 188.43], [506.45, 188.48]]}, {\"category\": \"van\", \"corners_3d\": [[508.46, 226.56], [457.3, 226.53], [480.1, 218.83], [523.93, 218.85], [508.46, 173.46], [457.3, 173.46], [480.1, 173.37], [523.93, 173.37]]}, {\"category\": \"car\", \"corners_3d\": [[553.87, 199.17], [527.49, 199.19], [533.18, 196.96], [557.32, 196.94], [553.87, 174.45], [527.49, 174.45], [533.18, 174.31], [557.32, 174.31]]}, {\"category\": \"van\", \"corners_3d\": [[563.09, 194.1], [537.47, 194.12], [541.43, 192.43], [565.02, 192.41], [563.09, 166.23], [537.47, 166.22], [541.43, 166.75], [565.02, 166.75]]}]\n```", - "options": null, - "id": 145 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000411", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[669.49, 192.19], [688.57, 192.12], [701.65, 193.52], [681.22, 193.6], [669.49, 173.76], [688.57, 173.76], [701.65, 173.82], [681.22, 173.83]]}, {\"category\": \"car\", \"corners_3d\": [[960.38, 352.96], [1118.44, 352.2], [1616.64, 523.13], [1310.85, 526.05], [960.38, 196.02], [1118.44, 195.93], [1616.64, 217.91], [1310.85, 218.29]]}]\n```", - "options": null, - "id": 146 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000412", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[832.15, 316.84], [987.88, 319.0], [1305.8, 461.02], [993.95, 452.75], [832.15, 167.25], [987.88, 167.16], [1305.8, 161.63], [993.95, 161.95]]}, {\"category\": \"car\", \"corners_3d\": [[528.06, 211.89], [490.9, 211.9], [504.08, 207.54], [537.09, 207.54], [528.06, 167.88], [490.9, 167.88], [504.08, 168.43], [537.09, 168.43]]}, {\"category\": \"car\", \"corners_3d\": [[226.52, 335.59], [96.32, 335.02], [262.89, 283.78], [352.24, 284.04], [226.52, 199.95], [96.32, 199.86], [262.89, 191.32], [352.24, 191.37]]}]\n```", - "options": null, - "id": 147 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000412", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[802.58, 229.21], [795.02, 230.18], [758.75, 228.12], [766.63, 227.22], [802.58, 155.85], [795.02, 155.56], [758.75, 156.18], [766.63, 156.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[808.92, 227.04], [799.57, 228.05], [770.07, 226.31], [779.65, 225.36], [808.92, 158.09], [799.57, 157.82], [770.07, 158.29], [779.65, 158.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[658.74, 199.3], [661.76, 199.66], [644.66, 199.81], [641.85, 199.46], [658.74, 166.37], [661.76, 166.29], [644.66, 166.25], [641.85, 166.33]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[383.62, 229.53], [365.46, 229.53], [369.79, 228.52], [387.63, 228.52], [383.62, 180.49], [365.46, 180.49], [369.79, 180.35], [387.63, 180.35]]}]\n```", - "options": null, - "id": 148 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000412", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[832.15, 316.84], [987.88, 319.0], [1305.8, 461.02], [993.95, 452.75], [832.15, 167.25], [987.88, 167.16], [1305.8, 161.63], [993.95, 161.95]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[802.58, 229.21], [795.02, 230.18], [758.75, 228.12], [766.63, 227.22], [802.58, 155.85], [795.02, 155.56], [758.75, 156.18], [766.63, 156.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[808.92, 227.04], [799.57, 228.05], [770.07, 226.31], [779.65, 225.36], [808.92, 158.09], [799.57, 157.82], [770.07, 158.29], [779.65, 158.55]]}, {\"category\": \"car\", \"corners_3d\": [[528.06, 211.89], [490.9, 211.9], [504.08, 207.54], [537.09, 207.54], [528.06, 167.88], [490.9, 167.88], [504.08, 168.43], [537.09, 168.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[658.74, 199.3], [661.76, 199.66], [644.66, 199.81], [641.85, 199.46], [658.74, 166.37], [661.76, 166.29], [644.66, 166.25], [641.85, 166.33]]}, {\"category\": \"car\", \"corners_3d\": [[226.52, 335.59], [96.32, 335.02], [262.89, 283.78], [352.24, 284.04], [226.52, 199.95], [96.32, 199.86], [262.89, 191.32], [352.24, 191.37]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[383.62, 229.53], [365.46, 229.53], [369.79, 228.52], [387.63, 228.52], [383.62, 180.49], [365.46, 180.49], [369.79, 180.35], [387.63, 180.35]]}]\n```", - "options": null, - "id": 149 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000416", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[434.78, 207.88], [401.48, 207.95], [420.63, 204.19], [450.33, 204.14], [434.78, 170.45], [401.48, 170.45], [420.63, 170.71], [450.33, 170.71]]}]\n```", - "options": null, - "id": 150 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000416", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[642.36, 195.71], [657.39, 195.64], [665.9, 196.94], [650.04, 197.02], [642.36, 179.04], [657.39, 179.03], [665.9, 179.38], [650.04, 179.4]]}]\n```", - "options": null, - "id": 151 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000416", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[642.36, 195.71], [657.39, 195.64], [665.9, 196.94], [650.04, 197.02], [642.36, 179.04], [657.39, 179.03], [665.9, 179.38], [650.04, 179.4]]}, {\"category\": \"van\", \"corners_3d\": [[434.78, 207.88], [401.48, 207.95], [420.63, 204.19], [450.33, 204.14], [434.78, 170.45], [401.48, 170.45], [420.63, 170.71], [450.33, 170.71]]}]\n```", - "options": null, - "id": 152 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000420", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[62.63, 525.52], [-269.64, 522.36], [188.69, 342.99], [351.7, 343.74], [62.63, 203.9], [-269.64, 203.63], [188.69, 187.83], [351.7, 187.9]]}, {\"category\": \"car\", \"corners_3d\": [[529.75, 231.85], [600.03, 232.44], [572.96, 252.0], [480.26, 250.96], [529.75, 168.15], [600.03, 168.1], [572.96, 166.54], [480.26, 166.63]]}, {\"category\": \"car\", \"corners_3d\": [[824.19, 247.76], [911.65, 251.01], [915.22, 272.09], [805.21, 266.91], [824.19, 170.74], [911.65, 170.65], [915.22, 170.05], [805.21, 170.2]]}, {\"category\": \"car\", \"corners_3d\": [[423.94, 252.49], [345.26, 252.5], [403.19, 235.0], [464.59, 235.0], [423.94, 181.8], [345.26, 181.8], [403.19, 179.83], [464.59, 179.83]]}, {\"category\": \"car\", \"corners_3d\": [[935.91, 221.25], [997.02, 224.5], [909.87, 233.42], [845.35, 229.0], [935.91, 163.62], [997.02, 163.0], [909.87, 161.3], [845.35, 162.14]]}]\n```", - "options": null, - "id": 153 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000422", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[289.37, 318.4], [136.27, 321.77], [262.24, 274.89], [362.97, 273.39], [289.37, 183.95], [136.27, 184.01], [262.24, 183.18], [362.97, 183.15]]}]\n```", - "options": null, - "id": 154 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000422", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[422.39, 279.68], [366.06, 280.81], [384.97, 265.81], [432.69, 265.0], [422.39, 155.98], [366.06, 155.69], [384.97, 159.59], [432.69, 159.8]]}]\n```", - "options": null, - "id": 155 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000422", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[61.84, 378.34], [165.4, 357.76], [220.99, 371.62], [115.31, 395.79], [61.84, 163.42], [165.4, 165.32], [220.99, 164.04], [115.31, 161.82]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[27.78, 368.91], [122.52, 357.68], [126.61, 378.85], [19.97, 393.05], [27.78, 172.19], [122.52, 172.75], [126.61, 171.7], [19.97, 170.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[678.65, 227.92], [643.65, 229.52], [625.7, 227.84], [660.09, 226.35], [678.65, 153.49], [643.65, 152.52], [625.7, 153.53], [660.09, 154.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[472.76, 213.27], [495.45, 213.14], [495.31, 214.17], [471.88, 214.31], [472.76, 167.04], [495.45, 167.1], [495.31, 166.63], [471.88, 166.57]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[488.22, 212.72], [512.22, 212.57], [512.93, 213.68], [488.07, 213.84], [488.22, 164.79], [512.22, 164.87], [512.93, 164.28], [488.07, 164.19]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[350.96, 217.22], [379.37, 216.89], [377.79, 218.36], [348.17, 218.72], [350.96, 164.62], [379.37, 164.78], [377.79, 164.09], [348.17, 163.92]]}]\n```", - "options": null, - "id": 156 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000422", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[289.37, 318.4], [136.27, 321.77], [262.24, 274.89], [362.97, 273.39], [289.37, 183.95], [136.27, 184.01], [262.24, 183.18], [362.97, 183.15]]}, {\"category\": \"cyclist\", \"corners_3d\": [[422.39, 279.68], [366.06, 280.81], [384.97, 265.81], [432.69, 265.0], [422.39, 155.98], [366.06, 155.69], [384.97, 159.59], [432.69, 159.8]]}]\n```", - "options": null, - "id": 157 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000422", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[61.84, 378.34], [165.4, 357.76], [220.99, 371.62], [115.31, 395.79], [61.84, 163.42], [165.4, 165.32], [220.99, 164.04], [115.31, 161.82]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[27.78, 368.91], [122.52, 357.68], [126.61, 378.85], [19.97, 393.05], [27.78, 172.19], [122.52, 172.75], [126.61, 171.7], [19.97, 170.99]]}, {\"category\": \"car\", \"corners_3d\": [[289.37, 318.4], [136.27, 321.77], [262.24, 274.89], [362.97, 273.39], [289.37, 183.95], [136.27, 184.01], [262.24, 183.18], [362.97, 183.15]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[678.65, 227.92], [643.65, 229.52], [625.7, 227.84], [660.09, 226.35], [678.65, 153.49], [643.65, 152.52], [625.7, 153.53], [660.09, 154.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[472.76, 213.27], [495.45, 213.14], [495.31, 214.17], [471.88, 214.31], [472.76, 167.04], [495.45, 167.1], [495.31, 166.63], [471.88, 166.57]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[488.22, 212.72], [512.22, 212.57], [512.93, 213.68], [488.07, 213.84], [488.22, 164.79], [512.22, 164.87], [512.93, 164.28], [488.07, 164.19]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[350.96, 217.22], [379.37, 216.89], [377.79, 218.36], [348.17, 218.72], [350.96, 164.62], [379.37, 164.78], [377.79, 164.09], [348.17, 163.92]]}]\n```", - "options": null, - "id": 158 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000422", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[61.84, 378.34], [165.4, 357.76], [220.99, 371.62], [115.31, 395.79], [61.84, 163.42], [165.4, 165.32], [220.99, 164.04], [115.31, 161.82]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[27.78, 368.91], [122.52, 357.68], [126.61, 378.85], [19.97, 393.05], [27.78, 172.19], [122.52, 172.75], [126.61, 171.7], [19.97, 170.99]]}, {\"category\": \"cyclist\", \"corners_3d\": [[422.39, 279.68], [366.06, 280.81], [384.97, 265.81], [432.69, 265.0], [422.39, 155.98], [366.06, 155.69], [384.97, 159.59], [432.69, 159.8]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[678.65, 227.92], [643.65, 229.52], [625.7, 227.84], [660.09, 226.35], [678.65, 153.49], [643.65, 152.52], [625.7, 153.53], [660.09, 154.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[472.76, 213.27], [495.45, 213.14], [495.31, 214.17], [471.88, 214.31], [472.76, 167.04], [495.45, 167.1], [495.31, 166.63], [471.88, 166.57]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[488.22, 212.72], [512.22, 212.57], [512.93, 213.68], [488.07, 213.84], [488.22, 164.79], [512.22, 164.87], [512.93, 164.28], [488.07, 164.19]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[350.96, 217.22], [379.37, 216.89], [377.79, 218.36], [348.17, 218.72], [350.96, 164.62], [379.37, 164.78], [377.79, 164.09], [348.17, 163.92]]}]\n```", - "options": null, - "id": 159 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000422", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[61.84, 378.34], [165.4, 357.76], [220.99, 371.62], [115.31, 395.79], [61.84, 163.42], [165.4, 165.32], [220.99, 164.04], [115.31, 161.82]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[27.78, 368.91], [122.52, 357.68], [126.61, 378.85], [19.97, 393.05], [27.78, 172.19], [122.52, 172.75], [126.61, 171.7], [19.97, 170.99]]}, {\"category\": \"car\", \"corners_3d\": [[289.37, 318.4], [136.27, 321.77], [262.24, 274.89], [362.97, 273.39], [289.37, 183.95], [136.27, 184.01], [262.24, 183.18], [362.97, 183.15]]}, {\"category\": \"cyclist\", \"corners_3d\": [[422.39, 279.68], [366.06, 280.81], [384.97, 265.81], [432.69, 265.0], [422.39, 155.98], [366.06, 155.69], [384.97, 159.59], [432.69, 159.8]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[678.65, 227.92], [643.65, 229.52], [625.7, 227.84], [660.09, 226.35], [678.65, 153.49], [643.65, 152.52], [625.7, 153.53], [660.09, 154.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[472.76, 213.27], [495.45, 213.14], [495.31, 214.17], [471.88, 214.31], [472.76, 167.04], [495.45, 167.1], [495.31, 166.63], [471.88, 166.57]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[488.22, 212.72], [512.22, 212.57], [512.93, 213.68], [488.07, 213.84], [488.22, 164.79], [512.22, 164.87], [512.93, 164.28], [488.07, 164.19]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[350.96, 217.22], [379.37, 216.89], [377.79, 218.36], [348.17, 218.72], [350.96, 164.62], [379.37, 164.78], [377.79, 164.09], [348.17, 163.92]]}]\n```", - "options": null, - "id": 160 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000451", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[444.21, 230.76], [397.65, 230.92], [427.36, 221.46], [466.29, 221.36], [444.21, 164.85], [397.65, 164.83], [427.36, 166.13], [466.29, 166.15]]}]\n```", - "options": null, - "id": 161 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000451", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[578.96, 200.35], [594.51, 200.35], [593.72, 201.74], [577.39, 201.74], [578.96, 188.11], [594.51, 188.11], [593.72, 188.88], [577.39, 188.88]]}]\n```", - "options": null, - "id": 162 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000451", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[444.21, 230.76], [397.65, 230.92], [427.36, 221.46], [466.29, 221.36], [444.21, 164.85], [397.65, 164.83], [427.36, 166.13], [466.29, 166.15]]}, {\"category\": \"car\", \"corners_3d\": [[578.96, 200.35], [594.51, 200.35], [593.72, 201.74], [577.39, 201.74], [578.96, 188.11], [594.51, 188.11], [593.72, 188.88], [577.39, 188.88]]}]\n```", - "options": null, - "id": 163 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000454", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (bus). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[593.03, 189.8], [616.13, 189.8], [618.46, 192.79], [591.28, 192.8], [593.03, 164.77], [616.13, 164.77], [618.46, 163.34], [591.28, 163.34]]}]\n```", - "options": null, - "id": 164 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000454", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[873.06, 210.28], [878.77, 208.77], [960.41, 210.66], [958.0, 212.33], [873.06, 179.01], [878.77, 178.76], [960.41, 179.07], [958.0, 179.35]]}, {\"category\": \"car\", \"corners_3d\": [[278.31, 213.87], [247.61, 213.85], [283.7, 209.83], [311.41, 209.85], [278.31, 190.36], [247.61, 190.35], [283.7, 188.64], [311.41, 188.64]]}]\n```", - "options": null, - "id": 165 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000454", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (bus, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[593.03, 189.8], [616.13, 189.8], [618.46, 192.79], [591.28, 192.8], [593.03, 164.77], [616.13, 164.77], [618.46, 163.34], [591.28, 163.34]]}, {\"category\": \"car\", \"corners_3d\": [[873.06, 210.28], [878.77, 208.77], [960.41, 210.66], [958.0, 212.33], [873.06, 179.01], [878.77, 178.76], [960.41, 179.07], [958.0, 179.35]]}, {\"category\": \"car\", \"corners_3d\": [[278.31, 213.87], [247.61, 213.85], [283.7, 209.83], [311.41, 209.85], [278.31, 190.36], [247.61, 190.35], [283.7, 188.64], [311.41, 188.64]]}]\n```", - "options": null, - "id": 166 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000464", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[348.85, 312.72], [225.66, 310.77], [364.49, 267.37], [449.87, 268.28], [348.85, 213.42], [225.66, 212.85], [364.49, 200.27], [449.87, 200.53]]}, {\"category\": \"car\", \"corners_3d\": [[1184.97, 222.31], [1141.37, 219.32], [1244.21, 217.73], [1292.59, 220.51], [1184.97, 175.97], [1141.37, 175.79], [1244.21, 175.68], [1292.59, 175.86]]}]\n```", - "options": null, - "id": 167 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000464", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-2627.64, 1788.37], [-5401.77, 2085.35], [29.79, 344.59], [231.7, 341.8], [-2627.64, -475.64], [-5401.77, -594.85], [29.79, 103.92], [231.7, 105.04]]}]\n```", - "options": null, - "id": 168 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000464", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[660.52, 237.24], [666.24, 235.74], [715.07, 237.45], [710.38, 239.02], [660.52, 187.37], [666.24, 187.04], [715.07, 187.42], [710.38, 187.78]]}]\n```", - "options": null, - "id": 169 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000464", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[434.45, 279.73], [440.47, 274.81], [488.32, 274.4], [484.63, 279.28], [434.45, 185.01], [440.47, 184.45], [488.32, 184.4], [484.63, 184.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[530.96, 260.91], [517.64, 263.44], [490.47, 261.48], [504.25, 259.06], [530.96, 187.38], [517.64, 187.8], [490.47, 187.47], [504.25, 187.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[544.15, 259.73], [521.93, 261.71], [504.32, 258.49], [526.12, 256.65], [544.15, 184.56], [521.93, 184.83], [504.32, 184.39], [526.12, 184.14]]}]\n```", - "options": null, - "id": 170 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000464", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-2627.64, 1788.37], [-5401.77, 2085.35], [29.79, 344.59], [231.7, 341.8], [-2627.64, -475.64], [-5401.77, -594.85], [29.79, 103.92], [231.7, 105.04]]}, {\"category\": \"car\", \"corners_3d\": [[348.85, 312.72], [225.66, 310.77], [364.49, 267.37], [449.87, 268.28], [348.85, 213.42], [225.66, 212.85], [364.49, 200.27], [449.87, 200.53]]}, {\"category\": \"car\", \"corners_3d\": [[1184.97, 222.31], [1141.37, 219.32], [1244.21, 217.73], [1292.59, 220.51], [1184.97, 175.97], [1141.37, 175.79], [1244.21, 175.68], [1292.59, 175.86]]}]\n```", - "options": null, - "id": 171 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000464", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[348.85, 312.72], [225.66, 310.77], [364.49, 267.37], [449.87, 268.28], [348.85, 213.42], [225.66, 212.85], [364.49, 200.27], [449.87, 200.53]]}, {\"category\": \"cyclist\", \"corners_3d\": [[660.52, 237.24], [666.24, 235.74], [715.07, 237.45], [710.38, 239.02], [660.52, 187.37], [666.24, 187.04], [715.07, 187.42], [710.38, 187.78]]}, {\"category\": \"car\", \"corners_3d\": [[1184.97, 222.31], [1141.37, 219.32], [1244.21, 217.73], [1292.59, 220.51], [1184.97, 175.97], [1141.37, 175.79], [1244.21, 175.68], [1292.59, 175.86]]}]\n```", - "options": null, - "id": 172 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000464", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[348.85, 312.72], [225.66, 310.77], [364.49, 267.37], [449.87, 268.28], [348.85, 213.42], [225.66, 212.85], [364.49, 200.27], [449.87, 200.53]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[434.45, 279.73], [440.47, 274.81], [488.32, 274.4], [484.63, 279.28], [434.45, 185.01], [440.47, 184.45], [488.32, 184.4], [484.63, 184.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[530.96, 260.91], [517.64, 263.44], [490.47, 261.48], [504.25, 259.06], [530.96, 187.38], [517.64, 187.8], [490.47, 187.47], [504.25, 187.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[544.15, 259.73], [521.93, 261.71], [504.32, 258.49], [526.12, 256.65], [544.15, 184.56], [521.93, 184.83], [504.32, 184.39], [526.12, 184.14]]}, {\"category\": \"car\", \"corners_3d\": [[1184.97, 222.31], [1141.37, 219.32], [1244.21, 217.73], [1292.59, 220.51], [1184.97, 175.97], [1141.37, 175.79], [1244.21, 175.68], [1292.59, 175.86]]}]\n```", - "options": null, - "id": 173 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000464", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-2627.64, 1788.37], [-5401.77, 2085.35], [29.79, 344.59], [231.7, 341.8], [-2627.64, -475.64], [-5401.77, -594.85], [29.79, 103.92], [231.7, 105.04]]}, {\"category\": \"cyclist\", \"corners_3d\": [[660.52, 237.24], [666.24, 235.74], [715.07, 237.45], [710.38, 239.02], [660.52, 187.37], [666.24, 187.04], [715.07, 187.42], [710.38, 187.78]]}]\n```", - "options": null, - "id": 174 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000464", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-2627.64, 1788.37], [-5401.77, 2085.35], [29.79, 344.59], [231.7, 341.8], [-2627.64, -475.64], [-5401.77, -594.85], [29.79, 103.92], [231.7, 105.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[434.45, 279.73], [440.47, 274.81], [488.32, 274.4], [484.63, 279.28], [434.45, 185.01], [440.47, 184.45], [488.32, 184.4], [484.63, 184.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[530.96, 260.91], [517.64, 263.44], [490.47, 261.48], [504.25, 259.06], [530.96, 187.38], [517.64, 187.8], [490.47, 187.47], [504.25, 187.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[544.15, 259.73], [521.93, 261.71], [504.32, 258.49], [526.12, 256.65], [544.15, 184.56], [521.93, 184.83], [504.32, 184.39], [526.12, 184.14]]}]\n```", - "options": null, - "id": 175 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000464", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[434.45, 279.73], [440.47, 274.81], [488.32, 274.4], [484.63, 279.28], [434.45, 185.01], [440.47, 184.45], [488.32, 184.4], [484.63, 184.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[530.96, 260.91], [517.64, 263.44], [490.47, 261.48], [504.25, 259.06], [530.96, 187.38], [517.64, 187.8], [490.47, 187.47], [504.25, 187.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[544.15, 259.73], [521.93, 261.71], [504.32, 258.49], [526.12, 256.65], [544.15, 184.56], [521.93, 184.83], [504.32, 184.39], [526.12, 184.14]]}, {\"category\": \"cyclist\", \"corners_3d\": [[660.52, 237.24], [666.24, 235.74], [715.07, 237.45], [710.38, 239.02], [660.52, 187.37], [666.24, 187.04], [715.07, 187.42], [710.38, 187.78]]}]\n```", - "options": null, - "id": 176 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000464", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-2627.64, 1788.37], [-5401.77, 2085.35], [29.79, 344.59], [231.7, 341.8], [-2627.64, -475.64], [-5401.77, -594.85], [29.79, 103.92], [231.7, 105.04]]}, {\"category\": \"car\", \"corners_3d\": [[348.85, 312.72], [225.66, 310.77], [364.49, 267.37], [449.87, 268.28], [348.85, 213.42], [225.66, 212.85], [364.49, 200.27], [449.87, 200.53]]}, {\"category\": \"cyclist\", \"corners_3d\": [[660.52, 237.24], [666.24, 235.74], [715.07, 237.45], [710.38, 239.02], [660.52, 187.37], [666.24, 187.04], [715.07, 187.42], [710.38, 187.78]]}, {\"category\": \"car\", \"corners_3d\": [[1184.97, 222.31], [1141.37, 219.32], [1244.21, 217.73], [1292.59, 220.51], [1184.97, 175.97], [1141.37, 175.79], [1244.21, 175.68], [1292.59, 175.86]]}]\n```", - "options": null, - "id": 177 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000464", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-2627.64, 1788.37], [-5401.77, 2085.35], [29.79, 344.59], [231.7, 341.8], [-2627.64, -475.64], [-5401.77, -594.85], [29.79, 103.92], [231.7, 105.04]]}, {\"category\": \"car\", \"corners_3d\": [[348.85, 312.72], [225.66, 310.77], [364.49, 267.37], [449.87, 268.28], [348.85, 213.42], [225.66, 212.85], [364.49, 200.27], [449.87, 200.53]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[434.45, 279.73], [440.47, 274.81], [488.32, 274.4], [484.63, 279.28], [434.45, 185.01], [440.47, 184.45], [488.32, 184.4], [484.63, 184.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[530.96, 260.91], [517.64, 263.44], [490.47, 261.48], [504.25, 259.06], [530.96, 187.38], [517.64, 187.8], [490.47, 187.47], [504.25, 187.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[544.15, 259.73], [521.93, 261.71], [504.32, 258.49], [526.12, 256.65], [544.15, 184.56], [521.93, 184.83], [504.32, 184.39], [526.12, 184.14]]}, {\"category\": \"car\", \"corners_3d\": [[1184.97, 222.31], [1141.37, 219.32], [1244.21, 217.73], [1292.59, 220.51], [1184.97, 175.97], [1141.37, 175.79], [1244.21, 175.68], [1292.59, 175.86]]}]\n```", - "options": null, - "id": 178 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000464", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[348.85, 312.72], [225.66, 310.77], [364.49, 267.37], [449.87, 268.28], [348.85, 213.42], [225.66, 212.85], [364.49, 200.27], [449.87, 200.53]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[434.45, 279.73], [440.47, 274.81], [488.32, 274.4], [484.63, 279.28], [434.45, 185.01], [440.47, 184.45], [488.32, 184.4], [484.63, 184.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[530.96, 260.91], [517.64, 263.44], [490.47, 261.48], [504.25, 259.06], [530.96, 187.38], [517.64, 187.8], [490.47, 187.47], [504.25, 187.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[544.15, 259.73], [521.93, 261.71], [504.32, 258.49], [526.12, 256.65], [544.15, 184.56], [521.93, 184.83], [504.32, 184.39], [526.12, 184.14]]}, {\"category\": \"cyclist\", \"corners_3d\": [[660.52, 237.24], [666.24, 235.74], [715.07, 237.45], [710.38, 239.02], [660.52, 187.37], [666.24, 187.04], [715.07, 187.42], [710.38, 187.78]]}, {\"category\": \"car\", \"corners_3d\": [[1184.97, 222.31], [1141.37, 219.32], [1244.21, 217.73], [1292.59, 220.51], [1184.97, 175.97], [1141.37, 175.79], [1244.21, 175.68], [1292.59, 175.86]]}]\n```", - "options": null, - "id": 179 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000464", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, cyclist, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-2627.64, 1788.37], [-5401.77, 2085.35], [29.79, 344.59], [231.7, 341.8], [-2627.64, -475.64], [-5401.77, -594.85], [29.79, 103.92], [231.7, 105.04]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[434.45, 279.73], [440.47, 274.81], [488.32, 274.4], [484.63, 279.28], [434.45, 185.01], [440.47, 184.45], [488.32, 184.4], [484.63, 184.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[530.96, 260.91], [517.64, 263.44], [490.47, 261.48], [504.25, 259.06], [530.96, 187.38], [517.64, 187.8], [490.47, 187.47], [504.25, 187.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[544.15, 259.73], [521.93, 261.71], [504.32, 258.49], [526.12, 256.65], [544.15, 184.56], [521.93, 184.83], [504.32, 184.39], [526.12, 184.14]]}, {\"category\": \"cyclist\", \"corners_3d\": [[660.52, 237.24], [666.24, 235.74], [715.07, 237.45], [710.38, 239.02], [660.52, 187.37], [666.24, 187.04], [715.07, 187.42], [710.38, 187.78]]}]\n```", - "options": null, - "id": 180 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000464", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van, cyclist, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-2627.64, 1788.37], [-5401.77, 2085.35], [29.79, 344.59], [231.7, 341.8], [-2627.64, -475.64], [-5401.77, -594.85], [29.79, 103.92], [231.7, 105.04]]}, {\"category\": \"car\", \"corners_3d\": [[348.85, 312.72], [225.66, 310.77], [364.49, 267.37], [449.87, 268.28], [348.85, 213.42], [225.66, 212.85], [364.49, 200.27], [449.87, 200.53]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[434.45, 279.73], [440.47, 274.81], [488.32, 274.4], [484.63, 279.28], [434.45, 185.01], [440.47, 184.45], [488.32, 184.4], [484.63, 184.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[530.96, 260.91], [517.64, 263.44], [490.47, 261.48], [504.25, 259.06], [530.96, 187.38], [517.64, 187.8], [490.47, 187.47], [504.25, 187.07]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[544.15, 259.73], [521.93, 261.71], [504.32, 258.49], [526.12, 256.65], [544.15, 184.56], [521.93, 184.83], [504.32, 184.39], [526.12, 184.14]]}, {\"category\": \"cyclist\", \"corners_3d\": [[660.52, 237.24], [666.24, 235.74], [715.07, 237.45], [710.38, 239.02], [660.52, 187.37], [666.24, 187.04], [715.07, 187.42], [710.38, 187.78]]}, {\"category\": \"car\", \"corners_3d\": [[1184.97, 222.31], [1141.37, 219.32], [1244.21, 217.73], [1292.59, 220.51], [1184.97, 175.97], [1141.37, 175.79], [1244.21, 175.68], [1292.59, 175.86]]}]\n```", - "options": null, - "id": 181 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000477", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[181.06, 290.96], [148.15, 298.32], [59.74, 296.74], [97.69, 289.57], [181.06, 161.57], [148.15, 160.31], [59.74, 160.58], [97.69, 161.81]]}]\n```", - "options": null, - "id": 182 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000540", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[709.3, 255.77], [811.39, 255.55], [945.66, 307.02], [780.33, 307.62], [709.3, 133.39], [811.39, 133.5], [945.66, 109.0], [780.33, 108.72]]}, {\"category\": \"van\", \"corners_3d\": [[261.74, 249.37], [179.77, 249.29], [272.43, 233.06], [337.06, 233.11], [261.74, 161.06], [179.77, 161.08], [272.43, 163.58], [337.06, 163.57]]}]\n```", - "options": null, - "id": 183 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000540", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[585.26, 219.93], [635.28, 219.9], [640.49, 227.47], [582.41, 227.51], [585.26, 179.65], [635.28, 179.65], [640.49, 180.74], [582.41, 180.75]]}, {\"category\": \"car\", \"corners_3d\": [[240.7, 253.21], [309.51, 246.89], [423.91, 258.51], [354.67, 267.09], [240.7, 185.45], [309.51, 184.46], [423.91, 186.29], [354.67, 187.63]]}, {\"category\": \"car\", \"corners_3d\": [[319.6, 240.5], [373.2, 236.24], [464.13, 244.59], [409.9, 250.09], [319.6, 187.31], [373.2, 186.4], [464.13, 188.19], [409.9, 189.36]]}, {\"category\": \"car\", \"corners_3d\": [[482.83, 222.66], [443.64, 224.98], [391.38, 220.15], [429.2, 218.23], [482.83, 180.04], [443.64, 180.38], [391.38, 179.68], [429.2, 179.4]]}, {\"category\": \"car\", \"corners_3d\": [[515.57, 221.0], [565.15, 221.0], [557.54, 229.13], [499.59, 229.13], [515.57, 185.38], [565.15, 185.38], [557.54, 187.5], [499.59, 187.5]]}, {\"category\": \"car\", \"corners_3d\": [[676.42, 229.42], [734.95, 229.38], [765.21, 242.33], [693.28, 242.4], [676.42, 179.66], [734.95, 179.65], [765.21, 181.21], [693.28, 181.22]]}, {\"category\": \"car\", \"corners_3d\": [[663.29, 219.61], [710.34, 219.61], [728.71, 228.19], [673.03, 228.19], [663.29, 183.08], [710.34, 183.08], [728.71, 184.96], [673.03, 184.96]]}, {\"category\": \"car\", \"corners_3d\": [[657.87, 210.09], [691.51, 210.02], [704.59, 214.43], [666.98, 214.51], [657.87, 174.45], [691.51, 174.44], [704.59, 174.63], [666.98, 174.63]]}, {\"category\": \"car\", \"corners_3d\": [[632.9, 197.89], [655.29, 197.87], [660.48, 199.81], [636.36, 199.84], [632.9, 177.91], [655.29, 177.91], [660.48, 178.3], [636.36, 178.31]]}, {\"category\": \"car\", \"corners_3d\": [[137.88, 273.79], [48.45, 273.92], [164.0, 252.82], [234.64, 252.73], [137.88, 198.49], [48.45, 198.52], [164.0, 193.16], [234.64, 193.14]]}, {\"category\": \"car\", \"corners_3d\": [[592.07, 208.37], [626.76, 208.36], [629.52, 212.47], [590.81, 212.49], [592.07, 176.79], [626.76, 176.79], [629.52, 177.25], [590.81, 177.25]]}, {\"category\": \"car\", \"corners_3d\": [[100.19, 235.19], [46.43, 235.38], [121.86, 226.54], [167.84, 226.4], [100.19, 187.06], [46.43, 187.1], [121.86, 185.09], [167.84, 185.06]]}]\n```", - "options": null, - "id": 184 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000540", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[585.26, 219.93], [635.28, 219.9], [640.49, 227.47], [582.41, 227.51], [585.26, 179.65], [635.28, 179.65], [640.49, 180.74], [582.41, 180.75]]}, {\"category\": \"car\", \"corners_3d\": [[240.7, 253.21], [309.51, 246.89], [423.91, 258.51], [354.67, 267.09], [240.7, 185.45], [309.51, 184.46], [423.91, 186.29], [354.67, 187.63]]}, {\"category\": \"car\", \"corners_3d\": [[319.6, 240.5], [373.2, 236.24], [464.13, 244.59], [409.9, 250.09], [319.6, 187.31], [373.2, 186.4], [464.13, 188.19], [409.9, 189.36]]}, {\"category\": \"car\", \"corners_3d\": [[482.83, 222.66], [443.64, 224.98], [391.38, 220.15], [429.2, 218.23], [482.83, 180.04], [443.64, 180.38], [391.38, 179.68], [429.2, 179.4]]}, {\"category\": \"car\", \"corners_3d\": [[515.57, 221.0], [565.15, 221.0], [557.54, 229.13], [499.59, 229.13], [515.57, 185.38], [565.15, 185.38], [557.54, 187.5], [499.59, 187.5]]}, {\"category\": \"van\", \"corners_3d\": [[709.3, 255.77], [811.39, 255.55], [945.66, 307.02], [780.33, 307.62], [709.3, 133.39], [811.39, 133.5], [945.66, 109.0], [780.33, 108.72]]}, {\"category\": \"car\", \"corners_3d\": [[676.42, 229.42], [734.95, 229.38], [765.21, 242.33], [693.28, 242.4], [676.42, 179.66], [734.95, 179.65], [765.21, 181.21], [693.28, 181.22]]}, {\"category\": \"car\", \"corners_3d\": [[663.29, 219.61], [710.34, 219.61], [728.71, 228.19], [673.03, 228.19], [663.29, 183.08], [710.34, 183.08], [728.71, 184.96], [673.03, 184.96]]}, {\"category\": \"car\", \"corners_3d\": [[657.87, 210.09], [691.51, 210.02], [704.59, 214.43], [666.98, 214.51], [657.87, 174.45], [691.51, 174.44], [704.59, 174.63], [666.98, 174.63]]}, {\"category\": \"car\", \"corners_3d\": [[632.9, 197.89], [655.29, 197.87], [660.48, 199.81], [636.36, 199.84], [632.9, 177.91], [655.29, 177.91], [660.48, 178.3], [636.36, 178.31]]}, {\"category\": \"car\", \"corners_3d\": [[137.88, 273.79], [48.45, 273.92], [164.0, 252.82], [234.64, 252.73], [137.88, 198.49], [48.45, 198.52], [164.0, 193.16], [234.64, 193.14]]}, {\"category\": \"van\", \"corners_3d\": [[261.74, 249.37], [179.77, 249.29], [272.43, 233.06], [337.06, 233.11], [261.74, 161.06], [179.77, 161.08], [272.43, 163.58], [337.06, 163.57]]}, {\"category\": \"car\", \"corners_3d\": [[592.07, 208.37], [626.76, 208.36], [629.52, 212.47], [590.81, 212.49], [592.07, 176.79], [626.76, 176.79], [629.52, 177.25], [590.81, 177.25]]}, {\"category\": \"car\", \"corners_3d\": [[100.19, 235.19], [46.43, 235.38], [121.86, 226.54], [167.84, 226.4], [100.19, 187.06], [46.43, 187.1], [121.86, 185.09], [167.84, 185.06]]}]\n```", - "options": null, - "id": 185 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000577", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[473.99, 311.97], [355.56, 310.6], [438.8, 271.97], [524.37, 272.68], [473.99, 190.02], [355.56, 189.86], [438.8, 185.09], [524.37, 185.17]]}, {\"category\": \"car\", \"corners_3d\": [[555.46, 265.2], [473.26, 264.79], [510.31, 244.09], [574.08, 244.34], [555.46, 169.84], [473.26, 169.85], [510.31, 170.53], [574.08, 170.52]]}, {\"category\": \"car\", \"corners_3d\": [[598.26, 240.48], [539.17, 239.98], [562.37, 228.44], [611.38, 228.78], [598.26, 175.67], [539.17, 175.65], [562.37, 175.17], [611.38, 175.19]]}, {\"category\": \"car\", \"corners_3d\": [[652.01, 215.94], [620.32, 215.41], [639.24, 211.3], [668.05, 211.74], [652.01, 185.53], [620.32, 185.38], [639.24, 184.17], [668.05, 184.3]]}]\n```", - "options": null, - "id": 186 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000619", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[530.17, 211.45], [487.08, 211.43], [502.94, 206.69], [540.74, 206.71], [530.17, 171.99], [487.08, 171.99], [502.94, 172.1], [540.74, 172.1]]}, {\"category\": \"car\", \"corners_3d\": [[287.98, 376.19], [91.5, 377.33], [273.82, 303.22], [398.7, 302.76], [287.98, 198.39], [91.5, 198.53], [273.82, 189.22], [398.7, 189.16]]}, {\"category\": \"car\", \"corners_3d\": [[184.95, 228.95], [188.85, 225.08], [354.74, 220.36], [362.32, 223.54], [184.95, 169.33], [188.85, 169.58], [354.74, 169.87], [362.32, 169.67]]}, {\"category\": \"car\", \"corners_3d\": [[466.38, 188.29], [462.33, 187.89], [521.59, 187.42], [527.04, 187.8], [466.38, 165.72], [462.33, 165.91], [521.59, 166.12], [527.04, 165.95]]}]\n```", - "options": null, - "id": 187 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000641", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[595.18, 211.43], [635.96, 211.83], [617.14, 218.78], [569.41, 218.23], [595.18, 169.79], [635.96, 169.76], [617.14, 169.2], [569.41, 169.25]]}, {\"category\": \"car\", \"corners_3d\": [[190.43, 266.49], [233.26, 262.05], [286.04, 268.8], [242.62, 273.95], [190.43, 212.98], [233.26, 211.08], [286.04, 213.97], [242.62, 216.17]]}, {\"category\": \"car\", \"corners_3d\": [[38.05, 284.22], [116.13, 275.82], [176.56, 285.72], [95.73, 295.89], [38.05, 217.02], [116.13, 213.69], [176.56, 217.62], [95.73, 221.65]]}, {\"category\": \"car\", \"corners_3d\": [[1199.84, 234.28], [1269.54, 242.44], [1023.61, 240.46], [983.76, 232.74], [1199.84, 157.21], [1269.54, 155.14], [1023.61, 155.64], [983.76, 157.61]]}, {\"category\": \"car\", \"corners_3d\": [[1052.99, 223.32], [1089.95, 228.24], [877.64, 226.86], [859.99, 222.16], [1052.99, 160.96], [1089.95, 159.8], [877.64, 160.13], [859.99, 161.23]]}, {\"category\": \"car\", \"corners_3d\": [[653.59, 192.44], [621.0, 192.3], [630.91, 190.45], [660.45, 190.57], [653.59, 161.09], [621.0, 161.18], [630.91, 162.28], [660.45, 162.21]]}, {\"category\": \"car\", \"corners_3d\": [[685.48, 179.71], [667.12, 179.68], [670.52, 179.28], [687.82, 179.31], [685.48, 164.0], [667.12, 164.04], [670.52, 164.55], [687.82, 164.52]]}]\n```", - "options": null, - "id": 188 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000666", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[174.24, 269.6], [224.91, 258.73], [400.82, 259.18], [372.25, 270.18], [174.24, 169.05], [224.91, 169.48], [400.82, 169.46], [372.25, 169.03]]}, {\"category\": \"car\", \"corners_3d\": [[270.51, 241.49], [300.37, 235.6], [449.34, 235.75], [433.44, 241.68], [270.51, 174.04], [300.37, 173.94], [449.34, 173.94], [433.44, 174.05]]}, {\"category\": \"car\", \"corners_3d\": [[821.52, 196.43], [863.25, 196.49], [899.28, 200.29], [850.77, 200.21], [821.52, 154.22], [863.25, 154.18], [899.28, 151.18], [850.77, 151.24]]}, {\"category\": \"car\", \"corners_3d\": [[878.97, 192.97], [911.28, 192.99], [946.9, 195.49], [910.54, 195.47], [878.97, 162.37], [911.28, 162.36], [946.9, 161.06], [910.54, 161.07]]}, {\"category\": \"car\", \"corners_3d\": [[343.79, 224.87], [360.59, 221.57], [481.56, 221.56], [472.97, 224.86], [343.79, 178.48], [360.59, 178.13], [481.56, 178.13], [472.97, 178.48]]}, {\"category\": \"car\", \"corners_3d\": [[372.34, 219.64], [384.88, 216.91], [478.27, 216.73], [471.56, 219.44], [372.34, 177.07], [384.88, 176.83], [478.27, 176.81], [471.56, 177.05]]}, {\"category\": \"car\", \"corners_3d\": [[382.54, 214.82], [395.95, 212.35], [506.68, 212.36], [500.18, 214.83], [382.54, 170.01], [395.95, 170.18], [506.68, 170.18], [500.18, 170.01]]}]\n```", - "options": null, - "id": 189 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000676", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[524.95, 212.45], [565.91, 212.5], [556.31, 217.76], [508.44, 217.68], [524.95, 174.5], [565.91, 174.49], [556.31, 173.3], [508.44, 173.32]]}, {\"category\": \"car\", \"corners_3d\": [[550.78, 199.56], [579.75, 199.57], [576.31, 201.32], [544.55, 201.3], [550.78, 173.01], [579.75, 173.01], [576.31, 172.18], [544.55, 172.19]]}, {\"category\": \"car\", \"corners_3d\": [[610.07, 197.59], [639.68, 197.63], [640.01, 199.58], [606.83, 199.54], [610.07, 170.69], [639.68, 170.67], [640.01, 169.35], [606.83, 169.38]]}, {\"category\": \"car\", \"corners_3d\": [[579.76, 191.44], [602.33, 191.46], [599.64, 192.23], [575.31, 192.21], [579.76, 169.27], [602.33, 169.25], [599.64, 168.29], [575.31, 168.31]]}]\n```", - "options": null, - "id": 190 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000685", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[511.91, 256.43], [553.03, 255.49], [565.31, 266.87], [518.61, 268.09], [511.91, 168.81], [553.03, 168.85], [565.31, 168.3], [518.61, 168.24]]}]\n```", - "options": null, - "id": 191 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000685", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[764.78, 230.79], [733.36, 231.36], [720.06, 228.56], [750.12, 228.05], [764.78, 160.33], [733.36, 160.21], [720.06, 160.81], [750.12, 160.92]]}]\n```", - "options": null, - "id": 192 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000685", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[452.66, 202.69], [484.14, 202.43], [487.28, 205.77], [452.24, 206.1], [452.66, 168.58], [484.14, 168.62], [487.28, 168.14], [452.24, 168.1]]}, {\"category\": \"van\", \"corners_3d\": [[294.6, 266.53], [205.84, 268.68], [262.77, 247.44], [331.21, 246.13], [294.6, 182.35], [205.84, 182.57], [262.77, 180.41], [331.21, 180.28]]}]\n```", - "options": null, - "id": 193 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000685", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[938.12, 309.68], [1121.5, 324.05], [1262.33, 471.21], [906.93, 420.0], [938.12, 174.61], [1121.5, 174.8], [1262.33, 176.69], [906.93, 176.03]]}, {\"category\": \"car\", \"corners_3d\": [[646.77, 269.14], [738.35, 266.65], [866.93, 310.59], [735.83, 316.03], [646.77, 179.58], [738.35, 179.41], [866.93, 182.47], [735.83, 182.86]]}, {\"category\": \"car\", \"corners_3d\": [[196.31, 312.57], [62.89, 316.18], [189.54, 274.65], [282.71, 272.81], [196.31, 204.52], [62.89, 205.34], [189.54, 195.93], [282.71, 195.51]]}, {\"category\": \"car\", \"corners_3d\": [[577.88, 235.05], [642.63, 233.99], [681.86, 248.0], [602.79, 249.6], [577.88, 172.45], [642.63, 172.46], [681.86, 172.36], [602.79, 172.35]]}, {\"category\": \"car\", \"corners_3d\": [[543.16, 222.61], [595.05, 221.76], [625.38, 230.94], [564.17, 232.15], [543.16, 175.92], [595.05, 175.86], [625.38, 176.43], [564.17, 176.5]]}, {\"category\": \"car\", \"corners_3d\": [[341.75, 245.64], [280.47, 247.14], [301.59, 235.65], [353.19, 234.57], [341.75, 194.75], [280.47, 195.2], [301.59, 191.74], [353.19, 191.42]]}, {\"category\": \"car\", \"corners_3d\": [[356.74, 227.15], [312.0, 227.91], [326.81, 221.14], [365.93, 220.55], [356.74, 187.81], [312.0, 188.02], [326.81, 186.15], [365.93, 185.99]]}, {\"category\": \"car\", \"corners_3d\": [[527.62, 212.46], [565.56, 212.01], [581.26, 217.27], [538.35, 217.86], [527.62, 177.67], [565.56, 177.62], [581.26, 178.26], [538.35, 178.33]]}]\n```", - "options": null, - "id": 194 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000685", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[511.91, 256.43], [553.03, 255.49], [565.31, 266.87], [518.61, 268.09], [511.91, 168.81], [553.03, 168.85], [565.31, 168.3], [518.61, 168.24]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[764.78, 230.79], [733.36, 231.36], [720.06, 228.56], [750.12, 228.05], [764.78, 160.33], [733.36, 160.21], [720.06, 160.81], [750.12, 160.92]]}]\n```", - "options": null, - "id": 195 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000685", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[452.66, 202.69], [484.14, 202.43], [487.28, 205.77], [452.24, 206.1], [452.66, 168.58], [484.14, 168.62], [487.28, 168.14], [452.24, 168.1]]}, {\"category\": \"cyclist\", \"corners_3d\": [[511.91, 256.43], [553.03, 255.49], [565.31, 266.87], [518.61, 268.09], [511.91, 168.81], [553.03, 168.85], [565.31, 168.3], [518.61, 168.24]]}, {\"category\": \"van\", \"corners_3d\": [[294.6, 266.53], [205.84, 268.68], [262.77, 247.44], [331.21, 246.13], [294.6, 182.35], [205.84, 182.57], [262.77, 180.41], [331.21, 180.28]]}]\n```", - "options": null, - "id": 196 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000685", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[511.91, 256.43], [553.03, 255.49], [565.31, 266.87], [518.61, 268.09], [511.91, 168.81], [553.03, 168.85], [565.31, 168.3], [518.61, 168.24]]}, {\"category\": \"car\", \"corners_3d\": [[938.12, 309.68], [1121.5, 324.05], [1262.33, 471.21], [906.93, 420.0], [938.12, 174.61], [1121.5, 174.8], [1262.33, 176.69], [906.93, 176.03]]}, {\"category\": \"car\", \"corners_3d\": [[646.77, 269.14], [738.35, 266.65], [866.93, 310.59], [735.83, 316.03], [646.77, 179.58], [738.35, 179.41], [866.93, 182.47], [735.83, 182.86]]}, {\"category\": \"car\", \"corners_3d\": [[196.31, 312.57], [62.89, 316.18], [189.54, 274.65], [282.71, 272.81], [196.31, 204.52], [62.89, 205.34], [189.54, 195.93], [282.71, 195.51]]}, {\"category\": \"car\", \"corners_3d\": [[577.88, 235.05], [642.63, 233.99], [681.86, 248.0], [602.79, 249.6], [577.88, 172.45], [642.63, 172.46], [681.86, 172.36], [602.79, 172.35]]}, {\"category\": \"car\", \"corners_3d\": [[543.16, 222.61], [595.05, 221.76], [625.38, 230.94], [564.17, 232.15], [543.16, 175.92], [595.05, 175.86], [625.38, 176.43], [564.17, 176.5]]}, {\"category\": \"car\", \"corners_3d\": [[341.75, 245.64], [280.47, 247.14], [301.59, 235.65], [353.19, 234.57], [341.75, 194.75], [280.47, 195.2], [301.59, 191.74], [353.19, 191.42]]}, {\"category\": \"car\", \"corners_3d\": [[356.74, 227.15], [312.0, 227.91], [326.81, 221.14], [365.93, 220.55], [356.74, 187.81], [312.0, 188.02], [326.81, 186.15], [365.93, 185.99]]}, {\"category\": \"car\", \"corners_3d\": [[527.62, 212.46], [565.56, 212.01], [581.26, 217.27], [538.35, 217.86], [527.62, 177.67], [565.56, 177.62], [581.26, 178.26], [538.35, 178.33]]}]\n```", - "options": null, - "id": 197 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000685", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[452.66, 202.69], [484.14, 202.43], [487.28, 205.77], [452.24, 206.1], [452.66, 168.58], [484.14, 168.62], [487.28, 168.14], [452.24, 168.1]]}, {\"category\": \"van\", \"corners_3d\": [[294.6, 266.53], [205.84, 268.68], [262.77, 247.44], [331.21, 246.13], [294.6, 182.35], [205.84, 182.57], [262.77, 180.41], [331.21, 180.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[764.78, 230.79], [733.36, 231.36], [720.06, 228.56], [750.12, 228.05], [764.78, 160.33], [733.36, 160.21], [720.06, 160.81], [750.12, 160.92]]}]\n```", - "options": null, - "id": 198 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000685", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[938.12, 309.68], [1121.5, 324.05], [1262.33, 471.21], [906.93, 420.0], [938.12, 174.61], [1121.5, 174.8], [1262.33, 176.69], [906.93, 176.03]]}, {\"category\": \"car\", \"corners_3d\": [[646.77, 269.14], [738.35, 266.65], [866.93, 310.59], [735.83, 316.03], [646.77, 179.58], [738.35, 179.41], [866.93, 182.47], [735.83, 182.86]]}, {\"category\": \"car\", \"corners_3d\": [[196.31, 312.57], [62.89, 316.18], [189.54, 274.65], [282.71, 272.81], [196.31, 204.52], [62.89, 205.34], [189.54, 195.93], [282.71, 195.51]]}, {\"category\": \"car\", \"corners_3d\": [[577.88, 235.05], [642.63, 233.99], [681.86, 248.0], [602.79, 249.6], [577.88, 172.45], [642.63, 172.46], [681.86, 172.36], [602.79, 172.35]]}, {\"category\": \"car\", \"corners_3d\": [[543.16, 222.61], [595.05, 221.76], [625.38, 230.94], [564.17, 232.15], [543.16, 175.92], [595.05, 175.86], [625.38, 176.43], [564.17, 176.5]]}, {\"category\": \"car\", \"corners_3d\": [[341.75, 245.64], [280.47, 247.14], [301.59, 235.65], [353.19, 234.57], [341.75, 194.75], [280.47, 195.2], [301.59, 191.74], [353.19, 191.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[764.78, 230.79], [733.36, 231.36], [720.06, 228.56], [750.12, 228.05], [764.78, 160.33], [733.36, 160.21], [720.06, 160.81], [750.12, 160.92]]}, {\"category\": \"car\", \"corners_3d\": [[356.74, 227.15], [312.0, 227.91], [326.81, 221.14], [365.93, 220.55], [356.74, 187.81], [312.0, 188.02], [326.81, 186.15], [365.93, 185.99]]}, {\"category\": \"car\", \"corners_3d\": [[527.62, 212.46], [565.56, 212.01], [581.26, 217.27], [538.35, 217.86], [527.62, 177.67], [565.56, 177.62], [581.26, 178.26], [538.35, 178.33]]}]\n```", - "options": null, - "id": 199 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000685", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[452.66, 202.69], [484.14, 202.43], [487.28, 205.77], [452.24, 206.1], [452.66, 168.58], [484.14, 168.62], [487.28, 168.14], [452.24, 168.1]]}, {\"category\": \"car\", \"corners_3d\": [[938.12, 309.68], [1121.5, 324.05], [1262.33, 471.21], [906.93, 420.0], [938.12, 174.61], [1121.5, 174.8], [1262.33, 176.69], [906.93, 176.03]]}, {\"category\": \"car\", \"corners_3d\": [[646.77, 269.14], [738.35, 266.65], [866.93, 310.59], [735.83, 316.03], [646.77, 179.58], [738.35, 179.41], [866.93, 182.47], [735.83, 182.86]]}, {\"category\": \"car\", \"corners_3d\": [[196.31, 312.57], [62.89, 316.18], [189.54, 274.65], [282.71, 272.81], [196.31, 204.52], [62.89, 205.34], [189.54, 195.93], [282.71, 195.51]]}, {\"category\": \"van\", \"corners_3d\": [[294.6, 266.53], [205.84, 268.68], [262.77, 247.44], [331.21, 246.13], [294.6, 182.35], [205.84, 182.57], [262.77, 180.41], [331.21, 180.28]]}, {\"category\": \"car\", \"corners_3d\": [[577.88, 235.05], [642.63, 233.99], [681.86, 248.0], [602.79, 249.6], [577.88, 172.45], [642.63, 172.46], [681.86, 172.36], [602.79, 172.35]]}, {\"category\": \"car\", \"corners_3d\": [[543.16, 222.61], [595.05, 221.76], [625.38, 230.94], [564.17, 232.15], [543.16, 175.92], [595.05, 175.86], [625.38, 176.43], [564.17, 176.5]]}, {\"category\": \"car\", \"corners_3d\": [[341.75, 245.64], [280.47, 247.14], [301.59, 235.65], [353.19, 234.57], [341.75, 194.75], [280.47, 195.2], [301.59, 191.74], [353.19, 191.42]]}, {\"category\": \"car\", \"corners_3d\": [[356.74, 227.15], [312.0, 227.91], [326.81, 221.14], [365.93, 220.55], [356.74, 187.81], [312.0, 188.02], [326.81, 186.15], [365.93, 185.99]]}, {\"category\": \"car\", \"corners_3d\": [[527.62, 212.46], [565.56, 212.01], [581.26, 217.27], [538.35, 217.86], [527.62, 177.67], [565.56, 177.62], [581.26, 178.26], [538.35, 178.33]]}]\n```", - "options": null, - "id": 200 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000685", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, pedestrian, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[452.66, 202.69], [484.14, 202.43], [487.28, 205.77], [452.24, 206.1], [452.66, 168.58], [484.14, 168.62], [487.28, 168.14], [452.24, 168.1]]}, {\"category\": \"cyclist\", \"corners_3d\": [[511.91, 256.43], [553.03, 255.49], [565.31, 266.87], [518.61, 268.09], [511.91, 168.81], [553.03, 168.85], [565.31, 168.3], [518.61, 168.24]]}, {\"category\": \"van\", \"corners_3d\": [[294.6, 266.53], [205.84, 268.68], [262.77, 247.44], [331.21, 246.13], [294.6, 182.35], [205.84, 182.57], [262.77, 180.41], [331.21, 180.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[764.78, 230.79], [733.36, 231.36], [720.06, 228.56], [750.12, 228.05], [764.78, 160.33], [733.36, 160.21], [720.06, 160.81], [750.12, 160.92]]}]\n```", - "options": null, - "id": 201 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000685", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, pedestrian, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[511.91, 256.43], [553.03, 255.49], [565.31, 266.87], [518.61, 268.09], [511.91, 168.81], [553.03, 168.85], [565.31, 168.3], [518.61, 168.24]]}, {\"category\": \"car\", \"corners_3d\": [[938.12, 309.68], [1121.5, 324.05], [1262.33, 471.21], [906.93, 420.0], [938.12, 174.61], [1121.5, 174.8], [1262.33, 176.69], [906.93, 176.03]]}, {\"category\": \"car\", \"corners_3d\": [[646.77, 269.14], [738.35, 266.65], [866.93, 310.59], [735.83, 316.03], [646.77, 179.58], [738.35, 179.41], [866.93, 182.47], [735.83, 182.86]]}, {\"category\": \"car\", \"corners_3d\": [[196.31, 312.57], [62.89, 316.18], [189.54, 274.65], [282.71, 272.81], [196.31, 204.52], [62.89, 205.34], [189.54, 195.93], [282.71, 195.51]]}, {\"category\": \"car\", \"corners_3d\": [[577.88, 235.05], [642.63, 233.99], [681.86, 248.0], [602.79, 249.6], [577.88, 172.45], [642.63, 172.46], [681.86, 172.36], [602.79, 172.35]]}, {\"category\": \"car\", \"corners_3d\": [[543.16, 222.61], [595.05, 221.76], [625.38, 230.94], [564.17, 232.15], [543.16, 175.92], [595.05, 175.86], [625.38, 176.43], [564.17, 176.5]]}, {\"category\": \"car\", \"corners_3d\": [[341.75, 245.64], [280.47, 247.14], [301.59, 235.65], [353.19, 234.57], [341.75, 194.75], [280.47, 195.2], [301.59, 191.74], [353.19, 191.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[764.78, 230.79], [733.36, 231.36], [720.06, 228.56], [750.12, 228.05], [764.78, 160.33], [733.36, 160.21], [720.06, 160.81], [750.12, 160.92]]}, {\"category\": \"car\", \"corners_3d\": [[356.74, 227.15], [312.0, 227.91], [326.81, 221.14], [365.93, 220.55], [356.74, 187.81], [312.0, 188.02], [326.81, 186.15], [365.93, 185.99]]}, {\"category\": \"car\", \"corners_3d\": [[527.62, 212.46], [565.56, 212.01], [581.26, 217.27], [538.35, 217.86], [527.62, 177.67], [565.56, 177.62], [581.26, 178.26], [538.35, 178.33]]}]\n```", - "options": null, - "id": 202 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000685", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[452.66, 202.69], [484.14, 202.43], [487.28, 205.77], [452.24, 206.1], [452.66, 168.58], [484.14, 168.62], [487.28, 168.14], [452.24, 168.1]]}, {\"category\": \"cyclist\", \"corners_3d\": [[511.91, 256.43], [553.03, 255.49], [565.31, 266.87], [518.61, 268.09], [511.91, 168.81], [553.03, 168.85], [565.31, 168.3], [518.61, 168.24]]}, {\"category\": \"car\", \"corners_3d\": [[938.12, 309.68], [1121.5, 324.05], [1262.33, 471.21], [906.93, 420.0], [938.12, 174.61], [1121.5, 174.8], [1262.33, 176.69], [906.93, 176.03]]}, {\"category\": \"car\", \"corners_3d\": [[646.77, 269.14], [738.35, 266.65], [866.93, 310.59], [735.83, 316.03], [646.77, 179.58], [738.35, 179.41], [866.93, 182.47], [735.83, 182.86]]}, {\"category\": \"car\", \"corners_3d\": [[196.31, 312.57], [62.89, 316.18], [189.54, 274.65], [282.71, 272.81], [196.31, 204.52], [62.89, 205.34], [189.54, 195.93], [282.71, 195.51]]}, {\"category\": \"van\", \"corners_3d\": [[294.6, 266.53], [205.84, 268.68], [262.77, 247.44], [331.21, 246.13], [294.6, 182.35], [205.84, 182.57], [262.77, 180.41], [331.21, 180.28]]}, {\"category\": \"car\", \"corners_3d\": [[577.88, 235.05], [642.63, 233.99], [681.86, 248.0], [602.79, 249.6], [577.88, 172.45], [642.63, 172.46], [681.86, 172.36], [602.79, 172.35]]}, {\"category\": \"car\", \"corners_3d\": [[543.16, 222.61], [595.05, 221.76], [625.38, 230.94], [564.17, 232.15], [543.16, 175.92], [595.05, 175.86], [625.38, 176.43], [564.17, 176.5]]}, {\"category\": \"car\", \"corners_3d\": [[341.75, 245.64], [280.47, 247.14], [301.59, 235.65], [353.19, 234.57], [341.75, 194.75], [280.47, 195.2], [301.59, 191.74], [353.19, 191.42]]}, {\"category\": \"car\", \"corners_3d\": [[356.74, 227.15], [312.0, 227.91], [326.81, 221.14], [365.93, 220.55], [356.74, 187.81], [312.0, 188.02], [326.81, 186.15], [365.93, 185.99]]}, {\"category\": \"car\", \"corners_3d\": [[527.62, 212.46], [565.56, 212.01], [581.26, 217.27], [538.35, 217.86], [527.62, 177.67], [565.56, 177.62], [581.26, 178.26], [538.35, 178.33]]}]\n```", - "options": null, - "id": 203 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000685", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[452.66, 202.69], [484.14, 202.43], [487.28, 205.77], [452.24, 206.1], [452.66, 168.58], [484.14, 168.62], [487.28, 168.14], [452.24, 168.1]]}, {\"category\": \"car\", \"corners_3d\": [[938.12, 309.68], [1121.5, 324.05], [1262.33, 471.21], [906.93, 420.0], [938.12, 174.61], [1121.5, 174.8], [1262.33, 176.69], [906.93, 176.03]]}, {\"category\": \"car\", \"corners_3d\": [[646.77, 269.14], [738.35, 266.65], [866.93, 310.59], [735.83, 316.03], [646.77, 179.58], [738.35, 179.41], [866.93, 182.47], [735.83, 182.86]]}, {\"category\": \"car\", \"corners_3d\": [[196.31, 312.57], [62.89, 316.18], [189.54, 274.65], [282.71, 272.81], [196.31, 204.52], [62.89, 205.34], [189.54, 195.93], [282.71, 195.51]]}, {\"category\": \"van\", \"corners_3d\": [[294.6, 266.53], [205.84, 268.68], [262.77, 247.44], [331.21, 246.13], [294.6, 182.35], [205.84, 182.57], [262.77, 180.41], [331.21, 180.28]]}, {\"category\": \"car\", \"corners_3d\": [[577.88, 235.05], [642.63, 233.99], [681.86, 248.0], [602.79, 249.6], [577.88, 172.45], [642.63, 172.46], [681.86, 172.36], [602.79, 172.35]]}, {\"category\": \"car\", \"corners_3d\": [[543.16, 222.61], [595.05, 221.76], [625.38, 230.94], [564.17, 232.15], [543.16, 175.92], [595.05, 175.86], [625.38, 176.43], [564.17, 176.5]]}, {\"category\": \"car\", \"corners_3d\": [[341.75, 245.64], [280.47, 247.14], [301.59, 235.65], [353.19, 234.57], [341.75, 194.75], [280.47, 195.2], [301.59, 191.74], [353.19, 191.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[764.78, 230.79], [733.36, 231.36], [720.06, 228.56], [750.12, 228.05], [764.78, 160.33], [733.36, 160.21], [720.06, 160.81], [750.12, 160.92]]}, {\"category\": \"car\", \"corners_3d\": [[356.74, 227.15], [312.0, 227.91], [326.81, 221.14], [365.93, 220.55], [356.74, 187.81], [312.0, 188.02], [326.81, 186.15], [365.93, 185.99]]}, {\"category\": \"car\", \"corners_3d\": [[527.62, 212.46], [565.56, 212.01], [581.26, 217.27], [538.35, 217.86], [527.62, 177.67], [565.56, 177.62], [581.26, 178.26], [538.35, 178.33]]}]\n```", - "options": null, - "id": 204 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000685", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, pedestrian, van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[452.66, 202.69], [484.14, 202.43], [487.28, 205.77], [452.24, 206.1], [452.66, 168.58], [484.14, 168.62], [487.28, 168.14], [452.24, 168.1]]}, {\"category\": \"cyclist\", \"corners_3d\": [[511.91, 256.43], [553.03, 255.49], [565.31, 266.87], [518.61, 268.09], [511.91, 168.81], [553.03, 168.85], [565.31, 168.3], [518.61, 168.24]]}, {\"category\": \"car\", \"corners_3d\": [[938.12, 309.68], [1121.5, 324.05], [1262.33, 471.21], [906.93, 420.0], [938.12, 174.61], [1121.5, 174.8], [1262.33, 176.69], [906.93, 176.03]]}, {\"category\": \"car\", \"corners_3d\": [[646.77, 269.14], [738.35, 266.65], [866.93, 310.59], [735.83, 316.03], [646.77, 179.58], [738.35, 179.41], [866.93, 182.47], [735.83, 182.86]]}, {\"category\": \"car\", \"corners_3d\": [[196.31, 312.57], [62.89, 316.18], [189.54, 274.65], [282.71, 272.81], [196.31, 204.52], [62.89, 205.34], [189.54, 195.93], [282.71, 195.51]]}, {\"category\": \"van\", \"corners_3d\": [[294.6, 266.53], [205.84, 268.68], [262.77, 247.44], [331.21, 246.13], [294.6, 182.35], [205.84, 182.57], [262.77, 180.41], [331.21, 180.28]]}, {\"category\": \"car\", \"corners_3d\": [[577.88, 235.05], [642.63, 233.99], [681.86, 248.0], [602.79, 249.6], [577.88, 172.45], [642.63, 172.46], [681.86, 172.36], [602.79, 172.35]]}, {\"category\": \"car\", \"corners_3d\": [[543.16, 222.61], [595.05, 221.76], [625.38, 230.94], [564.17, 232.15], [543.16, 175.92], [595.05, 175.86], [625.38, 176.43], [564.17, 176.5]]}, {\"category\": \"car\", \"corners_3d\": [[341.75, 245.64], [280.47, 247.14], [301.59, 235.65], [353.19, 234.57], [341.75, 194.75], [280.47, 195.2], [301.59, 191.74], [353.19, 191.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[764.78, 230.79], [733.36, 231.36], [720.06, 228.56], [750.12, 228.05], [764.78, 160.33], [733.36, 160.21], [720.06, 160.81], [750.12, 160.92]]}, {\"category\": \"car\", \"corners_3d\": [[356.74, 227.15], [312.0, 227.91], [326.81, 221.14], [365.93, 220.55], [356.74, 187.81], [312.0, 188.02], [326.81, 186.15], [365.93, 185.99]]}, {\"category\": \"car\", \"corners_3d\": [[527.62, 212.46], [565.56, 212.01], [581.26, 217.27], [538.35, 217.86], [527.62, 177.67], [565.56, 177.62], [581.26, 178.26], [538.35, 178.33]]}]\n```", - "options": null, - "id": 205 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000690", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-2024.65, 2106.3], [-3637.04, 2017.06], [10.0, 438.38], [257.31, 440.16], [-2024.65, 508.8], [-3637.04, 493.29], [10.0, 218.99], [257.31, 219.3]]}, {\"category\": \"car\", \"corners_3d\": [[532.8, 233.04], [480.35, 232.91], [500.69, 224.78], [546.07, 224.88], [532.8, 188.36], [480.35, 188.32], [500.69, 186.23], [546.07, 186.25]]}]\n```", - "options": null, - "id": 206 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000694", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-53.36, 730.82], [-697.36, 763.98], [160.89, 359.97], [357.03, 356.51], [-53.36, 77.96], [-697.36, 72.32], [160.89, 141.03], [357.03, 141.62]]}, {\"category\": \"car\", \"corners_3d\": [[773.04, 284.58], [884.39, 284.42], [1079.35, 361.69], [891.16, 362.15], [773.04, 189.49], [884.39, 189.46], [1079.35, 200.97], [891.16, 201.04]]}, {\"category\": \"car\", \"corners_3d\": [[516.57, 216.21], [468.97, 216.29], [485.84, 210.1], [526.63, 210.04], [516.57, 170.09], [468.97, 170.09], [485.84, 170.48], [526.63, 170.49]]}, {\"category\": \"car\", \"corners_3d\": [[382.89, 316.01], [234.17, 317.49], [344.88, 270.22], [444.56, 269.54], [382.89, 174.69], [234.17, 174.71], [344.88, 174.1], [444.56, 174.1]]}, {\"category\": \"car\", \"corners_3d\": [[721.26, 252.92], [793.84, 252.6], [854.82, 275.48], [761.63, 276.01], [721.26, 180.99], [793.84, 180.96], [854.82, 183.28], [761.63, 183.33]]}, {\"category\": \"car\", \"corners_3d\": [[679.2, 224.68], [730.64, 224.42], [759.75, 233.19], [699.68, 233.55], [679.2, 175.52], [730.64, 175.5], [759.75, 175.96], [699.68, 175.97]]}, {\"category\": \"car\", \"corners_3d\": [[652.16, 213.75], [694.69, 213.63], [713.87, 220.11], [664.62, 220.27], [652.16, 173.12], [694.69, 173.12], [713.87, 173.16], [664.62, 173.16]]}, {\"category\": \"car\", \"corners_3d\": [[638.57, 204.47], [672.02, 204.42], [686.64, 209.5], [647.82, 209.57], [638.57, 174.15], [672.02, 174.15], [686.64, 174.36], [647.82, 174.36]]}]\n```", - "options": null, - "id": 207 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000748", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[575.86, 214.97], [611.76, 214.98], [611.92, 219.03], [572.56, 219.02], [575.86, 180.53], [611.76, 180.53], [611.92, 181.26], [572.56, 181.26]]}, {\"category\": \"car\", \"corners_3d\": [[186.9, 253.29], [124.24, 253.5], [206.96, 239.1], [258.27, 238.95], [186.9, 200.5], [124.24, 200.57], [206.96, 195.62], [258.27, 195.57]]}, {\"category\": \"car\", \"corners_3d\": [[372.44, 218.35], [339.16, 218.33], [365.54, 214.0], [395.65, 214.01], [372.44, 188.66], [339.16, 188.66], [365.54, 187.15], [395.65, 187.16]]}]\n```", - "options": null, - "id": 208 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000748", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[520.84, 197.65], [550.89, 197.62], [538.84, 205.56], [499.13, 205.61], [520.84, 155.89], [550.89, 155.92], [538.84, 150.49], [499.13, 150.45]]}]\n```", - "options": null, - "id": 209 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000748", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[575.86, 214.97], [611.76, 214.98], [611.92, 219.03], [572.56, 219.02], [575.86, 180.53], [611.76, 180.53], [611.92, 181.26], [572.56, 181.26]]}, {\"category\": \"tram\", \"corners_3d\": [[520.84, 197.65], [550.89, 197.62], [538.84, 205.56], [499.13, 205.61], [520.84, 155.89], [550.89, 155.92], [538.84, 150.49], [499.13, 150.45]]}, {\"category\": \"car\", \"corners_3d\": [[186.9, 253.29], [124.24, 253.5], [206.96, 239.1], [258.27, 238.95], [186.9, 200.5], [124.24, 200.57], [206.96, 195.62], [258.27, 195.57]]}, {\"category\": \"car\", \"corners_3d\": [[372.44, 218.35], [339.16, 218.33], [365.54, 214.0], [395.65, 214.01], [372.44, 188.66], [339.16, 188.66], [365.54, 187.15], [395.65, 187.16]]}]\n```", - "options": null, - "id": 210 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000767", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (bus). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[626.2, 194.51], [661.24, 194.58], [663.53, 200.97], [618.2, 200.87], [626.2, 156.64], [661.24, 156.59], [663.53, 151.8], [618.2, 151.88]]}]\n```", - "options": null, - "id": 211 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000767", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[332.0, 211.74], [299.81, 211.72], [328.86, 208.15], [358.11, 208.17], [332.0, 187.05], [299.81, 187.05], [328.86, 185.75], [358.11, 185.75]]}]\n```", - "options": null, - "id": 212 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000767", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (bus, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[626.2, 194.51], [661.24, 194.58], [663.53, 200.97], [618.2, 200.87], [626.2, 156.64], [661.24, 156.59], [663.53, 151.8], [618.2, 151.88]]}, {\"category\": \"car\", \"corners_3d\": [[332.0, 211.74], [299.81, 211.72], [328.86, 208.15], [358.11, 208.17], [332.0, 187.05], [299.81, 187.05], [328.86, 185.75], [358.11, 185.75]]}]\n```", - "options": null, - "id": 213 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000774", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (motorcyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"motorcyclist\", \"corners_3d\": [[260.14, 452.69], [138.43, 452.25], [270.0, 375.31], [358.3, 375.54], [260.14, 158.8], [138.43, 158.82], [270.0, 162.68], [358.3, 162.67]]}]\n```", - "options": null, - "id": 214 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000774", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[661.22, 174.57], [633.37, 174.56], [634.43, 174.41], [659.76, 174.41], [661.22, 148.79], [633.37, 148.83], [634.43, 151.0], [659.76, 150.97]]}]\n```", - "options": null, - "id": 215 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000774", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (motorcyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"motorcyclist\", \"corners_3d\": [[260.14, 452.69], [138.43, 452.25], [270.0, 375.31], [358.3, 375.54], [260.14, 158.8], [138.43, 158.82], [270.0, 162.68], [358.3, 162.67]]}, {\"category\": \"car\", \"corners_3d\": [[661.22, 174.57], [633.37, 174.56], [634.43, 174.41], [659.76, 174.41], [661.22, 148.79], [633.37, 148.83], [634.43, 151.0], [659.76, 150.97]]}]\n```", - "options": null, - "id": 216 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000777", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[580.04, 200.65], [605.84, 200.65], [605.41, 203.4], [577.05, 203.4], [580.04, 176.84], [605.84, 176.84], [605.41, 177.24], [577.05, 177.24]]}, {\"category\": \"car\", \"corners_3d\": [[652.26, 202.05], [648.48, 200.89], [721.25, 200.68], [728.01, 201.83], [652.26, 174.49], [648.48, 174.42], [721.25, 174.41], [728.01, 174.48]]}]\n```", - "options": null, - "id": 217 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000864", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[575.81, 212.94], [603.91, 212.86], [607.15, 217.21], [576.01, 217.3], [575.81, 187.0], [603.91, 186.97], [607.15, 188.5], [576.01, 188.54]]}, {\"category\": \"car\", \"corners_3d\": [[802.84, 274.68], [774.5, 262.17], [994.52, 260.74], [1052.61, 272.83], [802.84, 182.04], [774.5, 180.91], [994.52, 180.78], [1052.61, 181.87]]}, {\"category\": \"car\", \"corners_3d\": [[947.54, 249.56], [978.73, 257.49], [768.32, 256.28], [756.98, 248.57], [947.54, 182.37], [978.73, 183.36], [768.32, 183.21], [756.98, 182.25]]}, {\"category\": \"car\", \"corners_3d\": [[848.62, 228.95], [860.36, 233.28], [712.14, 231.84], [711.05, 227.71], [848.62, 182.4], [860.36, 183.14], [712.14, 182.89], [711.05, 182.19]]}, {\"category\": \"car\", \"corners_3d\": [[755.31, 247.2], [745.88, 241.14], [897.39, 241.73], [920.49, 247.91], [755.31, 180.53], [745.88, 179.91], [897.39, 179.97], [920.49, 180.6]]}]\n```", - "options": null, - "id": 218 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000889", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[280.07, 356.2], [217.31, 355.42], [266.31, 335.09], [321.94, 335.7], [280.07, 154.71], [217.31, 154.82], [266.31, 157.81], [321.94, 157.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[206.46, 370.47], [131.1, 369.45], [197.62, 344.7], [263.37, 345.47], [206.46, 154.69], [131.1, 154.83], [197.62, 158.19], [263.37, 158.09]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[378.23, 231.95], [351.88, 231.91], [363.64, 229.64], [388.83, 229.67], [378.23, 167.17], [351.88, 167.18], [363.64, 167.77], [388.83, 167.76]]}]\n```", - "options": null, - "id": 219 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000889", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[582.76, 206.95], [623.85, 206.86], [635.14, 211.69], [586.54, 211.81], [582.76, 159.22], [623.85, 159.29], [635.14, 155.41], [586.54, 155.31]]}]\n```", - "options": null, - "id": 220 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000889", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[527.57, 192.64], [518.36, 192.65], [520.27, 192.18], [529.13, 192.17], [527.57, 164.31], [518.36, 164.3], [520.27, 164.93], [529.13, 164.94]]}]\n```", - "options": null, - "id": 221 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000889", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[582.76, 206.95], [623.85, 206.86], [635.14, 211.69], [586.54, 211.81], [582.76, 159.22], [623.85, 159.29], [635.14, 155.41], [586.54, 155.31]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[280.07, 356.2], [217.31, 355.42], [266.31, 335.09], [321.94, 335.7], [280.07, 154.71], [217.31, 154.82], [266.31, 157.81], [321.94, 157.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[206.46, 370.47], [131.1, 369.45], [197.62, 344.7], [263.37, 345.47], [206.46, 154.69], [131.1, 154.83], [197.62, 158.19], [263.37, 158.09]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[378.23, 231.95], [351.88, 231.91], [363.64, 229.64], [388.83, 229.67], [378.23, 167.17], [351.88, 167.18], [363.64, 167.77], [388.83, 167.76]]}]\n```", - "options": null, - "id": 222 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000889", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[280.07, 356.2], [217.31, 355.42], [266.31, 335.09], [321.94, 335.7], [280.07, 154.71], [217.31, 154.82], [266.31, 157.81], [321.94, 157.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[206.46, 370.47], [131.1, 369.45], [197.62, 344.7], [263.37, 345.47], [206.46, 154.69], [131.1, 154.83], [197.62, 158.19], [263.37, 158.09]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[378.23, 231.95], [351.88, 231.91], [363.64, 229.64], [388.83, 229.67], [378.23, 167.17], [351.88, 167.18], [363.64, 167.77], [388.83, 167.76]]}, {\"category\": \"cyclist\", \"corners_3d\": [[527.57, 192.64], [518.36, 192.65], [520.27, 192.18], [529.13, 192.17], [527.57, 164.31], [518.36, 164.3], [520.27, 164.93], [529.13, 164.94]]}]\n```", - "options": null, - "id": 223 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000889", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[582.76, 206.95], [623.85, 206.86], [635.14, 211.69], [586.54, 211.81], [582.76, 159.22], [623.85, 159.29], [635.14, 155.41], [586.54, 155.31]]}, {\"category\": \"cyclist\", \"corners_3d\": [[527.57, 192.64], [518.36, 192.65], [520.27, 192.18], [529.13, 192.17], [527.57, 164.31], [518.36, 164.3], [520.27, 164.93], [529.13, 164.94]]}]\n```", - "options": null, - "id": 224 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000889", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[582.76, 206.95], [623.85, 206.86], [635.14, 211.69], [586.54, 211.81], [582.76, 159.22], [623.85, 159.29], [635.14, 155.41], [586.54, 155.31]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[280.07, 356.2], [217.31, 355.42], [266.31, 335.09], [321.94, 335.7], [280.07, 154.71], [217.31, 154.82], [266.31, 157.81], [321.94, 157.72]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[206.46, 370.47], [131.1, 369.45], [197.62, 344.7], [263.37, 345.47], [206.46, 154.69], [131.1, 154.83], [197.62, 158.19], [263.37, 158.09]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[378.23, 231.95], [351.88, 231.91], [363.64, 229.64], [388.83, 229.67], [378.23, 167.17], [351.88, 167.18], [363.64, 167.77], [388.83, 167.76]]}, {\"category\": \"cyclist\", \"corners_3d\": [[527.57, 192.64], [518.36, 192.65], [520.27, 192.18], [529.13, 192.17], [527.57, 164.31], [518.36, 164.3], [520.27, 164.93], [529.13, 164.94]]}]\n```", - "options": null, - "id": 225 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000892", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[903.57, 240.87], [880.55, 239.81], [891.32, 235.58], [913.03, 236.51], [903.57, 178.98], [880.55, 178.89], [891.32, 178.51], [913.03, 178.59]]}, {\"category\": \"cyclist\", \"corners_3d\": [[929.2, 233.69], [908.36, 232.28], [941.17, 228.98], [961.56, 230.23], [929.2, 169.59], [908.36, 169.67], [941.17, 169.85], [961.56, 169.78]]}]\n```", - "options": null, - "id": 226 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000892", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[391.4, 306.27], [443.59, 303.99], [445.13, 315.45], [388.32, 318.15], [391.4, 178.66], [443.59, 178.56], [445.13, 179.06], [388.32, 179.18]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[354.18, 305.98], [400.96, 304.59], [391.63, 318.98], [339.58, 320.69], [354.18, 171.15], [400.96, 171.17], [391.63, 170.98], [339.58, 170.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[775.69, 257.26], [747.21, 258.25], [727.0, 253.15], [754.02, 252.27], [775.69, 160.56], [747.21, 160.41], [727.0, 161.16], [754.02, 161.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[819.16, 255.62], [788.06, 256.89], [767.4, 252.66], [797.24, 251.52], [819.16, 157.41], [788.06, 157.17], [767.4, 157.96], [797.24, 158.18]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[367.03, 258.11], [361.45, 260.13], [320.2, 260.18], [326.74, 258.15], [367.03, 168.93], [361.45, 168.84], [320.2, 168.84], [326.74, 168.93]]}]\n```", - "options": null, - "id": 227 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000892", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[391.4, 306.27], [443.59, 303.99], [445.13, 315.45], [388.32, 318.15], [391.4, 178.66], [443.59, 178.56], [445.13, 179.06], [388.32, 179.18]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[354.18, 305.98], [400.96, 304.59], [391.63, 318.98], [339.58, 320.69], [354.18, 171.15], [400.96, 171.17], [391.63, 170.98], [339.58, 170.96]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[775.69, 257.26], [747.21, 258.25], [727.0, 253.15], [754.02, 252.27], [775.69, 160.56], [747.21, 160.41], [727.0, 161.16], [754.02, 161.28]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[819.16, 255.62], [788.06, 256.89], [767.4, 252.66], [797.24, 251.52], [819.16, 157.41], [788.06, 157.17], [767.4, 157.96], [797.24, 158.18]]}, {\"category\": \"cyclist\", \"corners_3d\": [[903.57, 240.87], [880.55, 239.81], [891.32, 235.58], [913.03, 236.51], [903.57, 178.98], [880.55, 178.89], [891.32, 178.51], [913.03, 178.59]]}, {\"category\": \"cyclist\", \"corners_3d\": [[929.2, 233.69], [908.36, 232.28], [941.17, 228.98], [961.56, 230.23], [929.2, 169.59], [908.36, 169.67], [941.17, 169.85], [961.56, 169.78]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[367.03, 258.11], [361.45, 260.13], [320.2, 260.18], [326.74, 258.15], [367.03, 168.93], [361.45, 168.84], [320.2, 168.84], [326.74, 168.93]]}]\n```", - "options": null, - "id": 228 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000914", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[758.07, 220.75], [816.52, 221.43], [832.4, 231.71], [761.53, 230.71], [758.07, 167.37], [816.52, 167.29], [832.4, 166.11], [761.53, 166.23]]}, {\"category\": \"car\", \"corners_3d\": [[-407.81, 415.35], [-570.36, 404.01], [-13.68, 304.7], [92.38, 308.31], [-407.81, 200.37], [-570.36, 199.08], [-13.68, 187.81], [92.38, 188.22]]}, {\"category\": \"car\", \"corners_3d\": [[427.91, 241.81], [369.62, 239.8], [461.91, 226.89], [510.91, 228.19], [427.91, 183.16], [369.62, 182.86], [461.91, 180.93], [510.91, 181.13]]}, {\"category\": \"car\", \"corners_3d\": [[1056.62, 189.32], [1076.48, 190.1], [996.15, 190.03], [979.99, 189.25], [1056.62, 158.38], [1076.48, 157.69], [996.15, 157.76], [979.99, 158.44]]}]\n```", - "options": null, - "id": 229 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000917", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[916.0, 204.0], [886.26, 201.89], [1001.6, 200.97], [1038.43, 202.95], [916.0, 139.44], [886.26, 141.7], [1001.6, 142.69], [1038.43, 140.57]]}]\n```", - "options": null, - "id": 230 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000917", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[286.22, 229.1], [278.52, 224.81], [446.46, 219.21], [465.52, 222.59], [286.22, 167.46], [278.52, 167.87], [446.46, 168.41], [465.52, 168.09]]}, {\"category\": \"car\", \"corners_3d\": [[1006.8, 235.54], [1093.57, 241.5], [1002.09, 253.43], [910.94, 245.34], [1006.8, 159.47], [1093.57, 158.19], [1002.09, 155.65], [910.94, 157.38]]}, {\"category\": \"car\", \"corners_3d\": [[1247.17, 196.11], [1204.67, 194.73], [1301.25, 194.37], [1348.96, 195.7], [1247.17, 149.29], [1204.67, 150.68], [1301.25, 151.05], [1348.96, 149.7]]}]\n```", - "options": null, - "id": 231 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000917", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[286.22, 229.1], [278.52, 224.81], [446.46, 219.21], [465.52, 222.59], [286.22, 167.46], [278.52, 167.87], [446.46, 168.41], [465.52, 168.09]]}, {\"category\": \"car\", \"corners_3d\": [[1006.8, 235.54], [1093.57, 241.5], [1002.09, 253.43], [910.94, 245.34], [1006.8, 159.47], [1093.57, 158.19], [1002.09, 155.65], [910.94, 157.38]]}, {\"category\": \"van\", \"corners_3d\": [[916.0, 204.0], [886.26, 201.89], [1001.6, 200.97], [1038.43, 202.95], [916.0, 139.44], [886.26, 141.7], [1001.6, 142.69], [1038.43, 140.57]]}, {\"category\": \"car\", \"corners_3d\": [[1247.17, 196.11], [1204.67, 194.73], [1301.25, 194.37], [1348.96, 195.7], [1247.17, 149.29], [1204.67, 150.68], [1301.25, 151.05], [1348.96, 149.7]]}]\n```", - "options": null, - "id": 232 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000929", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[618.13, 220.03], [664.89, 220.25], [665.49, 227.01], [612.09, 226.73], [618.13, 174.81], [664.89, 174.82], [665.49, 175.1], [612.09, 175.09]]}, {\"category\": \"car\", \"corners_3d\": [[539.7, 219.53], [496.58, 219.34], [516.19, 213.74], [554.18, 213.9], [539.7, 180.44], [496.58, 180.41], [516.19, 179.5], [554.18, 179.53]]}, {\"category\": \"car\", \"corners_3d\": [[484.23, 214.99], [449.1, 214.76], [472.71, 210.63], [504.48, 210.82], [484.23, 181.61], [449.1, 181.56], [472.71, 180.7], [504.48, 180.74]]}]\n```", - "options": null, - "id": 233 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000936", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[738.49, 240.14], [767.39, 240.12], [777.61, 244.3], [746.92, 244.33], [738.49, 177.14], [767.39, 177.14], [777.61, 177.41], [746.92, 177.41]]}]\n```", - "options": null, - "id": 234 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000936", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[452.2, 209.87], [486.63, 208.34], [558.52, 212.27], [523.55, 214.17], [452.2, 158.08], [486.63, 158.69], [558.52, 157.12], [523.55, 156.37]]}, {\"category\": \"van\", \"corners_3d\": [[1332.98, 249.47], [1451.71, 261.68], [1184.83, 262.7], [1101.97, 250.23], [1332.98, 148.24], [1451.71, 144.31], [1184.83, 143.99], [1101.97, 147.99]]}]\n```", - "options": null, - "id": 235 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000936", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.45, 221.18], [637.68, 221.18], [641.82, 228.46], [587.49, 228.46], [590.45, 183.18], [637.68, 183.18], [641.82, 184.74], [587.49, 184.74]]}, {\"category\": \"car\", \"corners_3d\": [[-97.34, 346.64], [96.82, 321.97], [212.81, 384.87], [-54.64, 438.49], [-97.34, 188.02], [96.82, 185.87], [212.81, 191.36], [-54.64, 196.04]]}, {\"category\": \"car\", \"corners_3d\": [[127.81, 290.21], [253.95, 277.1], [376.49, 305.62], [230.45, 327.63], [127.81, 182.61], [253.95, 181.52], [376.49, 183.89], [230.45, 185.72]]}, {\"category\": \"car\", \"corners_3d\": [[331.92, 239.71], [388.54, 235.51], [487.65, 245.59], [429.0, 251.32], [331.92, 183.16], [388.54, 182.51], [487.65, 184.07], [429.0, 184.95]]}, {\"category\": \"car\", \"corners_3d\": [[383.65, 232.02], [433.4, 228.99], [503.84, 236.56], [451.31, 240.49], [383.65, 181.3], [433.4, 180.86], [503.84, 181.94], [451.31, 182.5]]}, {\"category\": \"car\", \"corners_3d\": [[431.45, 216.47], [465.32, 214.79], [525.37, 219.01], [490.61, 221.05], [431.45, 180.54], [465.32, 180.25], [525.37, 180.99], [490.61, 181.35]]}, {\"category\": \"car\", \"corners_3d\": [[973.14, 360.51], [1176.51, 360.02], [2125.61, 669.72], [1589.99, 673.18], [973.14, 163.66], [1176.51, 163.68], [2125.61, 148.5], [1589.99, 148.33]]}, {\"category\": \"car\", \"corners_3d\": [[1071.43, 246.55], [1134.07, 254.74], [962.81, 257.79], [915.79, 249.01], [1071.43, 176.55], [1134.07, 176.96], [962.81, 177.11], [915.79, 176.67]]}, {\"category\": \"car\", \"corners_3d\": [[595.34, 208.98], [626.29, 209.01], [626.47, 212.7], [592.36, 212.66], [595.34, 180.81], [626.29, 180.82], [626.47, 181.64], [592.36, 181.63]]}, {\"category\": \"car\", \"corners_3d\": [[-32.87, 311.4], [-161.64, 312.14], [58.45, 271.21], [148.69, 270.84], [-32.87, 190.15], [-161.64, 190.24], [58.45, 185.13], [148.69, 185.08]]}, {\"category\": \"car\", \"corners_3d\": [[217.24, 221.22], [166.19, 221.25], [229.98, 214.16], [273.52, 214.14], [217.24, 174.9], [166.19, 174.9], [229.98, 174.6], [273.52, 174.6]]}, {\"category\": \"car\", \"corners_3d\": [[288.91, 211.41], [255.84, 211.3], [294.18, 207.58], [324.14, 207.67], [288.91, 181.96], [255.84, 181.94], [294.18, 181.06], [324.14, 181.08]]}]\n```", - "options": null, - "id": 236 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000936", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[422.62, 201.54], [387.85, 201.54], [416.17, 197.87], [446.49, 197.86], [422.62, 159.42], [387.85, 159.42], [416.17, 161.14], [446.49, 161.14]]}]\n```", - "options": null, - "id": 237 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000936", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[452.2, 209.87], [486.63, 208.34], [558.52, 212.27], [523.55, 214.17], [452.2, 158.08], [486.63, 158.69], [558.52, 157.12], [523.55, 156.37]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[738.49, 240.14], [767.39, 240.12], [777.61, 244.3], [746.92, 244.33], [738.49, 177.14], [767.39, 177.14], [777.61, 177.41], [746.92, 177.41]]}, {\"category\": \"van\", \"corners_3d\": [[1332.98, 249.47], [1451.71, 261.68], [1184.83, 262.7], [1101.97, 250.23], [1332.98, 148.24], [1451.71, 144.31], [1184.83, 143.99], [1101.97, 147.99]]}]\n```", - "options": null, - "id": 238 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000936", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.45, 221.18], [637.68, 221.18], [641.82, 228.46], [587.49, 228.46], [590.45, 183.18], [637.68, 183.18], [641.82, 184.74], [587.49, 184.74]]}, {\"category\": \"car\", \"corners_3d\": [[-97.34, 346.64], [96.82, 321.97], [212.81, 384.87], [-54.64, 438.49], [-97.34, 188.02], [96.82, 185.87], [212.81, 191.36], [-54.64, 196.04]]}, {\"category\": \"car\", \"corners_3d\": [[127.81, 290.21], [253.95, 277.1], [376.49, 305.62], [230.45, 327.63], [127.81, 182.61], [253.95, 181.52], [376.49, 183.89], [230.45, 185.72]]}, {\"category\": \"car\", \"corners_3d\": [[331.92, 239.71], [388.54, 235.51], [487.65, 245.59], [429.0, 251.32], [331.92, 183.16], [388.54, 182.51], [487.65, 184.07], [429.0, 184.95]]}, {\"category\": \"car\", \"corners_3d\": [[383.65, 232.02], [433.4, 228.99], [503.84, 236.56], [451.31, 240.49], [383.65, 181.3], [433.4, 180.86], [503.84, 181.94], [451.31, 182.5]]}, {\"category\": \"car\", \"corners_3d\": [[431.45, 216.47], [465.32, 214.79], [525.37, 219.01], [490.61, 221.05], [431.45, 180.54], [465.32, 180.25], [525.37, 180.99], [490.61, 181.35]]}, {\"category\": \"car\", \"corners_3d\": [[973.14, 360.51], [1176.51, 360.02], [2125.61, 669.72], [1589.99, 673.18], [973.14, 163.66], [1176.51, 163.68], [2125.61, 148.5], [1589.99, 148.33]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[738.49, 240.14], [767.39, 240.12], [777.61, 244.3], [746.92, 244.33], [738.49, 177.14], [767.39, 177.14], [777.61, 177.41], [746.92, 177.41]]}, {\"category\": \"car\", \"corners_3d\": [[1071.43, 246.55], [1134.07, 254.74], [962.81, 257.79], [915.79, 249.01], [1071.43, 176.55], [1134.07, 176.96], [962.81, 177.11], [915.79, 176.67]]}, {\"category\": \"car\", \"corners_3d\": [[595.34, 208.98], [626.29, 209.01], [626.47, 212.7], [592.36, 212.66], [595.34, 180.81], [626.29, 180.82], [626.47, 181.64], [592.36, 181.63]]}, {\"category\": \"car\", \"corners_3d\": [[-32.87, 311.4], [-161.64, 312.14], [58.45, 271.21], [148.69, 270.84], [-32.87, 190.15], [-161.64, 190.24], [58.45, 185.13], [148.69, 185.08]]}, {\"category\": \"car\", \"corners_3d\": [[217.24, 221.22], [166.19, 221.25], [229.98, 214.16], [273.52, 214.14], [217.24, 174.9], [166.19, 174.9], [229.98, 174.6], [273.52, 174.6]]}, {\"category\": \"car\", \"corners_3d\": [[288.91, 211.41], [255.84, 211.3], [294.18, 207.58], [324.14, 207.67], [288.91, 181.96], [255.84, 181.94], [294.18, 181.06], [324.14, 181.08]]}]\n```", - "options": null, - "id": 239 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000936", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[738.49, 240.14], [767.39, 240.12], [777.61, 244.3], [746.92, 244.33], [738.49, 177.14], [767.39, 177.14], [777.61, 177.41], [746.92, 177.41]]}, {\"category\": \"truck\", \"corners_3d\": [[422.62, 201.54], [387.85, 201.54], [416.17, 197.87], [446.49, 197.86], [422.62, 159.42], [387.85, 159.42], [416.17, 161.14], [446.49, 161.14]]}]\n```", - "options": null, - "id": 240 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000936", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.45, 221.18], [637.68, 221.18], [641.82, 228.46], [587.49, 228.46], [590.45, 183.18], [637.68, 183.18], [641.82, 184.74], [587.49, 184.74]]}, {\"category\": \"car\", \"corners_3d\": [[-97.34, 346.64], [96.82, 321.97], [212.81, 384.87], [-54.64, 438.49], [-97.34, 188.02], [96.82, 185.87], [212.81, 191.36], [-54.64, 196.04]]}, {\"category\": \"car\", \"corners_3d\": [[127.81, 290.21], [253.95, 277.1], [376.49, 305.62], [230.45, 327.63], [127.81, 182.61], [253.95, 181.52], [376.49, 183.89], [230.45, 185.72]]}, {\"category\": \"car\", \"corners_3d\": [[331.92, 239.71], [388.54, 235.51], [487.65, 245.59], [429.0, 251.32], [331.92, 183.16], [388.54, 182.51], [487.65, 184.07], [429.0, 184.95]]}, {\"category\": \"car\", \"corners_3d\": [[383.65, 232.02], [433.4, 228.99], [503.84, 236.56], [451.31, 240.49], [383.65, 181.3], [433.4, 180.86], [503.84, 181.94], [451.31, 182.5]]}, {\"category\": \"car\", \"corners_3d\": [[431.45, 216.47], [465.32, 214.79], [525.37, 219.01], [490.61, 221.05], [431.45, 180.54], [465.32, 180.25], [525.37, 180.99], [490.61, 181.35]]}, {\"category\": \"van\", \"corners_3d\": [[452.2, 209.87], [486.63, 208.34], [558.52, 212.27], [523.55, 214.17], [452.2, 158.08], [486.63, 158.69], [558.52, 157.12], [523.55, 156.37]]}, {\"category\": \"car\", \"corners_3d\": [[973.14, 360.51], [1176.51, 360.02], [2125.61, 669.72], [1589.99, 673.18], [973.14, 163.66], [1176.51, 163.68], [2125.61, 148.5], [1589.99, 148.33]]}, {\"category\": \"car\", \"corners_3d\": [[1071.43, 246.55], [1134.07, 254.74], [962.81, 257.79], [915.79, 249.01], [1071.43, 176.55], [1134.07, 176.96], [962.81, 177.11], [915.79, 176.67]]}, {\"category\": \"car\", \"corners_3d\": [[595.34, 208.98], [626.29, 209.01], [626.47, 212.7], [592.36, 212.66], [595.34, 180.81], [626.29, 180.82], [626.47, 181.64], [592.36, 181.63]]}, {\"category\": \"car\", \"corners_3d\": [[-32.87, 311.4], [-161.64, 312.14], [58.45, 271.21], [148.69, 270.84], [-32.87, 190.15], [-161.64, 190.24], [58.45, 185.13], [148.69, 185.08]]}, {\"category\": \"van\", \"corners_3d\": [[1332.98, 249.47], [1451.71, 261.68], [1184.83, 262.7], [1101.97, 250.23], [1332.98, 148.24], [1451.71, 144.31], [1184.83, 143.99], [1101.97, 147.99]]}, {\"category\": \"car\", \"corners_3d\": [[217.24, 221.22], [166.19, 221.25], [229.98, 214.16], [273.52, 214.14], [217.24, 174.9], [166.19, 174.9], [229.98, 174.6], [273.52, 174.6]]}, {\"category\": \"car\", \"corners_3d\": [[288.91, 211.41], [255.84, 211.3], [294.18, 207.58], [324.14, 207.67], [288.91, 181.96], [255.84, 181.94], [294.18, 181.06], [324.14, 181.08]]}]\n```", - "options": null, - "id": 241 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000936", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[452.2, 209.87], [486.63, 208.34], [558.52, 212.27], [523.55, 214.17], [452.2, 158.08], [486.63, 158.69], [558.52, 157.12], [523.55, 156.37]]}, {\"category\": \"van\", \"corners_3d\": [[1332.98, 249.47], [1451.71, 261.68], [1184.83, 262.7], [1101.97, 250.23], [1332.98, 148.24], [1451.71, 144.31], [1184.83, 143.99], [1101.97, 147.99]]}, {\"category\": \"truck\", \"corners_3d\": [[422.62, 201.54], [387.85, 201.54], [416.17, 197.87], [446.49, 197.86], [422.62, 159.42], [387.85, 159.42], [416.17, 161.14], [446.49, 161.14]]}]\n```", - "options": null, - "id": 242 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000936", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.45, 221.18], [637.68, 221.18], [641.82, 228.46], [587.49, 228.46], [590.45, 183.18], [637.68, 183.18], [641.82, 184.74], [587.49, 184.74]]}, {\"category\": \"car\", \"corners_3d\": [[-97.34, 346.64], [96.82, 321.97], [212.81, 384.87], [-54.64, 438.49], [-97.34, 188.02], [96.82, 185.87], [212.81, 191.36], [-54.64, 196.04]]}, {\"category\": \"car\", \"corners_3d\": [[127.81, 290.21], [253.95, 277.1], [376.49, 305.62], [230.45, 327.63], [127.81, 182.61], [253.95, 181.52], [376.49, 183.89], [230.45, 185.72]]}, {\"category\": \"car\", \"corners_3d\": [[331.92, 239.71], [388.54, 235.51], [487.65, 245.59], [429.0, 251.32], [331.92, 183.16], [388.54, 182.51], [487.65, 184.07], [429.0, 184.95]]}, {\"category\": \"car\", \"corners_3d\": [[383.65, 232.02], [433.4, 228.99], [503.84, 236.56], [451.31, 240.49], [383.65, 181.3], [433.4, 180.86], [503.84, 181.94], [451.31, 182.5]]}, {\"category\": \"car\", \"corners_3d\": [[431.45, 216.47], [465.32, 214.79], [525.37, 219.01], [490.61, 221.05], [431.45, 180.54], [465.32, 180.25], [525.37, 180.99], [490.61, 181.35]]}, {\"category\": \"car\", \"corners_3d\": [[973.14, 360.51], [1176.51, 360.02], [2125.61, 669.72], [1589.99, 673.18], [973.14, 163.66], [1176.51, 163.68], [2125.61, 148.5], [1589.99, 148.33]]}, {\"category\": \"car\", \"corners_3d\": [[1071.43, 246.55], [1134.07, 254.74], [962.81, 257.79], [915.79, 249.01], [1071.43, 176.55], [1134.07, 176.96], [962.81, 177.11], [915.79, 176.67]]}, {\"category\": \"car\", \"corners_3d\": [[595.34, 208.98], [626.29, 209.01], [626.47, 212.7], [592.36, 212.66], [595.34, 180.81], [626.29, 180.82], [626.47, 181.64], [592.36, 181.63]]}, {\"category\": \"car\", \"corners_3d\": [[-32.87, 311.4], [-161.64, 312.14], [58.45, 271.21], [148.69, 270.84], [-32.87, 190.15], [-161.64, 190.24], [58.45, 185.13], [148.69, 185.08]]}, {\"category\": \"car\", \"corners_3d\": [[217.24, 221.22], [166.19, 221.25], [229.98, 214.16], [273.52, 214.14], [217.24, 174.9], [166.19, 174.9], [229.98, 174.6], [273.52, 174.6]]}, {\"category\": \"car\", \"corners_3d\": [[288.91, 211.41], [255.84, 211.3], [294.18, 207.58], [324.14, 207.67], [288.91, 181.96], [255.84, 181.94], [294.18, 181.06], [324.14, 181.08]]}, {\"category\": \"truck\", \"corners_3d\": [[422.62, 201.54], [387.85, 201.54], [416.17, 197.87], [446.49, 197.86], [422.62, 159.42], [387.85, 159.42], [416.17, 161.14], [446.49, 161.14]]}]\n```", - "options": null, - "id": 243 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000936", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.45, 221.18], [637.68, 221.18], [641.82, 228.46], [587.49, 228.46], [590.45, 183.18], [637.68, 183.18], [641.82, 184.74], [587.49, 184.74]]}, {\"category\": \"car\", \"corners_3d\": [[-97.34, 346.64], [96.82, 321.97], [212.81, 384.87], [-54.64, 438.49], [-97.34, 188.02], [96.82, 185.87], [212.81, 191.36], [-54.64, 196.04]]}, {\"category\": \"car\", \"corners_3d\": [[127.81, 290.21], [253.95, 277.1], [376.49, 305.62], [230.45, 327.63], [127.81, 182.61], [253.95, 181.52], [376.49, 183.89], [230.45, 185.72]]}, {\"category\": \"car\", \"corners_3d\": [[331.92, 239.71], [388.54, 235.51], [487.65, 245.59], [429.0, 251.32], [331.92, 183.16], [388.54, 182.51], [487.65, 184.07], [429.0, 184.95]]}, {\"category\": \"car\", \"corners_3d\": [[383.65, 232.02], [433.4, 228.99], [503.84, 236.56], [451.31, 240.49], [383.65, 181.3], [433.4, 180.86], [503.84, 181.94], [451.31, 182.5]]}, {\"category\": \"car\", \"corners_3d\": [[431.45, 216.47], [465.32, 214.79], [525.37, 219.01], [490.61, 221.05], [431.45, 180.54], [465.32, 180.25], [525.37, 180.99], [490.61, 181.35]]}, {\"category\": \"van\", \"corners_3d\": [[452.2, 209.87], [486.63, 208.34], [558.52, 212.27], [523.55, 214.17], [452.2, 158.08], [486.63, 158.69], [558.52, 157.12], [523.55, 156.37]]}, {\"category\": \"car\", \"corners_3d\": [[973.14, 360.51], [1176.51, 360.02], [2125.61, 669.72], [1589.99, 673.18], [973.14, 163.66], [1176.51, 163.68], [2125.61, 148.5], [1589.99, 148.33]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[738.49, 240.14], [767.39, 240.12], [777.61, 244.3], [746.92, 244.33], [738.49, 177.14], [767.39, 177.14], [777.61, 177.41], [746.92, 177.41]]}, {\"category\": \"car\", \"corners_3d\": [[1071.43, 246.55], [1134.07, 254.74], [962.81, 257.79], [915.79, 249.01], [1071.43, 176.55], [1134.07, 176.96], [962.81, 177.11], [915.79, 176.67]]}, {\"category\": \"car\", \"corners_3d\": [[595.34, 208.98], [626.29, 209.01], [626.47, 212.7], [592.36, 212.66], [595.34, 180.81], [626.29, 180.82], [626.47, 181.64], [592.36, 181.63]]}, {\"category\": \"car\", \"corners_3d\": [[-32.87, 311.4], [-161.64, 312.14], [58.45, 271.21], [148.69, 270.84], [-32.87, 190.15], [-161.64, 190.24], [58.45, 185.13], [148.69, 185.08]]}, {\"category\": \"van\", \"corners_3d\": [[1332.98, 249.47], [1451.71, 261.68], [1184.83, 262.7], [1101.97, 250.23], [1332.98, 148.24], [1451.71, 144.31], [1184.83, 143.99], [1101.97, 147.99]]}, {\"category\": \"car\", \"corners_3d\": [[217.24, 221.22], [166.19, 221.25], [229.98, 214.16], [273.52, 214.14], [217.24, 174.9], [166.19, 174.9], [229.98, 174.6], [273.52, 174.6]]}, {\"category\": \"car\", \"corners_3d\": [[288.91, 211.41], [255.84, 211.3], [294.18, 207.58], [324.14, 207.67], [288.91, 181.96], [255.84, 181.94], [294.18, 181.06], [324.14, 181.08]]}]\n```", - "options": null, - "id": 244 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000936", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, van, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[452.2, 209.87], [486.63, 208.34], [558.52, 212.27], [523.55, 214.17], [452.2, 158.08], [486.63, 158.69], [558.52, 157.12], [523.55, 156.37]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[738.49, 240.14], [767.39, 240.12], [777.61, 244.3], [746.92, 244.33], [738.49, 177.14], [767.39, 177.14], [777.61, 177.41], [746.92, 177.41]]}, {\"category\": \"van\", \"corners_3d\": [[1332.98, 249.47], [1451.71, 261.68], [1184.83, 262.7], [1101.97, 250.23], [1332.98, 148.24], [1451.71, 144.31], [1184.83, 143.99], [1101.97, 147.99]]}, {\"category\": \"truck\", \"corners_3d\": [[422.62, 201.54], [387.85, 201.54], [416.17, 197.87], [446.49, 197.86], [422.62, 159.42], [387.85, 159.42], [416.17, 161.14], [446.49, 161.14]]}]\n```", - "options": null, - "id": 245 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000936", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.45, 221.18], [637.68, 221.18], [641.82, 228.46], [587.49, 228.46], [590.45, 183.18], [637.68, 183.18], [641.82, 184.74], [587.49, 184.74]]}, {\"category\": \"car\", \"corners_3d\": [[-97.34, 346.64], [96.82, 321.97], [212.81, 384.87], [-54.64, 438.49], [-97.34, 188.02], [96.82, 185.87], [212.81, 191.36], [-54.64, 196.04]]}, {\"category\": \"car\", \"corners_3d\": [[127.81, 290.21], [253.95, 277.1], [376.49, 305.62], [230.45, 327.63], [127.81, 182.61], [253.95, 181.52], [376.49, 183.89], [230.45, 185.72]]}, {\"category\": \"car\", \"corners_3d\": [[331.92, 239.71], [388.54, 235.51], [487.65, 245.59], [429.0, 251.32], [331.92, 183.16], [388.54, 182.51], [487.65, 184.07], [429.0, 184.95]]}, {\"category\": \"car\", \"corners_3d\": [[383.65, 232.02], [433.4, 228.99], [503.84, 236.56], [451.31, 240.49], [383.65, 181.3], [433.4, 180.86], [503.84, 181.94], [451.31, 182.5]]}, {\"category\": \"car\", \"corners_3d\": [[431.45, 216.47], [465.32, 214.79], [525.37, 219.01], [490.61, 221.05], [431.45, 180.54], [465.32, 180.25], [525.37, 180.99], [490.61, 181.35]]}, {\"category\": \"car\", \"corners_3d\": [[973.14, 360.51], [1176.51, 360.02], [2125.61, 669.72], [1589.99, 673.18], [973.14, 163.66], [1176.51, 163.68], [2125.61, 148.5], [1589.99, 148.33]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[738.49, 240.14], [767.39, 240.12], [777.61, 244.3], [746.92, 244.33], [738.49, 177.14], [767.39, 177.14], [777.61, 177.41], [746.92, 177.41]]}, {\"category\": \"car\", \"corners_3d\": [[1071.43, 246.55], [1134.07, 254.74], [962.81, 257.79], [915.79, 249.01], [1071.43, 176.55], [1134.07, 176.96], [962.81, 177.11], [915.79, 176.67]]}, {\"category\": \"car\", \"corners_3d\": [[595.34, 208.98], [626.29, 209.01], [626.47, 212.7], [592.36, 212.66], [595.34, 180.81], [626.29, 180.82], [626.47, 181.64], [592.36, 181.63]]}, {\"category\": \"car\", \"corners_3d\": [[-32.87, 311.4], [-161.64, 312.14], [58.45, 271.21], [148.69, 270.84], [-32.87, 190.15], [-161.64, 190.24], [58.45, 185.13], [148.69, 185.08]]}, {\"category\": \"car\", \"corners_3d\": [[217.24, 221.22], [166.19, 221.25], [229.98, 214.16], [273.52, 214.14], [217.24, 174.9], [166.19, 174.9], [229.98, 174.6], [273.52, 174.6]]}, {\"category\": \"car\", \"corners_3d\": [[288.91, 211.41], [255.84, 211.3], [294.18, 207.58], [324.14, 207.67], [288.91, 181.96], [255.84, 181.94], [294.18, 181.06], [324.14, 181.08]]}, {\"category\": \"truck\", \"corners_3d\": [[422.62, 201.54], [387.85, 201.54], [416.17, 197.87], [446.49, 197.86], [422.62, 159.42], [387.85, 159.42], [416.17, 161.14], [446.49, 161.14]]}]\n```", - "options": null, - "id": 246 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000936", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.45, 221.18], [637.68, 221.18], [641.82, 228.46], [587.49, 228.46], [590.45, 183.18], [637.68, 183.18], [641.82, 184.74], [587.49, 184.74]]}, {\"category\": \"car\", \"corners_3d\": [[-97.34, 346.64], [96.82, 321.97], [212.81, 384.87], [-54.64, 438.49], [-97.34, 188.02], [96.82, 185.87], [212.81, 191.36], [-54.64, 196.04]]}, {\"category\": \"car\", \"corners_3d\": [[127.81, 290.21], [253.95, 277.1], [376.49, 305.62], [230.45, 327.63], [127.81, 182.61], [253.95, 181.52], [376.49, 183.89], [230.45, 185.72]]}, {\"category\": \"car\", \"corners_3d\": [[331.92, 239.71], [388.54, 235.51], [487.65, 245.59], [429.0, 251.32], [331.92, 183.16], [388.54, 182.51], [487.65, 184.07], [429.0, 184.95]]}, {\"category\": \"car\", \"corners_3d\": [[383.65, 232.02], [433.4, 228.99], [503.84, 236.56], [451.31, 240.49], [383.65, 181.3], [433.4, 180.86], [503.84, 181.94], [451.31, 182.5]]}, {\"category\": \"car\", \"corners_3d\": [[431.45, 216.47], [465.32, 214.79], [525.37, 219.01], [490.61, 221.05], [431.45, 180.54], [465.32, 180.25], [525.37, 180.99], [490.61, 181.35]]}, {\"category\": \"van\", \"corners_3d\": [[452.2, 209.87], [486.63, 208.34], [558.52, 212.27], [523.55, 214.17], [452.2, 158.08], [486.63, 158.69], [558.52, 157.12], [523.55, 156.37]]}, {\"category\": \"car\", \"corners_3d\": [[973.14, 360.51], [1176.51, 360.02], [2125.61, 669.72], [1589.99, 673.18], [973.14, 163.66], [1176.51, 163.68], [2125.61, 148.5], [1589.99, 148.33]]}, {\"category\": \"car\", \"corners_3d\": [[1071.43, 246.55], [1134.07, 254.74], [962.81, 257.79], [915.79, 249.01], [1071.43, 176.55], [1134.07, 176.96], [962.81, 177.11], [915.79, 176.67]]}, {\"category\": \"car\", \"corners_3d\": [[595.34, 208.98], [626.29, 209.01], [626.47, 212.7], [592.36, 212.66], [595.34, 180.81], [626.29, 180.82], [626.47, 181.64], [592.36, 181.63]]}, {\"category\": \"car\", \"corners_3d\": [[-32.87, 311.4], [-161.64, 312.14], [58.45, 271.21], [148.69, 270.84], [-32.87, 190.15], [-161.64, 190.24], [58.45, 185.13], [148.69, 185.08]]}, {\"category\": \"van\", \"corners_3d\": [[1332.98, 249.47], [1451.71, 261.68], [1184.83, 262.7], [1101.97, 250.23], [1332.98, 148.24], [1451.71, 144.31], [1184.83, 143.99], [1101.97, 147.99]]}, {\"category\": \"car\", \"corners_3d\": [[217.24, 221.22], [166.19, 221.25], [229.98, 214.16], [273.52, 214.14], [217.24, 174.9], [166.19, 174.9], [229.98, 174.6], [273.52, 174.6]]}, {\"category\": \"car\", \"corners_3d\": [[288.91, 211.41], [255.84, 211.3], [294.18, 207.58], [324.14, 207.67], [288.91, 181.96], [255.84, 181.94], [294.18, 181.06], [324.14, 181.08]]}, {\"category\": \"truck\", \"corners_3d\": [[422.62, 201.54], [387.85, 201.54], [416.17, 197.87], [446.49, 197.86], [422.62, 159.42], [387.85, 159.42], [416.17, 161.14], [446.49, 161.14]]}]\n```", - "options": null, - "id": 247 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000936", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, van, car, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[590.45, 221.18], [637.68, 221.18], [641.82, 228.46], [587.49, 228.46], [590.45, 183.18], [637.68, 183.18], [641.82, 184.74], [587.49, 184.74]]}, {\"category\": \"car\", \"corners_3d\": [[-97.34, 346.64], [96.82, 321.97], [212.81, 384.87], [-54.64, 438.49], [-97.34, 188.02], [96.82, 185.87], [212.81, 191.36], [-54.64, 196.04]]}, {\"category\": \"car\", \"corners_3d\": [[127.81, 290.21], [253.95, 277.1], [376.49, 305.62], [230.45, 327.63], [127.81, 182.61], [253.95, 181.52], [376.49, 183.89], [230.45, 185.72]]}, {\"category\": \"car\", \"corners_3d\": [[331.92, 239.71], [388.54, 235.51], [487.65, 245.59], [429.0, 251.32], [331.92, 183.16], [388.54, 182.51], [487.65, 184.07], [429.0, 184.95]]}, {\"category\": \"car\", \"corners_3d\": [[383.65, 232.02], [433.4, 228.99], [503.84, 236.56], [451.31, 240.49], [383.65, 181.3], [433.4, 180.86], [503.84, 181.94], [451.31, 182.5]]}, {\"category\": \"car\", \"corners_3d\": [[431.45, 216.47], [465.32, 214.79], [525.37, 219.01], [490.61, 221.05], [431.45, 180.54], [465.32, 180.25], [525.37, 180.99], [490.61, 181.35]]}, {\"category\": \"van\", \"corners_3d\": [[452.2, 209.87], [486.63, 208.34], [558.52, 212.27], [523.55, 214.17], [452.2, 158.08], [486.63, 158.69], [558.52, 157.12], [523.55, 156.37]]}, {\"category\": \"car\", \"corners_3d\": [[973.14, 360.51], [1176.51, 360.02], [2125.61, 669.72], [1589.99, 673.18], [973.14, 163.66], [1176.51, 163.68], [2125.61, 148.5], [1589.99, 148.33]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[738.49, 240.14], [767.39, 240.12], [777.61, 244.3], [746.92, 244.33], [738.49, 177.14], [767.39, 177.14], [777.61, 177.41], [746.92, 177.41]]}, {\"category\": \"car\", \"corners_3d\": [[1071.43, 246.55], [1134.07, 254.74], [962.81, 257.79], [915.79, 249.01], [1071.43, 176.55], [1134.07, 176.96], [962.81, 177.11], [915.79, 176.67]]}, {\"category\": \"car\", \"corners_3d\": [[595.34, 208.98], [626.29, 209.01], [626.47, 212.7], [592.36, 212.66], [595.34, 180.81], [626.29, 180.82], [626.47, 181.64], [592.36, 181.63]]}, {\"category\": \"car\", \"corners_3d\": [[-32.87, 311.4], [-161.64, 312.14], [58.45, 271.21], [148.69, 270.84], [-32.87, 190.15], [-161.64, 190.24], [58.45, 185.13], [148.69, 185.08]]}, {\"category\": \"van\", \"corners_3d\": [[1332.98, 249.47], [1451.71, 261.68], [1184.83, 262.7], [1101.97, 250.23], [1332.98, 148.24], [1451.71, 144.31], [1184.83, 143.99], [1101.97, 147.99]]}, {\"category\": \"car\", \"corners_3d\": [[217.24, 221.22], [166.19, 221.25], [229.98, 214.16], [273.52, 214.14], [217.24, 174.9], [166.19, 174.9], [229.98, 174.6], [273.52, 174.6]]}, {\"category\": \"car\", \"corners_3d\": [[288.91, 211.41], [255.84, 211.3], [294.18, 207.58], [324.14, 207.67], [288.91, 181.96], [255.84, 181.94], [294.18, 181.06], [324.14, 181.08]]}, {\"category\": \"truck\", \"corners_3d\": [[422.62, 201.54], [387.85, 201.54], [416.17, 197.87], [446.49, 197.86], [422.62, 159.42], [387.85, 159.42], [416.17, 161.14], [446.49, 161.14]]}]\n```", - "options": null, - "id": 248 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000940", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1106.23, 300.71], [1149.74, 300.72], [1267.88, 328.71], [1214.84, 328.7], [1106.23, 151.51], [1149.74, 151.51], [1267.88, 146.83], [1214.84, 146.84]]}]\n```", - "options": null, - "id": 249 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000940", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-55.05, 258.88], [-126.81, 258.89], [28.45, 240.74], [85.06, 240.73], [-55.05, 200.65], [-126.81, 200.65], [28.45, 194.79], [85.06, 194.79]]}, {\"category\": \"car\", \"corners_3d\": [[340.98, 212.96], [308.88, 212.97], [336.49, 209.28], [365.64, 209.27], [340.98, 185.79], [308.88, 185.79], [336.49, 184.6], [365.64, 184.6]]}, {\"category\": \"car\", \"corners_3d\": [[405.32, 204.05], [380.76, 204.04], [395.06, 202.14], [418.13, 202.15], [405.32, 185.16], [380.76, 185.16], [395.06, 184.41], [418.13, 184.42]]}]\n```", - "options": null, - "id": 250 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000940", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1106.23, 300.71], [1149.74, 300.72], [1267.88, 328.71], [1214.84, 328.7], [1106.23, 151.51], [1149.74, 151.51], [1267.88, 146.83], [1214.84, 146.84]]}, {\"category\": \"car\", \"corners_3d\": [[-55.05, 258.88], [-126.81, 258.89], [28.45, 240.74], [85.06, 240.73], [-55.05, 200.65], [-126.81, 200.65], [28.45, 194.79], [85.06, 194.79]]}, {\"category\": \"car\", \"corners_3d\": [[340.98, 212.96], [308.88, 212.97], [336.49, 209.28], [365.64, 209.27], [340.98, 185.79], [308.88, 185.79], [336.49, 184.6], [365.64, 184.6]]}, {\"category\": \"car\", \"corners_3d\": [[405.32, 204.05], [380.76, 204.04], [395.06, 202.14], [418.13, 202.15], [405.32, 185.16], [380.76, 185.16], [395.06, 184.41], [418.13, 184.42]]}]\n```", - "options": null, - "id": 251 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000984", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[399.71, 204.2], [373.62, 204.14], [395.13, 201.66], [419.19, 201.71], [399.71, 183.16], [373.62, 183.14], [395.13, 182.32], [419.19, 182.34]]}]\n```", - "options": null, - "id": 252 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000984", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[205.24, 215.8], [165.1, 215.75], [223.54, 210.28], [258.61, 210.32], [205.24, 169.67], [165.1, 169.68], [223.54, 170.08], [258.61, 170.08]]}, {\"category\": \"van\", \"corners_3d\": [[350.03, 200.68], [325.51, 200.63], [353.91, 198.16], [376.29, 198.2], [350.03, 174.85], [325.51, 174.84], [353.91, 174.67], [376.29, 174.67]]}]\n```", - "options": null, - "id": 253 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000984", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[205.24, 215.8], [165.1, 215.75], [223.54, 210.28], [258.61, 210.32], [205.24, 169.67], [165.1, 169.68], [223.54, 170.08], [258.61, 170.08]]}, {\"category\": \"car\", \"corners_3d\": [[399.71, 204.2], [373.62, 204.14], [395.13, 201.66], [419.19, 201.71], [399.71, 183.16], [373.62, 183.14], [395.13, 182.32], [419.19, 182.34]]}, {\"category\": \"van\", \"corners_3d\": [[350.03, 200.68], [325.51, 200.63], [353.91, 198.16], [376.29, 198.2], [350.03, 174.85], [325.51, 174.84], [353.91, 174.67], [376.29, 174.67]]}]\n```", - "options": null, - "id": 254 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000994", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[148.03, 222.64], [107.38, 222.65], [166.35, 216.79], [202.22, 216.79], [148.03, 183.5], [107.38, 183.5], [166.35, 182.25], [202.22, 182.25]]}, {\"category\": \"car\", \"corners_3d\": [[222.09, 214.6], [184.78, 214.54], [232.81, 210.05], [266.16, 210.1], [222.09, 184.6], [184.78, 184.58], [232.81, 183.32], [266.16, 183.33]]}, {\"category\": \"car\", \"corners_3d\": [[291.82, 207.82], [263.76, 207.75], [299.24, 204.51], [324.75, 204.56], [291.82, 182.5], [263.76, 182.48], [299.24, 181.58], [324.75, 181.6]]}, {\"category\": \"car\", \"corners_3d\": [[332.07, 202.61], [310.79, 202.56], [334.1, 200.49], [353.93, 200.53], [332.07, 180.25], [310.79, 180.24], [334.1, 179.72], [353.93, 179.73]]}, {\"category\": \"car\", \"corners_3d\": [[365.97, 199.59], [343.7, 199.54], [362.09, 197.95], [383.07, 198.0], [365.97, 180.32], [343.7, 180.3], [362.09, 179.86], [383.07, 179.87]]}, {\"category\": \"car\", \"corners_3d\": [[411.67, 196.44], [395.65, 196.38], [409.77, 195.22], [425.03, 195.27], [411.67, 178.72], [395.65, 178.71], [409.77, 178.42], [425.03, 178.43]]}]\n```", - "options": null, - "id": 255 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000994", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[648.25, 199.96], [693.07, 200.15], [694.04, 208.35], [635.87, 208.03], [648.25, 145.73], [693.07, 145.55], [694.04, 137.35], [635.87, 137.66]]}]\n```", - "options": null, - "id": 256 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000994", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[648.25, 199.96], [693.07, 200.15], [694.04, 208.35], [635.87, 208.03], [648.25, 145.73], [693.07, 145.55], [694.04, 137.35], [635.87, 137.66]]}, {\"category\": \"car\", \"corners_3d\": [[148.03, 222.64], [107.38, 222.65], [166.35, 216.79], [202.22, 216.79], [148.03, 183.5], [107.38, 183.5], [166.35, 182.25], [202.22, 182.25]]}, {\"category\": \"car\", \"corners_3d\": [[222.09, 214.6], [184.78, 214.54], [232.81, 210.05], [266.16, 210.1], [222.09, 184.6], [184.78, 184.58], [232.81, 183.32], [266.16, 183.33]]}, {\"category\": \"car\", \"corners_3d\": [[291.82, 207.82], [263.76, 207.75], [299.24, 204.51], [324.75, 204.56], [291.82, 182.5], [263.76, 182.48], [299.24, 181.58], [324.75, 181.6]]}, {\"category\": \"car\", \"corners_3d\": [[332.07, 202.61], [310.79, 202.56], [334.1, 200.49], [353.93, 200.53], [332.07, 180.25], [310.79, 180.24], [334.1, 179.72], [353.93, 179.73]]}, {\"category\": \"car\", \"corners_3d\": [[365.97, 199.59], [343.7, 199.54], [362.09, 197.95], [383.07, 198.0], [365.97, 180.32], [343.7, 180.3], [362.09, 179.86], [383.07, 179.87]]}, {\"category\": \"car\", \"corners_3d\": [[411.67, 196.44], [395.65, 196.38], [409.77, 195.22], [425.03, 195.27], [411.67, 178.72], [395.65, 178.71], [409.77, 178.42], [425.03, 178.43]]}]\n```", - "options": null, - "id": 257 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000995", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[208.57, 275.98], [91.69, 275.68], [266.33, 244.2], [344.43, 244.34], [208.57, 127.84], [91.69, 128.0], [266.33, 145.89], [344.43, 145.81]]}]\n```", - "options": null, - "id": 258 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000995", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[309.1, 299.07], [426.46, 299.28], [325.14, 364.96], [142.79, 364.46], [309.1, 191.33], [426.46, 191.35], [325.14, 196.84], [142.79, 196.79]]}, {\"category\": \"car\", \"corners_3d\": [[523.92, 228.7], [584.15, 228.55], [587.5, 241.74], [510.32, 241.99], [523.92, 173.88], [584.15, 173.91], [587.5, 171.78], [510.32, 171.74]]}, {\"category\": \"car\", \"corners_3d\": [[520.22, 214.82], [563.95, 214.89], [554.4, 220.33], [503.56, 220.24], [520.22, 171.82], [563.95, 171.81], [554.4, 170.22], [503.56, 170.25]]}, {\"category\": \"car\", \"corners_3d\": [[598.87, 205.72], [633.48, 205.72], [636.04, 207.63], [598.71, 207.63], [598.87, 170.89], [633.48, 170.88], [636.04, 170.05], [598.71, 170.05]]}, {\"category\": \"car\", \"corners_3d\": [[490.7, 202.33], [462.49, 202.28], [477.06, 200.6], [503.01, 200.64], [490.7, 174.14], [462.49, 174.16], [477.06, 174.75], [503.01, 174.73]]}]\n```", - "options": null, - "id": 259 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000995", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[309.1, 299.07], [426.46, 299.28], [325.14, 364.96], [142.79, 364.46], [309.1, 191.33], [426.46, 191.35], [325.14, 196.84], [142.79, 196.79]]}, {\"category\": \"car\", \"corners_3d\": [[523.92, 228.7], [584.15, 228.55], [587.5, 241.74], [510.32, 241.99], [523.92, 173.88], [584.15, 173.91], [587.5, 171.78], [510.32, 171.74]]}, {\"category\": \"car\", \"corners_3d\": [[520.22, 214.82], [563.95, 214.89], [554.4, 220.33], [503.56, 220.24], [520.22, 171.82], [563.95, 171.81], [554.4, 170.22], [503.56, 170.25]]}, {\"category\": \"van\", \"corners_3d\": [[208.57, 275.98], [91.69, 275.68], [266.33, 244.2], [344.43, 244.34], [208.57, 127.84], [91.69, 128.0], [266.33, 145.89], [344.43, 145.81]]}, {\"category\": \"car\", \"corners_3d\": [[598.87, 205.72], [633.48, 205.72], [636.04, 207.63], [598.71, 207.63], [598.87, 170.89], [633.48, 170.88], [636.04, 170.05], [598.71, 170.05]]}, {\"category\": \"car\", \"corners_3d\": [[490.7, 202.33], [462.49, 202.28], [477.06, 200.6], [503.01, 200.64], [490.7, 174.14], [462.49, 174.16], [477.06, 174.75], [503.01, 174.73]]}]\n```", - "options": null, - "id": 260 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001007", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-270.75, 340.08], [-109.01, 309.89], [166.42, 311.04], [64.32, 341.8], [-270.75, 176.95], [-109.01, 176.21], [166.42, 176.24], [64.32, 176.99]]}, {\"category\": \"car\", \"corners_3d\": [[327.03, 280.64], [274.7, 300.1], [-32.71, 299.56], [66.29, 280.24], [327.03, 172.19], [274.7, 172.06], [-32.71, 172.07], [66.29, 172.19]]}, {\"category\": \"car\", \"corners_3d\": [[86.94, 273.25], [148.39, 260.54], [384.26, 259.1], [357.51, 271.37], [86.94, 178.99], [148.39, 178.22], [384.26, 178.13], [357.51, 178.88]]}, {\"category\": \"car\", \"corners_3d\": [[464.38, 233.15], [448.76, 239.33], [275.03, 239.17], [306.74, 233.02], [464.38, 175.1], [448.76, 175.33], [275.03, 175.33], [306.74, 175.1]]}, {\"category\": \"car\", \"corners_3d\": [[516.45, 213.53], [512.94, 216.04], [393.27, 216.4], [403.81, 213.85], [516.45, 175.62], [512.94, 175.8], [393.27, 175.82], [403.81, 175.65]]}, {\"category\": \"car\", \"corners_3d\": [[657.19, 215.86], [698.43, 215.86], [713.75, 223.32], [665.36, 223.32], [657.19, 179.97], [698.43, 179.97], [713.75, 181.21], [665.36, 181.21]]}]\n```", - "options": null, - "id": 261 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001007", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[726.77, 269.44], [844.75, 269.46], [1126.35, 385.41], [866.72, 385.35], [726.77, 117.12], [844.75, 117.11], [1126.35, 50.2], [866.72, 50.24]]}, {\"category\": \"van\", \"corners_3d\": [[526.94, 214.08], [521.17, 216.73], [378.39, 216.65], [392.75, 214.01], [526.94, 147.82], [521.17, 146.22], [378.39, 146.27], [392.75, 147.87]]}]\n```", - "options": null, - "id": 262 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001007", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[726.77, 269.44], [844.75, 269.46], [1126.35, 385.41], [866.72, 385.35], [726.77, 117.12], [844.75, 117.11], [1126.35, 50.2], [866.72, 50.24]]}, {\"category\": \"car\", \"corners_3d\": [[-270.75, 340.08], [-109.01, 309.89], [166.42, 311.04], [64.32, 341.8], [-270.75, 176.95], [-109.01, 176.21], [166.42, 176.24], [64.32, 176.99]]}, {\"category\": \"car\", \"corners_3d\": [[327.03, 280.64], [274.7, 300.1], [-32.71, 299.56], [66.29, 280.24], [327.03, 172.19], [274.7, 172.06], [-32.71, 172.07], [66.29, 172.19]]}, {\"category\": \"car\", \"corners_3d\": [[86.94, 273.25], [148.39, 260.54], [384.26, 259.1], [357.51, 271.37], [86.94, 178.99], [148.39, 178.22], [384.26, 178.13], [357.51, 178.88]]}, {\"category\": \"car\", \"corners_3d\": [[464.38, 233.15], [448.76, 239.33], [275.03, 239.17], [306.74, 233.02], [464.38, 175.1], [448.76, 175.33], [275.03, 175.33], [306.74, 175.1]]}, {\"category\": \"car\", \"corners_3d\": [[516.45, 213.53], [512.94, 216.04], [393.27, 216.4], [403.81, 213.85], [516.45, 175.62], [512.94, 175.8], [393.27, 175.82], [403.81, 175.65]]}, {\"category\": \"van\", \"corners_3d\": [[526.94, 214.08], [521.17, 216.73], [378.39, 216.65], [392.75, 214.01], [526.94, 147.82], [521.17, 146.22], [378.39, 146.27], [392.75, 147.87]]}, {\"category\": \"car\", \"corners_3d\": [[657.19, 215.86], [698.43, 215.86], [713.75, 223.32], [665.36, 223.32], [657.19, 179.97], [698.43, 179.97], [713.75, 181.21], [665.36, 181.21]]}]\n```", - "options": null, - "id": 263 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001079", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[642.22, 212.82], [676.85, 212.66], [689.78, 216.8], [651.59, 216.98], [642.22, 178.85], [676.85, 178.83], [689.78, 179.45], [651.59, 179.48]]}]\n```", - "options": null, - "id": 264 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001090", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[572.44, 219.88], [542.13, 220.1], [540.96, 215.81], [568.54, 215.63], [572.44, 188.77], [542.13, 188.85], [540.96, 187.39], [568.54, 187.33]]}]\n```", - "options": null, - "id": 265 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001098", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[564.67, 220.18], [611.63, 220.0], [618.85, 227.59], [564.33, 227.83], [564.67, 174.83], [611.63, 174.82], [618.85, 175.14], [564.33, 175.15]]}]\n```", - "options": null, - "id": 266 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000992", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[834.04, 243.97], [870.93, 243.97], [879.59, 246.34], [841.47, 246.33], [834.04, 155.97], [870.93, 155.97], [879.59, 155.41], [841.47, 155.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[628.61, 194.97], [624.52, 195.12], [611.82, 194.81], [615.94, 194.65], [628.61, 165.16], [624.52, 165.11], [611.82, 165.22], [615.94, 165.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[635.08, 196.13], [630.19, 196.33], [619.38, 196.03], [624.3, 195.84], [635.08, 167.1], [630.19, 167.06], [619.38, 167.13], [624.3, 167.18]]}]\n```", - "options": null, - "id": 267 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000992", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[437.55, 277.91], [404.02, 277.96], [428.16, 265.13], [457.59, 265.09], [437.55, 175.82], [404.02, 175.82], [428.16, 175.46], [457.59, 175.46]]}, {\"category\": \"cyclist\", \"corners_3d\": [[482.85, 238.91], [463.56, 238.93], [474.91, 233.5], [492.62, 233.49], [482.85, 170.58], [463.56, 170.57], [474.91, 170.76], [492.62, 170.76]]}]\n```", - "options": null, - "id": 268 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000992", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[695.12, 244.46], [762.18, 244.2], [806.06, 261.5], [722.87, 261.91], [695.12, 175.47], [762.18, 175.47], [806.06, 176.1], [722.87, 176.11]]}, {\"category\": \"car\", \"corners_3d\": [[664.15, 226.35], [716.83, 226.27], [742.78, 237.71], [678.84, 237.82], [664.15, 168.88], [716.83, 168.88], [742.78, 168.03], [678.84, 168.03]]}, {\"category\": \"car\", \"corners_3d\": [[651.17, 215.33], [695.23, 215.33], [711.48, 223.45], [659.01, 223.44], [651.17, 167.63], [695.23, 167.63], [711.48, 166.63], [659.01, 166.63]]}, {\"category\": \"car\", \"corners_3d\": [[553.17, 193.81], [532.58, 193.82], [537.08, 192.45], [556.33, 192.45], [553.17, 169.44], [532.58, 169.44], [537.08, 169.66], [556.33, 169.66]]}, {\"category\": \"car\", \"corners_3d\": [[504.61, 210.12], [472.15, 210.14], [485.4, 206.33], [514.53, 206.31], [504.61, 176.84], [472.15, 176.84], [485.4, 176.43], [514.53, 176.43]]}, {\"category\": \"car\", \"corners_3d\": [[261.68, 319.83], [131.64, 319.58], [284.31, 273.35], [373.51, 273.47], [261.68, 201.3], [131.64, 201.26], [284.31, 192.31], [373.51, 192.33]]}, {\"category\": \"car\", \"corners_3d\": [[389.65, 262.32], [314.63, 262.24], [375.3, 244.25], [435.26, 244.31], [389.65, 187.02], [314.63, 187.0], [375.3, 184.16], [435.26, 184.17]]}]\n```", - "options": null, - "id": 269 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000992", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[834.04, 243.97], [870.93, 243.97], [879.59, 246.34], [841.47, 246.33], [834.04, 155.97], [870.93, 155.97], [879.59, 155.41], [841.47, 155.41]]}, {\"category\": \"cyclist\", \"corners_3d\": [[437.55, 277.91], [404.02, 277.96], [428.16, 265.13], [457.59, 265.09], [437.55, 175.82], [404.02, 175.82], [428.16, 175.46], [457.59, 175.46]]}, {\"category\": \"cyclist\", \"corners_3d\": [[482.85, 238.91], [463.56, 238.93], [474.91, 233.5], [492.62, 233.49], [482.85, 170.58], [463.56, 170.57], [474.91, 170.76], [492.62, 170.76]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[628.61, 194.97], [624.52, 195.12], [611.82, 194.81], [615.94, 194.65], [628.61, 165.16], [624.52, 165.11], [611.82, 165.22], [615.94, 165.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[635.08, 196.13], [630.19, 196.33], [619.38, 196.03], [624.3, 195.84], [635.08, 167.1], [630.19, 167.06], [619.38, 167.13], [624.3, 167.18]]}]\n```", - "options": null, - "id": 270 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000992", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[834.04, 243.97], [870.93, 243.97], [879.59, 246.34], [841.47, 246.33], [834.04, 155.97], [870.93, 155.97], [879.59, 155.41], [841.47, 155.41]]}, {\"category\": \"car\", \"corners_3d\": [[695.12, 244.46], [762.18, 244.2], [806.06, 261.5], [722.87, 261.91], [695.12, 175.47], [762.18, 175.47], [806.06, 176.1], [722.87, 176.11]]}, {\"category\": \"car\", \"corners_3d\": [[664.15, 226.35], [716.83, 226.27], [742.78, 237.71], [678.84, 237.82], [664.15, 168.88], [716.83, 168.88], [742.78, 168.03], [678.84, 168.03]]}, {\"category\": \"car\", \"corners_3d\": [[651.17, 215.33], [695.23, 215.33], [711.48, 223.45], [659.01, 223.44], [651.17, 167.63], [695.23, 167.63], [711.48, 166.63], [659.01, 166.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[628.61, 194.97], [624.52, 195.12], [611.82, 194.81], [615.94, 194.65], [628.61, 165.16], [624.52, 165.11], [611.82, 165.22], [615.94, 165.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[635.08, 196.13], [630.19, 196.33], [619.38, 196.03], [624.3, 195.84], [635.08, 167.1], [630.19, 167.06], [619.38, 167.13], [624.3, 167.18]]}, {\"category\": \"car\", \"corners_3d\": [[553.17, 193.81], [532.58, 193.82], [537.08, 192.45], [556.33, 192.45], [553.17, 169.44], [532.58, 169.44], [537.08, 169.66], [556.33, 169.66]]}, {\"category\": \"car\", \"corners_3d\": [[504.61, 210.12], [472.15, 210.14], [485.4, 206.33], [514.53, 206.31], [504.61, 176.84], [472.15, 176.84], [485.4, 176.43], [514.53, 176.43]]}, {\"category\": \"car\", \"corners_3d\": [[261.68, 319.83], [131.64, 319.58], [284.31, 273.35], [373.51, 273.47], [261.68, 201.3], [131.64, 201.26], [284.31, 192.31], [373.51, 192.33]]}, {\"category\": \"car\", \"corners_3d\": [[389.65, 262.32], [314.63, 262.24], [375.3, 244.25], [435.26, 244.31], [389.65, 187.02], [314.63, 187.0], [375.3, 184.16], [435.26, 184.17]]}]\n```", - "options": null, - "id": 271 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000992", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[437.55, 277.91], [404.02, 277.96], [428.16, 265.13], [457.59, 265.09], [437.55, 175.82], [404.02, 175.82], [428.16, 175.46], [457.59, 175.46]]}, {\"category\": \"cyclist\", \"corners_3d\": [[482.85, 238.91], [463.56, 238.93], [474.91, 233.5], [492.62, 233.49], [482.85, 170.58], [463.56, 170.57], [474.91, 170.76], [492.62, 170.76]]}, {\"category\": \"car\", \"corners_3d\": [[695.12, 244.46], [762.18, 244.2], [806.06, 261.5], [722.87, 261.91], [695.12, 175.47], [762.18, 175.47], [806.06, 176.1], [722.87, 176.11]]}, {\"category\": \"car\", \"corners_3d\": [[664.15, 226.35], [716.83, 226.27], [742.78, 237.71], [678.84, 237.82], [664.15, 168.88], [716.83, 168.88], [742.78, 168.03], [678.84, 168.03]]}, {\"category\": \"car\", \"corners_3d\": [[651.17, 215.33], [695.23, 215.33], [711.48, 223.45], [659.01, 223.44], [651.17, 167.63], [695.23, 167.63], [711.48, 166.63], [659.01, 166.63]]}, {\"category\": \"car\", \"corners_3d\": [[553.17, 193.81], [532.58, 193.82], [537.08, 192.45], [556.33, 192.45], [553.17, 169.44], [532.58, 169.44], [537.08, 169.66], [556.33, 169.66]]}, {\"category\": \"car\", \"corners_3d\": [[504.61, 210.12], [472.15, 210.14], [485.4, 206.33], [514.53, 206.31], [504.61, 176.84], [472.15, 176.84], [485.4, 176.43], [514.53, 176.43]]}, {\"category\": \"car\", \"corners_3d\": [[261.68, 319.83], [131.64, 319.58], [284.31, 273.35], [373.51, 273.47], [261.68, 201.3], [131.64, 201.26], [284.31, 192.31], [373.51, 192.33]]}, {\"category\": \"car\", \"corners_3d\": [[389.65, 262.32], [314.63, 262.24], [375.3, 244.25], [435.26, 244.31], [389.65, 187.02], [314.63, 187.0], [375.3, 184.16], [435.26, 184.17]]}]\n```", - "options": null, - "id": 272 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "000992", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[834.04, 243.97], [870.93, 243.97], [879.59, 246.34], [841.47, 246.33], [834.04, 155.97], [870.93, 155.97], [879.59, 155.41], [841.47, 155.41]]}, {\"category\": \"cyclist\", \"corners_3d\": [[437.55, 277.91], [404.02, 277.96], [428.16, 265.13], [457.59, 265.09], [437.55, 175.82], [404.02, 175.82], [428.16, 175.46], [457.59, 175.46]]}, {\"category\": \"cyclist\", \"corners_3d\": [[482.85, 238.91], [463.56, 238.93], [474.91, 233.5], [492.62, 233.49], [482.85, 170.58], [463.56, 170.57], [474.91, 170.76], [492.62, 170.76]]}, {\"category\": \"car\", \"corners_3d\": [[695.12, 244.46], [762.18, 244.2], [806.06, 261.5], [722.87, 261.91], [695.12, 175.47], [762.18, 175.47], [806.06, 176.1], [722.87, 176.11]]}, {\"category\": \"car\", \"corners_3d\": [[664.15, 226.35], [716.83, 226.27], [742.78, 237.71], [678.84, 237.82], [664.15, 168.88], [716.83, 168.88], [742.78, 168.03], [678.84, 168.03]]}, {\"category\": \"car\", \"corners_3d\": [[651.17, 215.33], [695.23, 215.33], [711.48, 223.45], [659.01, 223.44], [651.17, 167.63], [695.23, 167.63], [711.48, 166.63], [659.01, 166.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[628.61, 194.97], [624.52, 195.12], [611.82, 194.81], [615.94, 194.65], [628.61, 165.16], [624.52, 165.11], [611.82, 165.22], [615.94, 165.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[635.08, 196.13], [630.19, 196.33], [619.38, 196.03], [624.3, 195.84], [635.08, 167.1], [630.19, 167.06], [619.38, 167.13], [624.3, 167.18]]}, {\"category\": \"car\", \"corners_3d\": [[553.17, 193.81], [532.58, 193.82], [537.08, 192.45], [556.33, 192.45], [553.17, 169.44], [532.58, 169.44], [537.08, 169.66], [556.33, 169.66]]}, {\"category\": \"car\", \"corners_3d\": [[504.61, 210.12], [472.15, 210.14], [485.4, 206.33], [514.53, 206.31], [504.61, 176.84], [472.15, 176.84], [485.4, 176.43], [514.53, 176.43]]}, {\"category\": \"car\", \"corners_3d\": [[261.68, 319.83], [131.64, 319.58], [284.31, 273.35], [373.51, 273.47], [261.68, 201.3], [131.64, 201.26], [284.31, 192.31], [373.51, 192.33]]}, {\"category\": \"car\", \"corners_3d\": [[389.65, 262.32], [314.63, 262.24], [375.3, 244.25], [435.26, 244.31], [389.65, 187.02], [314.63, 187.0], [375.3, 184.16], [435.26, 184.17]]}]\n```", - "options": null, - "id": 273 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001117", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[630.95, 194.08], [649.78, 194.06], [653.97, 195.35], [634.0, 195.38], [630.95, 175.58], [649.78, 175.58], [653.97, 175.75], [634.0, 175.75]]}, {\"category\": \"car\", \"corners_3d\": [[668.06, 219.95], [710.66, 219.87], [729.23, 227.01], [680.17, 227.12], [668.06, 180.31], [710.66, 180.3], [729.23, 181.43], [680.17, 181.45]]}, {\"category\": \"car\", \"corners_3d\": [[764.29, 267.98], [856.73, 267.62], [977.0, 310.11], [843.55, 310.86], [764.29, 181.48], [856.73, 181.45], [977.0, 185.3], [843.55, 185.37]]}, {\"category\": \"car\", \"corners_3d\": [[729.49, 204.21], [759.29, 204.11], [783.89, 207.83], [750.63, 207.96], [729.49, 176.16], [759.29, 176.15], [783.89, 176.54], [750.63, 176.55]]}]\n```", - "options": null, - "id": 274 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001117", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[855.87, 240.23], [959.54, 239.95], [1265.33, 295.24], [1077.95, 296.19], [855.87, 94.35], [959.54, 94.68], [1265.33, 30.26], [1077.95, 29.15]]}]\n```", - "options": null, - "id": 275 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001117", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[630.95, 194.08], [649.78, 194.06], [653.97, 195.35], [634.0, 195.38], [630.95, 175.58], [649.78, 175.58], [653.97, 175.75], [634.0, 175.75]]}, {\"category\": \"car\", \"corners_3d\": [[668.06, 219.95], [710.66, 219.87], [729.23, 227.01], [680.17, 227.12], [668.06, 180.31], [710.66, 180.3], [729.23, 181.43], [680.17, 181.45]]}, {\"category\": \"car\", \"corners_3d\": [[764.29, 267.98], [856.73, 267.62], [977.0, 310.11], [843.55, 310.86], [764.29, 181.48], [856.73, 181.45], [977.0, 185.3], [843.55, 185.37]]}, {\"category\": \"car\", \"corners_3d\": [[729.49, 204.21], [759.29, 204.11], [783.89, 207.83], [750.63, 207.96], [729.49, 176.16], [759.29, 176.15], [783.89, 176.54], [750.63, 176.55]]}, {\"category\": \"truck\", \"corners_3d\": [[855.87, 240.23], [959.54, 239.95], [1265.33, 295.24], [1077.95, 296.19], [855.87, 94.35], [959.54, 94.68], [1265.33, 30.26], [1077.95, 29.15]]}]\n```", - "options": null, - "id": 276 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001153", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[624.0, 311.39], [744.84, 329.51], [520.4, 409.91], [384.88, 370.74], [624.0, 174.53], [744.84, 174.75], [520.4, 175.73], [384.88, 175.25]]}, {\"category\": \"car\", \"corners_3d\": [[801.94, 315.42], [689.36, 300.39], [812.32, 269.93], [906.38, 278.4], [801.94, 175.5], [689.36, 175.22], [812.32, 174.65], [906.38, 174.81]]}, {\"category\": \"car\", \"corners_3d\": [[942.7, 242.93], [1018.42, 247.46], [974.38, 262.93], [887.29, 256.41], [942.7, 181.23], [1018.42, 181.77], [974.38, 183.62], [887.29, 182.85]]}, {\"category\": \"car\", \"corners_3d\": [[1095.78, 219.97], [1049.21, 218.04], [1079.97, 214.22], [1123.65, 215.84], [1095.78, 176.67], [1049.21, 176.51], [1079.97, 176.2], [1123.65, 176.33]]}, {\"category\": \"car\", \"corners_3d\": [[1241.03, 250.65], [1326.95, 255.46], [1330.37, 264.73], [1235.21, 258.81], [1241.03, 189.7], [1326.95, 190.75], [1330.37, 192.75], [1235.21, 191.47]]}]\n```", - "options": null, - "id": 277 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001155", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[649.28, 218.25], [687.97, 218.21], [704.01, 226.09], [658.6, 226.16], [649.28, 183.64], [687.97, 183.63], [704.01, 185.51], [658.6, 185.52]]}]\n```", - "options": null, - "id": 278 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001166", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[393.2, 345.69], [222.55, 346.54], [359.4, 282.64], [467.05, 282.31], [393.2, 201.29], [222.55, 201.43], [359.4, 190.92], [467.05, 190.86]]}, {\"category\": \"car\", \"corners_3d\": [[576.47, 185.23], [550.69, 185.24], [553.61, 184.26], [577.33, 184.25], [576.47, 158.13], [550.69, 158.12], [553.61, 159.29], [577.33, 159.3]]}]\n```", - "options": null, - "id": 279 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001178", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[799.73, 185.68], [829.54, 185.93], [809.76, 187.34], [777.24, 187.02], [799.73, 147.49], [829.54, 146.98], [809.76, 144.2], [777.24, 144.83]]}]\n```", - "options": null, - "id": 280 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001178", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[754.97, 241.77], [837.27, 242.09], [900.63, 264.21], [791.79, 263.65], [754.97, 173.33], [837.27, 173.33], [900.63, 173.49], [791.79, 173.48]]}, {\"category\": \"car\", \"corners_3d\": [[741.06, 225.84], [796.6, 226.08], [833.86, 239.94], [763.72, 239.55], [741.06, 171.41], [796.6, 171.4], [833.86, 171.02], [763.72, 171.03]]}, {\"category\": \"car\", \"corners_3d\": [[739.68, 210.87], [788.87, 211.25], [805.13, 220.21], [744.4, 219.63], [739.68, 161.27], [788.87, 161.15], [805.13, 158.42], [744.4, 158.6]]}]\n```", - "options": null, - "id": 281 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001178", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[754.97, 241.77], [837.27, 242.09], [900.63, 264.21], [791.79, 263.65], [754.97, 173.33], [837.27, 173.33], [900.63, 173.49], [791.79, 173.48]]}, {\"category\": \"car\", \"corners_3d\": [[741.06, 225.84], [796.6, 226.08], [833.86, 239.94], [763.72, 239.55], [741.06, 171.41], [796.6, 171.4], [833.86, 171.02], [763.72, 171.03]]}, {\"category\": \"car\", \"corners_3d\": [[739.68, 210.87], [788.87, 211.25], [805.13, 220.21], [744.4, 219.63], [739.68, 161.27], [788.87, 161.15], [805.13, 158.42], [744.4, 158.6]]}, {\"category\": \"van\", \"corners_3d\": [[799.73, 185.68], [829.54, 185.93], [809.76, 187.34], [777.24, 187.02], [799.73, 147.49], [829.54, 146.98], [809.76, 144.2], [777.24, 144.83]]}]\n```", - "options": null, - "id": 282 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001202", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[573.27, 199.71], [550.6, 199.68], [556.75, 197.79], [577.83, 197.82], [573.27, 180.66], [550.6, 180.65], [556.75, 180.1], [577.83, 180.11]]}]\n```", - "options": null, - "id": 283 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001212", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[591.58, 199.47], [630.05, 199.47], [635.01, 206.1], [586.96, 206.1], [591.58, 152.62], [630.05, 152.62], [635.01, 147.58], [586.96, 147.58]]}]\n```", - "options": null, - "id": 284 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001212", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[243.49, 232.97], [203.24, 232.94], [250.73, 226.03], [286.37, 226.06], [243.49, 199.52], [203.24, 199.51], [250.73, 196.44], [286.37, 196.45]]}]\n```", - "options": null, - "id": 285 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001212", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[591.58, 199.47], [630.05, 199.47], [635.01, 206.1], [586.96, 206.1], [591.58, 152.62], [630.05, 152.62], [635.01, 147.58], [586.96, 147.58]]}, {\"category\": \"car\", \"corners_3d\": [[243.49, 232.97], [203.24, 232.94], [250.73, 226.03], [286.37, 226.06], [243.49, 199.52], [203.24, 199.51], [250.73, 196.44], [286.37, 196.45]]}]\n```", - "options": null, - "id": 286 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001219", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[563.02, 198.0], [543.56, 198.01], [546.9, 196.36], [565.09, 196.35], [563.02, 179.58], [543.56, 179.58], [546.9, 179.14], [565.09, 179.14]]}, {\"category\": \"car\", \"corners_3d\": [[535.65, 215.41], [501.47, 215.39], [513.85, 210.8], [544.35, 210.82], [535.65, 179.21], [501.47, 179.21], [513.85, 178.52], [544.35, 178.52]]}]\n```", - "options": null, - "id": 287 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001256", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[355.98, 293.15], [336.91, 300.37], [279.77, 299.67], [302.15, 292.53], [355.98, 159.04], [336.91, 157.66], [279.77, 157.79], [302.15, 159.15]]}]\n```", - "options": null, - "id": 288 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001261", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1005.71, 319.14], [851.88, 319.17], [767.07, 268.08], [867.2, 268.07], [1005.71, 152.05], [851.88, 152.05], [767.07, 159.31], [867.2, 159.31]]}, {\"category\": \"car\", \"corners_3d\": [[678.6, 212.22], [721.58, 212.19], [743.03, 219.3], [692.29, 219.33], [678.6, 167.86], [721.58, 167.86], [743.03, 166.96], [692.29, 166.95]]}, {\"category\": \"car\", \"corners_3d\": [[543.68, 215.84], [503.27, 215.8], [516.77, 210.97], [552.65, 211.01], [543.68, 177.54], [503.27, 177.54], [516.77, 177.01], [552.65, 177.01]]}, {\"category\": \"car\", \"corners_3d\": [[578.76, 200.52], [608.07, 200.63], [599.58, 203.8], [566.99, 203.66], [578.76, 173.99], [608.07, 173.99], [599.58, 174.12], [566.99, 174.12]]}]\n```", - "options": null, - "id": 289 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001312", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[564.18, 193.18], [547.16, 193.19], [549.89, 192.02], [565.94, 192.01], [564.18, 177.07], [547.16, 177.07], [549.89, 176.83], [565.94, 176.83]]}, {\"category\": \"car\", \"corners_3d\": [[546.42, 204.78], [519.2, 204.78], [527.08, 201.98], [551.91, 201.98], [546.42, 175.99], [519.2, 175.99], [527.08, 175.71], [551.91, 175.71]]}]\n```", - "options": null, - "id": 290 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001324", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[475.42, 246.28], [407.06, 247.17], [426.55, 233.7], [482.46, 233.1], [475.42, 177.74], [407.06, 177.79], [426.55, 176.9], [482.46, 176.86]]}, {\"category\": \"car\", \"corners_3d\": [[481.66, 211.83], [442.42, 212.43], [436.68, 207.77], [471.44, 207.3], [481.66, 178.7], [442.42, 178.79], [436.68, 178.1], [471.44, 178.02]]}]\n```", - "options": null, - "id": 291 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001345", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[421.97, 201.07], [399.29, 201.07], [413.3, 199.18], [434.47, 199.18], [421.97, 181.87], [399.29, 181.87], [413.3, 181.27], [434.47, 181.27]]}]\n```", - "options": null, - "id": 292 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001345", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[-159.45, 338.02], [-418.07, 338.65], [294.94, 222.74], [372.14, 222.68], [-159.45, 30.69], [-418.07, 30.15], [294.94, 129.92], [372.14, 129.97]]}]\n```", - "options": null, - "id": 293 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001345", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[898.88, 244.98], [924.09, 244.98], [960.56, 253.36], [932.43, 253.35], [898.88, 158.51], [924.09, 158.51], [960.56, 156.84], [932.43, 156.84]]}]\n```", - "options": null, - "id": 294 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001345", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[-159.45, 338.02], [-418.07, 338.65], [294.94, 222.74], [372.14, 222.68], [-159.45, 30.69], [-418.07, 30.15], [294.94, 129.92], [372.14, 129.97]]}, {\"category\": \"car\", \"corners_3d\": [[421.97, 201.07], [399.29, 201.07], [413.3, 199.18], [434.47, 199.18], [421.97, 181.87], [399.29, 181.87], [413.3, 181.27], [434.47, 181.27]]}]\n```", - "options": null, - "id": 295 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001345", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[898.88, 244.98], [924.09, 244.98], [960.56, 253.36], [932.43, 253.35], [898.88, 158.51], [924.09, 158.51], [960.56, 156.84], [932.43, 156.84]]}, {\"category\": \"car\", \"corners_3d\": [[421.97, 201.07], [399.29, 201.07], [413.3, 199.18], [434.47, 199.18], [421.97, 181.87], [399.29, 181.87], [413.3, 181.27], [434.47, 181.27]]}]\n```", - "options": null, - "id": 296 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001345", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (tram, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[-159.45, 338.02], [-418.07, 338.65], [294.94, 222.74], [372.14, 222.68], [-159.45, 30.69], [-418.07, 30.15], [294.94, 129.92], [372.14, 129.97]]}, {\"category\": \"cyclist\", \"corners_3d\": [[898.88, 244.98], [924.09, 244.98], [960.56, 253.36], [932.43, 253.35], [898.88, 158.51], [924.09, 158.51], [960.56, 156.84], [932.43, 156.84]]}]\n```", - "options": null, - "id": 297 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001345", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, tram, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[-159.45, 338.02], [-418.07, 338.65], [294.94, 222.74], [372.14, 222.68], [-159.45, 30.69], [-418.07, 30.15], [294.94, 129.92], [372.14, 129.97]]}, {\"category\": \"cyclist\", \"corners_3d\": [[898.88, 244.98], [924.09, 244.98], [960.56, 253.36], [932.43, 253.35], [898.88, 158.51], [924.09, 158.51], [960.56, 156.84], [932.43, 156.84]]}, {\"category\": \"car\", \"corners_3d\": [[421.97, 201.07], [399.29, 201.07], [413.3, 199.18], [434.47, 199.18], [421.97, 181.87], [399.29, 181.87], [413.3, 181.27], [434.47, 181.27]]}]\n```", - "options": null, - "id": 298 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001348", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[393.41, 350.8], [406.33, 334.92], [525.51, 332.34], [524.85, 347.66], [393.41, 145.99], [406.33, 149.21], [525.51, 149.73], [524.85, 146.63]]}]\n```", - "options": null, - "id": 299 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001358", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1219.28, 276.54], [1359.61, 298.19], [987.28, 298.82], [915.15, 276.95], [1219.28, 170.67], [1359.61, 168.45], [987.28, 168.38], [915.15, 170.62]]}, {\"category\": \"car\", \"corners_3d\": [[827.49, 206.35], [841.22, 207.99], [719.32, 207.95], [712.88, 206.31], [827.49, 169.85], [841.22, 169.17], [719.32, 169.19], [712.88, 169.86]]}]\n```", - "options": null, - "id": 300 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001358", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[656.5, 245.58], [682.18, 245.56], [692.27, 253.32], [663.53, 253.34], [656.5, 165.36], [682.18, 165.36], [692.27, 163.56], [663.53, 163.55]]}, {\"category\": \"cyclist\", \"corners_3d\": [[561.56, 207.55], [547.95, 207.54], [551.86, 206.16], [564.79, 206.17], [561.56, 168.83], [547.95, 168.83], [551.86, 169.43], [564.79, 169.42]]}]\n```", - "options": null, - "id": 301 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001358", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[217.41, 302.91], [166.45, 302.92], [206.66, 291.66], [252.92, 291.65], [217.41, 166.84], [166.45, 166.84], [206.66, 168.1], [252.92, 168.1]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[702.9, 219.56], [720.44, 219.55], [725.72, 221.23], [707.43, 221.24], [702.9, 159.24], [720.44, 159.24], [725.72, 158.33], [707.43, 158.32]]}]\n```", - "options": null, - "id": 302 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001358", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[656.5, 245.58], [682.18, 245.56], [692.27, 253.32], [663.53, 253.34], [656.5, 165.36], [682.18, 165.36], [692.27, 163.56], [663.53, 163.55]]}, {\"category\": \"car\", \"corners_3d\": [[1219.28, 276.54], [1359.61, 298.19], [987.28, 298.82], [915.15, 276.95], [1219.28, 170.67], [1359.61, 168.45], [987.28, 168.38], [915.15, 170.62]]}, {\"category\": \"cyclist\", \"corners_3d\": [[561.56, 207.55], [547.95, 207.54], [551.86, 206.16], [564.79, 206.17], [561.56, 168.83], [547.95, 168.83], [551.86, 169.43], [564.79, 169.42]]}, {\"category\": \"car\", \"corners_3d\": [[827.49, 206.35], [841.22, 207.99], [719.32, 207.95], [712.88, 206.31], [827.49, 169.85], [841.22, 169.17], [719.32, 169.19], [712.88, 169.86]]}]\n```", - "options": null, - "id": 303 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001358", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[217.41, 302.91], [166.45, 302.92], [206.66, 291.66], [252.92, 291.65], [217.41, 166.84], [166.45, 166.84], [206.66, 168.1], [252.92, 168.1]]}, {\"category\": \"car\", \"corners_3d\": [[1219.28, 276.54], [1359.61, 298.19], [987.28, 298.82], [915.15, 276.95], [1219.28, 170.67], [1359.61, 168.45], [987.28, 168.38], [915.15, 170.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[702.9, 219.56], [720.44, 219.55], [725.72, 221.23], [707.43, 221.24], [702.9, 159.24], [720.44, 159.24], [725.72, 158.33], [707.43, 158.32]]}, {\"category\": \"car\", \"corners_3d\": [[827.49, 206.35], [841.22, 207.99], [719.32, 207.95], [712.88, 206.31], [827.49, 169.85], [841.22, 169.17], [719.32, 169.19], [712.88, 169.86]]}]\n```", - "options": null, - "id": 304 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001358", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[217.41, 302.91], [166.45, 302.92], [206.66, 291.66], [252.92, 291.65], [217.41, 166.84], [166.45, 166.84], [206.66, 168.1], [252.92, 168.1]]}, {\"category\": \"cyclist\", \"corners_3d\": [[656.5, 245.58], [682.18, 245.56], [692.27, 253.32], [663.53, 253.34], [656.5, 165.36], [682.18, 165.36], [692.27, 163.56], [663.53, 163.55]]}, {\"category\": \"cyclist\", \"corners_3d\": [[561.56, 207.55], [547.95, 207.54], [551.86, 206.16], [564.79, 206.17], [561.56, 168.83], [547.95, 168.83], [551.86, 169.43], [564.79, 169.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[702.9, 219.56], [720.44, 219.55], [725.72, 221.23], [707.43, 221.24], [702.9, 159.24], [720.44, 159.24], [725.72, 158.33], [707.43, 158.32]]}]\n```", - "options": null, - "id": 305 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001358", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[217.41, 302.91], [166.45, 302.92], [206.66, 291.66], [252.92, 291.65], [217.41, 166.84], [166.45, 166.84], [206.66, 168.1], [252.92, 168.1]]}, {\"category\": \"cyclist\", \"corners_3d\": [[656.5, 245.58], [682.18, 245.56], [692.27, 253.32], [663.53, 253.34], [656.5, 165.36], [682.18, 165.36], [692.27, 163.56], [663.53, 163.55]]}, {\"category\": \"car\", \"corners_3d\": [[1219.28, 276.54], [1359.61, 298.19], [987.28, 298.82], [915.15, 276.95], [1219.28, 170.67], [1359.61, 168.45], [987.28, 168.38], [915.15, 170.62]]}, {\"category\": \"cyclist\", \"corners_3d\": [[561.56, 207.55], [547.95, 207.54], [551.86, 206.16], [564.79, 206.17], [561.56, 168.83], [547.95, 168.83], [551.86, 169.43], [564.79, 169.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[702.9, 219.56], [720.44, 219.55], [725.72, 221.23], [707.43, 221.24], [702.9, 159.24], [720.44, 159.24], [725.72, 158.33], [707.43, 158.32]]}, {\"category\": \"car\", \"corners_3d\": [[827.49, 206.35], [841.22, 207.99], [719.32, 207.95], [712.88, 206.31], [827.49, 169.85], [841.22, 169.17], [719.32, 169.19], [712.88, 169.86]]}]\n```", - "options": null, - "id": 306 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001381", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[434.22, 270.11], [333.99, 270.52], [400.58, 244.85], [474.34, 244.62], [434.22, 178.34], [333.99, 178.36], [400.58, 176.91], [474.34, 176.9]]}]\n```", - "options": null, - "id": 307 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001396", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[730.99, 202.41], [729.3, 201.82], [763.49, 201.89], [765.88, 202.49], [730.99, 173.54], [729.3, 173.52], [763.49, 173.52], [765.88, 173.54]]}]\n```", - "options": null, - "id": 308 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001396", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[819.24, 212.85], [805.67, 212.86], [799.61, 211.71], [812.79, 211.7], [819.24, 170.45], [805.67, 170.45], [799.61, 170.52], [812.79, 170.52]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.5, 201.92], [554.16, 202.35], [543.3, 202.41], [542.78, 201.98], [553.5, 171.21], [554.16, 171.18], [543.3, 171.18], [542.78, 171.2]]}]\n```", - "options": null, - "id": 309 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001396", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[599.1, 219.98], [647.68, 219.98], [653.52, 227.32], [597.39, 227.31], [599.1, 180.89], [647.68, 180.89], [653.52, 182.14], [597.39, 182.14]]}, {\"category\": \"car\", \"corners_3d\": [[444.47, 201.94], [419.48, 201.9], [435.36, 199.79], [458.55, 199.82], [444.47, 178.58], [419.48, 178.57], [435.36, 178.16], [458.55, 178.16]]}, {\"category\": \"car\", \"corners_3d\": [[586.11, 207.64], [620.9, 207.65], [621.32, 211.7], [582.48, 211.68], [586.11, 175.97], [620.9, 175.97], [621.32, 176.34], [582.48, 176.33]]}]\n```", - "options": null, - "id": 310 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001396", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[635.77, 199.77], [661.36, 199.8], [664.07, 202.27], [636.13, 202.23], [635.77, 170.65], [661.36, 170.65], [664.07, 170.45], [636.13, 170.45]]}]\n```", - "options": null, - "id": 311 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001396", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[819.24, 212.85], [805.67, 212.86], [799.61, 211.71], [812.79, 211.7], [819.24, 170.45], [805.67, 170.45], [799.61, 170.52], [812.79, 170.52]]}, {\"category\": \"cyclist\", \"corners_3d\": [[730.99, 202.41], [729.3, 201.82], [763.49, 201.89], [765.88, 202.49], [730.99, 173.54], [729.3, 173.52], [763.49, 173.52], [765.88, 173.54]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.5, 201.92], [554.16, 202.35], [543.3, 202.41], [542.78, 201.98], [553.5, 171.21], [554.16, 171.18], [543.3, 171.18], [542.78, 171.2]]}]\n```", - "options": null, - "id": 312 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001396", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[599.1, 219.98], [647.68, 219.98], [653.52, 227.32], [597.39, 227.31], [599.1, 180.89], [647.68, 180.89], [653.52, 182.14], [597.39, 182.14]]}, {\"category\": \"car\", \"corners_3d\": [[444.47, 201.94], [419.48, 201.9], [435.36, 199.79], [458.55, 199.82], [444.47, 178.58], [419.48, 178.57], [435.36, 178.16], [458.55, 178.16]]}, {\"category\": \"cyclist\", \"corners_3d\": [[730.99, 202.41], [729.3, 201.82], [763.49, 201.89], [765.88, 202.49], [730.99, 173.54], [729.3, 173.52], [763.49, 173.52], [765.88, 173.54]]}, {\"category\": \"car\", \"corners_3d\": [[586.11, 207.64], [620.9, 207.65], [621.32, 211.7], [582.48, 211.68], [586.11, 175.97], [620.9, 175.97], [621.32, 176.34], [582.48, 176.33]]}]\n```", - "options": null, - "id": 313 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001396", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[730.99, 202.41], [729.3, 201.82], [763.49, 201.89], [765.88, 202.49], [730.99, 173.54], [729.3, 173.52], [763.49, 173.52], [765.88, 173.54]]}, {\"category\": \"van\", \"corners_3d\": [[635.77, 199.77], [661.36, 199.8], [664.07, 202.27], [636.13, 202.23], [635.77, 170.65], [661.36, 170.65], [664.07, 170.45], [636.13, 170.45]]}]\n```", - "options": null, - "id": 314 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001396", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[599.1, 219.98], [647.68, 219.98], [653.52, 227.32], [597.39, 227.31], [599.1, 180.89], [647.68, 180.89], [653.52, 182.14], [597.39, 182.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[819.24, 212.85], [805.67, 212.86], [799.61, 211.71], [812.79, 211.7], [819.24, 170.45], [805.67, 170.45], [799.61, 170.52], [812.79, 170.52]]}, {\"category\": \"car\", \"corners_3d\": [[444.47, 201.94], [419.48, 201.9], [435.36, 199.79], [458.55, 199.82], [444.47, 178.58], [419.48, 178.57], [435.36, 178.16], [458.55, 178.16]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.5, 201.92], [554.16, 202.35], [543.3, 202.41], [542.78, 201.98], [553.5, 171.21], [554.16, 171.18], [543.3, 171.18], [542.78, 171.2]]}, {\"category\": \"car\", \"corners_3d\": [[586.11, 207.64], [620.9, 207.65], [621.32, 211.7], [582.48, 211.68], [586.11, 175.97], [620.9, 175.97], [621.32, 176.34], [582.48, 176.33]]}]\n```", - "options": null, - "id": 315 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001396", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[819.24, 212.85], [805.67, 212.86], [799.61, 211.71], [812.79, 211.7], [819.24, 170.45], [805.67, 170.45], [799.61, 170.52], [812.79, 170.52]]}, {\"category\": \"van\", \"corners_3d\": [[635.77, 199.77], [661.36, 199.8], [664.07, 202.27], [636.13, 202.23], [635.77, 170.65], [661.36, 170.65], [664.07, 170.45], [636.13, 170.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.5, 201.92], [554.16, 202.35], [543.3, 202.41], [542.78, 201.98], [553.5, 171.21], [554.16, 171.18], [543.3, 171.18], [542.78, 171.2]]}]\n```", - "options": null, - "id": 316 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001396", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[599.1, 219.98], [647.68, 219.98], [653.52, 227.32], [597.39, 227.31], [599.1, 180.89], [647.68, 180.89], [653.52, 182.14], [597.39, 182.14]]}, {\"category\": \"car\", \"corners_3d\": [[444.47, 201.94], [419.48, 201.9], [435.36, 199.79], [458.55, 199.82], [444.47, 178.58], [419.48, 178.57], [435.36, 178.16], [458.55, 178.16]]}, {\"category\": \"van\", \"corners_3d\": [[635.77, 199.77], [661.36, 199.8], [664.07, 202.27], [636.13, 202.23], [635.77, 170.65], [661.36, 170.65], [664.07, 170.45], [636.13, 170.45]]}, {\"category\": \"car\", \"corners_3d\": [[586.11, 207.64], [620.9, 207.65], [621.32, 211.7], [582.48, 211.68], [586.11, 175.97], [620.9, 175.97], [621.32, 176.34], [582.48, 176.33]]}]\n```", - "options": null, - "id": 317 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001396", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, pedestrian, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[599.1, 219.98], [647.68, 219.98], [653.52, 227.32], [597.39, 227.31], [599.1, 180.89], [647.68, 180.89], [653.52, 182.14], [597.39, 182.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[819.24, 212.85], [805.67, 212.86], [799.61, 211.71], [812.79, 211.7], [819.24, 170.45], [805.67, 170.45], [799.61, 170.52], [812.79, 170.52]]}, {\"category\": \"car\", \"corners_3d\": [[444.47, 201.94], [419.48, 201.9], [435.36, 199.79], [458.55, 199.82], [444.47, 178.58], [419.48, 178.57], [435.36, 178.16], [458.55, 178.16]]}, {\"category\": \"cyclist\", \"corners_3d\": [[730.99, 202.41], [729.3, 201.82], [763.49, 201.89], [765.88, 202.49], [730.99, 173.54], [729.3, 173.52], [763.49, 173.52], [765.88, 173.54]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.5, 201.92], [554.16, 202.35], [543.3, 202.41], [542.78, 201.98], [553.5, 171.21], [554.16, 171.18], [543.3, 171.18], [542.78, 171.2]]}, {\"category\": \"car\", \"corners_3d\": [[586.11, 207.64], [620.9, 207.65], [621.32, 211.7], [582.48, 211.68], [586.11, 175.97], [620.9, 175.97], [621.32, 176.34], [582.48, 176.33]]}]\n```", - "options": null, - "id": 318 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001396", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, pedestrian, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[819.24, 212.85], [805.67, 212.86], [799.61, 211.71], [812.79, 211.7], [819.24, 170.45], [805.67, 170.45], [799.61, 170.52], [812.79, 170.52]]}, {\"category\": \"cyclist\", \"corners_3d\": [[730.99, 202.41], [729.3, 201.82], [763.49, 201.89], [765.88, 202.49], [730.99, 173.54], [729.3, 173.52], [763.49, 173.52], [765.88, 173.54]]}, {\"category\": \"van\", \"corners_3d\": [[635.77, 199.77], [661.36, 199.8], [664.07, 202.27], [636.13, 202.23], [635.77, 170.65], [661.36, 170.65], [664.07, 170.45], [636.13, 170.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.5, 201.92], [554.16, 202.35], [543.3, 202.41], [542.78, 201.98], [553.5, 171.21], [554.16, 171.18], [543.3, 171.18], [542.78, 171.2]]}]\n```", - "options": null, - "id": 319 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001396", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[599.1, 219.98], [647.68, 219.98], [653.52, 227.32], [597.39, 227.31], [599.1, 180.89], [647.68, 180.89], [653.52, 182.14], [597.39, 182.14]]}, {\"category\": \"car\", \"corners_3d\": [[444.47, 201.94], [419.48, 201.9], [435.36, 199.79], [458.55, 199.82], [444.47, 178.58], [419.48, 178.57], [435.36, 178.16], [458.55, 178.16]]}, {\"category\": \"cyclist\", \"corners_3d\": [[730.99, 202.41], [729.3, 201.82], [763.49, 201.89], [765.88, 202.49], [730.99, 173.54], [729.3, 173.52], [763.49, 173.52], [765.88, 173.54]]}, {\"category\": \"van\", \"corners_3d\": [[635.77, 199.77], [661.36, 199.8], [664.07, 202.27], [636.13, 202.23], [635.77, 170.65], [661.36, 170.65], [664.07, 170.45], [636.13, 170.45]]}, {\"category\": \"car\", \"corners_3d\": [[586.11, 207.64], [620.9, 207.65], [621.32, 211.7], [582.48, 211.68], [586.11, 175.97], [620.9, 175.97], [621.32, 176.34], [582.48, 176.33]]}]\n```", - "options": null, - "id": 320 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001396", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[599.1, 219.98], [647.68, 219.98], [653.52, 227.32], [597.39, 227.31], [599.1, 180.89], [647.68, 180.89], [653.52, 182.14], [597.39, 182.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[819.24, 212.85], [805.67, 212.86], [799.61, 211.71], [812.79, 211.7], [819.24, 170.45], [805.67, 170.45], [799.61, 170.52], [812.79, 170.52]]}, {\"category\": \"car\", \"corners_3d\": [[444.47, 201.94], [419.48, 201.9], [435.36, 199.79], [458.55, 199.82], [444.47, 178.58], [419.48, 178.57], [435.36, 178.16], [458.55, 178.16]]}, {\"category\": \"van\", \"corners_3d\": [[635.77, 199.77], [661.36, 199.8], [664.07, 202.27], [636.13, 202.23], [635.77, 170.65], [661.36, 170.65], [664.07, 170.45], [636.13, 170.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.5, 201.92], [554.16, 202.35], [543.3, 202.41], [542.78, 201.98], [553.5, 171.21], [554.16, 171.18], [543.3, 171.18], [542.78, 171.2]]}, {\"category\": \"car\", \"corners_3d\": [[586.11, 207.64], [620.9, 207.65], [621.32, 211.7], [582.48, 211.68], [586.11, 175.97], [620.9, 175.97], [621.32, 176.34], [582.48, 176.33]]}]\n```", - "options": null, - "id": 321 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001396", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, pedestrian, car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[599.1, 219.98], [647.68, 219.98], [653.52, 227.32], [597.39, 227.31], [599.1, 180.89], [647.68, 180.89], [653.52, 182.14], [597.39, 182.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[819.24, 212.85], [805.67, 212.86], [799.61, 211.71], [812.79, 211.7], [819.24, 170.45], [805.67, 170.45], [799.61, 170.52], [812.79, 170.52]]}, {\"category\": \"car\", \"corners_3d\": [[444.47, 201.94], [419.48, 201.9], [435.36, 199.79], [458.55, 199.82], [444.47, 178.58], [419.48, 178.57], [435.36, 178.16], [458.55, 178.16]]}, {\"category\": \"cyclist\", \"corners_3d\": [[730.99, 202.41], [729.3, 201.82], [763.49, 201.89], [765.88, 202.49], [730.99, 173.54], [729.3, 173.52], [763.49, 173.52], [765.88, 173.54]]}, {\"category\": \"van\", \"corners_3d\": [[635.77, 199.77], [661.36, 199.8], [664.07, 202.27], [636.13, 202.23], [635.77, 170.65], [661.36, 170.65], [664.07, 170.45], [636.13, 170.45]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.5, 201.92], [554.16, 202.35], [543.3, 202.41], [542.78, 201.98], [553.5, 171.21], [554.16, 171.18], [543.3, 171.18], [542.78, 171.2]]}, {\"category\": \"car\", \"corners_3d\": [[586.11, 207.64], [620.9, 207.65], [621.32, 211.7], [582.48, 211.68], [586.11, 175.97], [620.9, 175.97], [621.32, 176.34], [582.48, 176.33]]}]\n```", - "options": null, - "id": 322 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001440", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[370.07, 213.35], [381.81, 211.29], [464.69, 211.24], [457.4, 213.29], [370.07, 178.26], [381.81, 177.99], [464.69, 177.98], [457.4, 178.25]]}]\n```", - "options": null, - "id": 323 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001460", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[673.48, 244.79], [740.03, 244.73], [776.97, 264.09], [692.51, 264.19], [673.48, 187.47], [740.03, 187.46], [776.97, 191.4], [692.51, 191.42]]}, {\"category\": \"car\", \"corners_3d\": [[523.73, 216.05], [491.21, 216.02], [503.43, 212.02], [532.96, 212.05], [523.73, 189.54], [491.21, 189.52], [503.43, 187.98], [532.96, 187.99]]}, {\"category\": \"car\", \"corners_3d\": [[429.89, 256.04], [366.0, 255.97], [410.65, 241.14], [463.16, 241.19], [429.89, 196.72], [366.0, 196.71], [410.65, 192.45], [463.16, 192.46]]}, {\"category\": \"car\", \"corners_3d\": [[466.73, 238.57], [416.01, 238.34], [448.76, 228.98], [492.3, 229.14], [466.73, 192.68], [416.01, 192.61], [448.76, 189.79], [492.3, 189.84]]}, {\"category\": \"car\", \"corners_3d\": [[265.15, 316.4], [151.26, 316.19], [289.8, 273.48], [369.86, 273.58], [265.15, 221.41], [151.26, 221.34], [289.8, 206.89], [369.86, 206.93]]}]\n```", - "options": null, - "id": 324 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001501", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[396.24, 209.84], [367.56, 209.84], [386.61, 206.93], [413.03, 206.92], [396.24, 187.34], [367.56, 187.34], [386.61, 186.2], [413.03, 186.2]]}]\n```", - "options": null, - "id": 325 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001501", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (bus). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[597.41, 191.1], [621.29, 191.11], [622.01, 194.46], [593.74, 194.46], [597.41, 165.22], [621.29, 165.22], [622.01, 163.82], [593.74, 163.82]]}]\n```", - "options": null, - "id": 326 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001501", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, bus). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[597.41, 191.1], [621.29, 191.11], [622.01, 194.46], [593.74, 194.46], [597.41, 165.22], [621.29, 165.22], [622.01, 163.82], [593.74, 163.82]]}, {\"category\": \"car\", \"corners_3d\": [[396.24, 209.84], [367.56, 209.84], [386.61, 206.93], [413.03, 206.92], [396.24, 187.34], [367.56, 187.34], [386.61, 186.2], [413.03, 186.2]]}]\n```", - "options": null, - "id": 327 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001510", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[673.61, 194.72], [708.17, 194.68], [730.5, 198.52], [689.89, 198.58], [673.61, 146.5], [708.17, 146.54], [730.5, 141.91], [689.89, 141.85]]}]\n```", - "options": null, - "id": 328 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001510", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[629.44, 194.98], [651.26, 194.96], [656.18, 196.7], [632.66, 196.72], [629.44, 174.74], [651.26, 174.74], [656.18, 174.88], [632.66, 174.89]]}, {\"category\": \"car\", \"corners_3d\": [[669.88, 215.75], [710.38, 215.63], [732.75, 222.65], [685.65, 222.81], [669.88, 172.07], [710.38, 172.07], [732.75, 171.94], [685.65, 171.94]]}, {\"category\": \"car\", \"corners_3d\": [[719.57, 238.69], [782.47, 238.63], [821.77, 253.03], [745.14, 253.11], [719.57, 182.07], [782.47, 182.07], [821.77, 184.08], [745.14, 184.09]]}, {\"category\": \"car\", \"corners_3d\": [[869.32, 236.75], [931.98, 236.59], [1033.99, 255.52], [952.99, 255.79], [869.32, 179.55], [931.98, 179.53], [1033.99, 181.51], [952.99, 181.54]]}, {\"category\": \"car\", \"corners_3d\": [[810.1, 201.2], [825.64, 200.69], [877.86, 202.74], [862.19, 203.33], [810.1, 175.46], [825.64, 175.41], [877.86, 175.6], [862.19, 175.65]]}]\n```", - "options": null, - "id": 329 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001510", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[629.44, 194.98], [651.26, 194.96], [656.18, 196.7], [632.66, 196.72], [629.44, 174.74], [651.26, 174.74], [656.18, 174.88], [632.66, 174.89]]}, {\"category\": \"car\", \"corners_3d\": [[669.88, 215.75], [710.38, 215.63], [732.75, 222.65], [685.65, 222.81], [669.88, 172.07], [710.38, 172.07], [732.75, 171.94], [685.65, 171.94]]}, {\"category\": \"car\", \"corners_3d\": [[719.57, 238.69], [782.47, 238.63], [821.77, 253.03], [745.14, 253.11], [719.57, 182.07], [782.47, 182.07], [821.77, 184.08], [745.14, 184.09]]}, {\"category\": \"truck\", \"corners_3d\": [[673.61, 194.72], [708.17, 194.68], [730.5, 198.52], [689.89, 198.58], [673.61, 146.5], [708.17, 146.54], [730.5, 141.91], [689.89, 141.85]]}, {\"category\": \"car\", \"corners_3d\": [[869.32, 236.75], [931.98, 236.59], [1033.99, 255.52], [952.99, 255.79], [869.32, 179.55], [931.98, 179.53], [1033.99, 181.51], [952.99, 181.54]]}, {\"category\": \"car\", \"corners_3d\": [[810.1, 201.2], [825.64, 200.69], [877.86, 202.74], [862.19, 203.33], [810.1, 175.46], [825.64, 175.41], [877.86, 175.6], [862.19, 175.65]]}]\n```", - "options": null, - "id": 330 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001531", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-120.97, 319.54], [-20.66, 295.82], [304.41, 288.35], [269.5, 309.03], [-120.97, 203.18], [-20.66, 198.27], [304.41, 196.73], [269.5, 201.01]]}, {\"category\": \"car\", \"corners_3d\": [[1077.54, 255.79], [1170.92, 255.79], [1336.7, 280.31], [1215.69, 280.3], [1077.54, 155.98], [1170.92, 155.98], [1336.7, 150.99], [1215.69, 150.99]]}, {\"category\": \"car\", \"corners_3d\": [[443.44, 244.62], [425.39, 251.02], [234.15, 249.97], [267.44, 243.73], [443.44, 185.01], [425.39, 186.1], [234.15, 185.92], [267.44, 184.86]]}]\n```", - "options": null, - "id": 331 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001531", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[312.54, 231.61], [335.07, 230.78], [341.08, 233.49], [317.57, 234.4], [312.54, 171.89], [335.07, 171.9], [341.08, 171.86], [317.57, 171.84]]}]\n```", - "options": null, - "id": 332 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001531", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-120.97, 319.54], [-20.66, 295.82], [304.41, 288.35], [269.5, 309.03], [-120.97, 203.18], [-20.66, 198.27], [304.41, 196.73], [269.5, 201.01]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[312.54, 231.61], [335.07, 230.78], [341.08, 233.49], [317.57, 234.4], [312.54, 171.89], [335.07, 171.9], [341.08, 171.86], [317.57, 171.84]]}, {\"category\": \"car\", \"corners_3d\": [[1077.54, 255.79], [1170.92, 255.79], [1336.7, 280.31], [1215.69, 280.3], [1077.54, 155.98], [1170.92, 155.98], [1336.7, 150.99], [1215.69, 150.99]]}, {\"category\": \"car\", \"corners_3d\": [[443.44, 244.62], [425.39, 251.02], [234.15, 249.97], [267.44, 243.73], [443.44, 185.01], [425.39, 186.1], [234.15, 185.92], [267.44, 184.86]]}]\n```", - "options": null, - "id": 333 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001540", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[398.93, 299.85], [381.55, 309.96], [268.92, 309.96], [295.09, 299.85], [398.93, 153.17], [381.55, 150.85], [268.92, 150.85], [295.09, 153.17]]}]\n```", - "options": null, - "id": 334 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001546", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[582.0, 189.93], [599.08, 189.92], [599.67, 190.86], [581.64, 190.88], [582.0, 176.48], [599.08, 176.48], [599.67, 176.68], [581.64, 176.68]]}, {\"category\": \"car\", \"corners_3d\": [[396.41, 233.51], [340.81, 233.51], [379.51, 224.76], [427.09, 224.76], [396.41, 183.35], [340.81, 183.35], [379.51, 181.83], [427.09, 181.83]]}]\n```", - "options": null, - "id": 335 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001594", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[652.77, 213.19], [692.08, 213.17], [708.4, 220.55], [661.9, 220.58], [652.77, 165.89], [692.08, 165.89], [708.4, 164.62], [661.9, 164.62]]}]\n```", - "options": null, - "id": 336 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001594", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-328.22, 898.36], [-1117.9, 914.44], [84.63, 393.68], [315.53, 392.23], [-328.22, 283.99], [-1117.9, 286.46], [84.63, 206.68], [315.53, 206.46]]}, {\"category\": \"car\", \"corners_3d\": [[683.33, 231.2], [732.28, 231.2], [756.53, 242.79], [697.86, 242.78], [683.33, 183.82], [732.28, 183.82], [756.53, 186.0], [697.86, 186.0]]}, {\"category\": \"car\", \"corners_3d\": [[668.87, 222.12], [711.64, 222.06], [729.79, 229.77], [680.34, 229.84], [668.87, 182.04], [711.64, 182.03], [729.79, 183.46], [680.34, 183.48]]}, {\"category\": \"car\", \"corners_3d\": [[566.18, 208.69], [536.67, 208.78], [539.42, 205.38], [566.13, 205.31], [566.18, 178.92], [536.67, 178.93], [539.42, 178.36], [566.13, 178.35]]}]\n```", - "options": null, - "id": 337 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001594", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-328.22, 898.36], [-1117.9, 914.44], [84.63, 393.68], [315.53, 392.23], [-328.22, 283.99], [-1117.9, 286.46], [84.63, 206.68], [315.53, 206.46]]}, {\"category\": \"car\", \"corners_3d\": [[683.33, 231.2], [732.28, 231.2], [756.53, 242.79], [697.86, 242.78], [683.33, 183.82], [732.28, 183.82], [756.53, 186.0], [697.86, 186.0]]}, {\"category\": \"car\", \"corners_3d\": [[668.87, 222.12], [711.64, 222.06], [729.79, 229.77], [680.34, 229.84], [668.87, 182.04], [711.64, 182.03], [729.79, 183.46], [680.34, 183.48]]}, {\"category\": \"van\", \"corners_3d\": [[652.77, 213.19], [692.08, 213.17], [708.4, 220.55], [661.9, 220.58], [652.77, 165.89], [692.08, 165.89], [708.4, 164.62], [661.9, 164.62]]}, {\"category\": \"car\", \"corners_3d\": [[566.18, 208.69], [536.67, 208.78], [539.42, 205.38], [566.13, 205.31], [566.18, 178.92], [536.67, 178.93], [539.42, 178.36], [566.13, 178.35]]}]\n```", - "options": null, - "id": 338 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001608", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[948.41, 224.51], [926.8, 224.54], [911.81, 222.21], [932.46, 222.18], [948.41, 156.77], [926.8, 156.76], [911.81, 157.48], [932.46, 157.49]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[548.44, 206.23], [549.06, 206.85], [535.18, 206.94], [534.81, 206.31], [548.44, 167.21], [549.06, 167.11], [535.18, 167.09], [534.81, 167.2]]}]\n```", - "options": null, - "id": 339 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001608", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[651.52, 198.72], [682.85, 198.75], [688.7, 201.72], [653.78, 201.67], [651.52, 163.11], [682.85, 163.09], [688.7, 161.97], [653.78, 161.99]]}]\n```", - "options": null, - "id": 340 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001608", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[752.32, 203.2], [749.76, 202.41], [794.38, 202.52], [798.15, 203.31], [752.32, 165.43], [749.76, 165.62], [794.38, 165.59], [798.15, 165.4]]}]\n```", - "options": null, - "id": 341 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001608", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.37, 222.78], [661.4, 222.78], [671.02, 232.16], [603.29, 232.15], [604.37, 176.89], [661.4, 176.89], [671.02, 177.65], [603.29, 177.65]]}, {\"category\": \"car\", \"corners_3d\": [[1043.84, 205.73], [1017.01, 203.83], [1121.67, 203.64], [1154.7, 205.51], [1043.84, 167.28], [1017.01, 167.6], [1121.67, 167.63], [1154.7, 167.31]]}, {\"category\": \"car\", \"corners_3d\": [[416.33, 207.03], [385.14, 207.02], [405.62, 203.99], [434.04, 204.0], [416.33, 178.12], [385.14, 178.12], [405.62, 177.65], [434.04, 177.65]]}, {\"category\": \"car\", \"corners_3d\": [[587.38, 207.86], [625.1, 207.84], [627.92, 212.3], [585.4, 212.32], [587.38, 173.52], [625.1, 173.52], [627.92, 173.61], [585.4, 173.61]]}]\n```", - "options": null, - "id": 342 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001608", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[948.41, 224.51], [926.8, 224.54], [911.81, 222.21], [932.46, 222.18], [948.41, 156.77], [926.8, 156.76], [911.81, 157.48], [932.46, 157.49]]}, {\"category\": \"van\", \"corners_3d\": [[651.52, 198.72], [682.85, 198.75], [688.7, 201.72], [653.78, 201.67], [651.52, 163.11], [682.85, 163.09], [688.7, 161.97], [653.78, 161.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[548.44, 206.23], [549.06, 206.85], [535.18, 206.94], [534.81, 206.31], [548.44, 167.21], [549.06, 167.11], [535.18, 167.09], [534.81, 167.2]]}]\n```", - "options": null, - "id": 343 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001608", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[948.41, 224.51], [926.8, 224.54], [911.81, 222.21], [932.46, 222.18], [948.41, 156.77], [926.8, 156.76], [911.81, 157.48], [932.46, 157.49]]}, {\"category\": \"cyclist\", \"corners_3d\": [[752.32, 203.2], [749.76, 202.41], [794.38, 202.52], [798.15, 203.31], [752.32, 165.43], [749.76, 165.62], [794.38, 165.59], [798.15, 165.4]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[548.44, 206.23], [549.06, 206.85], [535.18, 206.94], [534.81, 206.31], [548.44, 167.21], [549.06, 167.11], [535.18, 167.09], [534.81, 167.2]]}]\n```", - "options": null, - "id": 344 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001608", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.37, 222.78], [661.4, 222.78], [671.02, 232.16], [603.29, 232.15], [604.37, 176.89], [661.4, 176.89], [671.02, 177.65], [603.29, 177.65]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[948.41, 224.51], [926.8, 224.54], [911.81, 222.21], [932.46, 222.18], [948.41, 156.77], [926.8, 156.76], [911.81, 157.48], [932.46, 157.49]]}, {\"category\": \"car\", \"corners_3d\": [[1043.84, 205.73], [1017.01, 203.83], [1121.67, 203.64], [1154.7, 205.51], [1043.84, 167.28], [1017.01, 167.6], [1121.67, 167.63], [1154.7, 167.31]]}, {\"category\": \"car\", \"corners_3d\": [[416.33, 207.03], [385.14, 207.02], [405.62, 203.99], [434.04, 204.0], [416.33, 178.12], [385.14, 178.12], [405.62, 177.65], [434.04, 177.65]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[548.44, 206.23], [549.06, 206.85], [535.18, 206.94], [534.81, 206.31], [548.44, 167.21], [549.06, 167.11], [535.18, 167.09], [534.81, 167.2]]}, {\"category\": \"car\", \"corners_3d\": [[587.38, 207.86], [625.1, 207.84], [627.92, 212.3], [585.4, 212.32], [587.38, 173.52], [625.1, 173.52], [627.92, 173.61], [585.4, 173.61]]}]\n```", - "options": null, - "id": 345 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001608", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[752.32, 203.2], [749.76, 202.41], [794.38, 202.52], [798.15, 203.31], [752.32, 165.43], [749.76, 165.62], [794.38, 165.59], [798.15, 165.4]]}, {\"category\": \"van\", \"corners_3d\": [[651.52, 198.72], [682.85, 198.75], [688.7, 201.72], [653.78, 201.67], [651.52, 163.11], [682.85, 163.09], [688.7, 161.97], [653.78, 161.99]]}]\n```", - "options": null, - "id": 346 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001608", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.37, 222.78], [661.4, 222.78], [671.02, 232.16], [603.29, 232.15], [604.37, 176.89], [661.4, 176.89], [671.02, 177.65], [603.29, 177.65]]}, {\"category\": \"car\", \"corners_3d\": [[1043.84, 205.73], [1017.01, 203.83], [1121.67, 203.64], [1154.7, 205.51], [1043.84, 167.28], [1017.01, 167.6], [1121.67, 167.63], [1154.7, 167.31]]}, {\"category\": \"car\", \"corners_3d\": [[416.33, 207.03], [385.14, 207.02], [405.62, 203.99], [434.04, 204.0], [416.33, 178.12], [385.14, 178.12], [405.62, 177.65], [434.04, 177.65]]}, {\"category\": \"van\", \"corners_3d\": [[651.52, 198.72], [682.85, 198.75], [688.7, 201.72], [653.78, 201.67], [651.52, 163.11], [682.85, 163.09], [688.7, 161.97], [653.78, 161.99]]}, {\"category\": \"car\", \"corners_3d\": [[587.38, 207.86], [625.1, 207.84], [627.92, 212.3], [585.4, 212.32], [587.38, 173.52], [625.1, 173.52], [627.92, 173.61], [585.4, 173.61]]}]\n```", - "options": null, - "id": 347 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001608", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.37, 222.78], [661.4, 222.78], [671.02, 232.16], [603.29, 232.15], [604.37, 176.89], [661.4, 176.89], [671.02, 177.65], [603.29, 177.65]]}, {\"category\": \"car\", \"corners_3d\": [[1043.84, 205.73], [1017.01, 203.83], [1121.67, 203.64], [1154.7, 205.51], [1043.84, 167.28], [1017.01, 167.6], [1121.67, 167.63], [1154.7, 167.31]]}, {\"category\": \"car\", \"corners_3d\": [[416.33, 207.03], [385.14, 207.02], [405.62, 203.99], [434.04, 204.0], [416.33, 178.12], [385.14, 178.12], [405.62, 177.65], [434.04, 177.65]]}, {\"category\": \"cyclist\", \"corners_3d\": [[752.32, 203.2], [749.76, 202.41], [794.38, 202.52], [798.15, 203.31], [752.32, 165.43], [749.76, 165.62], [794.38, 165.59], [798.15, 165.4]]}, {\"category\": \"car\", \"corners_3d\": [[587.38, 207.86], [625.1, 207.84], [627.92, 212.3], [585.4, 212.32], [587.38, 173.52], [625.1, 173.52], [627.92, 173.61], [585.4, 173.61]]}]\n```", - "options": null, - "id": 348 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001608", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, van, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[948.41, 224.51], [926.8, 224.54], [911.81, 222.21], [932.46, 222.18], [948.41, 156.77], [926.8, 156.76], [911.81, 157.48], [932.46, 157.49]]}, {\"category\": \"cyclist\", \"corners_3d\": [[752.32, 203.2], [749.76, 202.41], [794.38, 202.52], [798.15, 203.31], [752.32, 165.43], [749.76, 165.62], [794.38, 165.59], [798.15, 165.4]]}, {\"category\": \"van\", \"corners_3d\": [[651.52, 198.72], [682.85, 198.75], [688.7, 201.72], [653.78, 201.67], [651.52, 163.11], [682.85, 163.09], [688.7, 161.97], [653.78, 161.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[548.44, 206.23], [549.06, 206.85], [535.18, 206.94], [534.81, 206.31], [548.44, 167.21], [549.06, 167.11], [535.18, 167.09], [534.81, 167.2]]}]\n```", - "options": null, - "id": 349 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001608", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.37, 222.78], [661.4, 222.78], [671.02, 232.16], [603.29, 232.15], [604.37, 176.89], [661.4, 176.89], [671.02, 177.65], [603.29, 177.65]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[948.41, 224.51], [926.8, 224.54], [911.81, 222.21], [932.46, 222.18], [948.41, 156.77], [926.8, 156.76], [911.81, 157.48], [932.46, 157.49]]}, {\"category\": \"car\", \"corners_3d\": [[1043.84, 205.73], [1017.01, 203.83], [1121.67, 203.64], [1154.7, 205.51], [1043.84, 167.28], [1017.01, 167.6], [1121.67, 167.63], [1154.7, 167.31]]}, {\"category\": \"car\", \"corners_3d\": [[416.33, 207.03], [385.14, 207.02], [405.62, 203.99], [434.04, 204.0], [416.33, 178.12], [385.14, 178.12], [405.62, 177.65], [434.04, 177.65]]}, {\"category\": \"van\", \"corners_3d\": [[651.52, 198.72], [682.85, 198.75], [688.7, 201.72], [653.78, 201.67], [651.52, 163.11], [682.85, 163.09], [688.7, 161.97], [653.78, 161.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[548.44, 206.23], [549.06, 206.85], [535.18, 206.94], [534.81, 206.31], [548.44, 167.21], [549.06, 167.11], [535.18, 167.09], [534.81, 167.2]]}, {\"category\": \"car\", \"corners_3d\": [[587.38, 207.86], [625.1, 207.84], [627.92, 212.3], [585.4, 212.32], [587.38, 173.52], [625.1, 173.52], [627.92, 173.61], [585.4, 173.61]]}]\n```", - "options": null, - "id": 350 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001608", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.37, 222.78], [661.4, 222.78], [671.02, 232.16], [603.29, 232.15], [604.37, 176.89], [661.4, 176.89], [671.02, 177.65], [603.29, 177.65]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[948.41, 224.51], [926.8, 224.54], [911.81, 222.21], [932.46, 222.18], [948.41, 156.77], [926.8, 156.76], [911.81, 157.48], [932.46, 157.49]]}, {\"category\": \"car\", \"corners_3d\": [[1043.84, 205.73], [1017.01, 203.83], [1121.67, 203.64], [1154.7, 205.51], [1043.84, 167.28], [1017.01, 167.6], [1121.67, 167.63], [1154.7, 167.31]]}, {\"category\": \"car\", \"corners_3d\": [[416.33, 207.03], [385.14, 207.02], [405.62, 203.99], [434.04, 204.0], [416.33, 178.12], [385.14, 178.12], [405.62, 177.65], [434.04, 177.65]]}, {\"category\": \"cyclist\", \"corners_3d\": [[752.32, 203.2], [749.76, 202.41], [794.38, 202.52], [798.15, 203.31], [752.32, 165.43], [749.76, 165.62], [794.38, 165.59], [798.15, 165.4]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[548.44, 206.23], [549.06, 206.85], [535.18, 206.94], [534.81, 206.31], [548.44, 167.21], [549.06, 167.11], [535.18, 167.09], [534.81, 167.2]]}, {\"category\": \"car\", \"corners_3d\": [[587.38, 207.86], [625.1, 207.84], [627.92, 212.3], [585.4, 212.32], [587.38, 173.52], [625.1, 173.52], [627.92, 173.61], [585.4, 173.61]]}]\n```", - "options": null, - "id": 351 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001608", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.37, 222.78], [661.4, 222.78], [671.02, 232.16], [603.29, 232.15], [604.37, 176.89], [661.4, 176.89], [671.02, 177.65], [603.29, 177.65]]}, {\"category\": \"car\", \"corners_3d\": [[1043.84, 205.73], [1017.01, 203.83], [1121.67, 203.64], [1154.7, 205.51], [1043.84, 167.28], [1017.01, 167.6], [1121.67, 167.63], [1154.7, 167.31]]}, {\"category\": \"car\", \"corners_3d\": [[416.33, 207.03], [385.14, 207.02], [405.62, 203.99], [434.04, 204.0], [416.33, 178.12], [385.14, 178.12], [405.62, 177.65], [434.04, 177.65]]}, {\"category\": \"cyclist\", \"corners_3d\": [[752.32, 203.2], [749.76, 202.41], [794.38, 202.52], [798.15, 203.31], [752.32, 165.43], [749.76, 165.62], [794.38, 165.59], [798.15, 165.4]]}, {\"category\": \"van\", \"corners_3d\": [[651.52, 198.72], [682.85, 198.75], [688.7, 201.72], [653.78, 201.67], [651.52, 163.11], [682.85, 163.09], [688.7, 161.97], [653.78, 161.99]]}, {\"category\": \"car\", \"corners_3d\": [[587.38, 207.86], [625.1, 207.84], [627.92, 212.3], [585.4, 212.32], [587.38, 173.52], [625.1, 173.52], [627.92, 173.61], [585.4, 173.61]]}]\n```", - "options": null, - "id": 352 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001608", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, van, cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[604.37, 222.78], [661.4, 222.78], [671.02, 232.16], [603.29, 232.15], [604.37, 176.89], [661.4, 176.89], [671.02, 177.65], [603.29, 177.65]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[948.41, 224.51], [926.8, 224.54], [911.81, 222.21], [932.46, 222.18], [948.41, 156.77], [926.8, 156.76], [911.81, 157.48], [932.46, 157.49]]}, {\"category\": \"car\", \"corners_3d\": [[1043.84, 205.73], [1017.01, 203.83], [1121.67, 203.64], [1154.7, 205.51], [1043.84, 167.28], [1017.01, 167.6], [1121.67, 167.63], [1154.7, 167.31]]}, {\"category\": \"car\", \"corners_3d\": [[416.33, 207.03], [385.14, 207.02], [405.62, 203.99], [434.04, 204.0], [416.33, 178.12], [385.14, 178.12], [405.62, 177.65], [434.04, 177.65]]}, {\"category\": \"cyclist\", \"corners_3d\": [[752.32, 203.2], [749.76, 202.41], [794.38, 202.52], [798.15, 203.31], [752.32, 165.43], [749.76, 165.62], [794.38, 165.59], [798.15, 165.4]]}, {\"category\": \"van\", \"corners_3d\": [[651.52, 198.72], [682.85, 198.75], [688.7, 201.72], [653.78, 201.67], [651.52, 163.11], [682.85, 163.09], [688.7, 161.97], [653.78, 161.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[548.44, 206.23], [549.06, 206.85], [535.18, 206.94], [534.81, 206.31], [548.44, 167.21], [549.06, 167.11], [535.18, 167.09], [534.81, 167.2]]}, {\"category\": \"car\", \"corners_3d\": [[587.38, 207.86], [625.1, 207.84], [627.92, 212.3], [585.4, 212.32], [587.38, 173.52], [625.1, 173.52], [627.92, 173.61], [585.4, 173.61]]}]\n```", - "options": null, - "id": 353 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001613", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[454.19, 201.23], [431.81, 201.24], [443.75, 199.24], [464.55, 199.24], [454.19, 182.82], [431.81, 182.83], [443.75, 182.13], [464.55, 182.12]]}]\n```", - "options": null, - "id": 354 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001613", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[355.52, 207.99], [315.36, 208.01], [356.01, 203.02], [390.46, 203.01], [355.52, 152.25], [315.36, 152.24], [356.01, 155.16], [390.46, 155.17]]}, {\"category\": \"truck\", \"corners_3d\": [[396.01, 203.43], [362.83, 203.43], [392.22, 199.78], [421.44, 199.78], [396.01, 156.15], [362.83, 156.15], [392.22, 158.14], [421.44, 158.15]]}]\n```", - "options": null, - "id": 355 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001613", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[454.19, 201.23], [431.81, 201.24], [443.75, 199.24], [464.55, 199.24], [454.19, 182.82], [431.81, 182.83], [443.75, 182.13], [464.55, 182.12]]}, {\"category\": \"truck\", \"corners_3d\": [[355.52, 207.99], [315.36, 208.01], [356.01, 203.02], [390.46, 203.01], [355.52, 152.25], [315.36, 152.24], [356.01, 155.16], [390.46, 155.17]]}, {\"category\": \"truck\", \"corners_3d\": [[396.01, 203.43], [362.83, 203.43], [392.22, 199.78], [421.44, 199.78], [396.01, 156.15], [362.83, 156.15], [392.22, 158.14], [421.44, 158.15]]}]\n```", - "options": null, - "id": 356 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001614", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[1156.98, 204.97], [1086.6, 203.68], [1077.9, 197.76], [1134.02, 198.6], [1156.98, 111.73], [1086.6, 114.18], [1077.9, 125.45], [1134.02, 123.86]]}]\n```", - "options": null, - "id": 357 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001614", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[688.62, 251.2], [761.87, 255.3], [676.44, 284.34], [584.75, 276.96], [688.62, 186.41], [761.87, 187.12], [676.44, 192.14], [584.75, 190.87]]}, {\"category\": \"car\", \"corners_3d\": [[577.82, 246.81], [635.32, 250.45], [515.08, 273.34], [448.83, 267.33], [577.82, 166.12], [635.32, 165.78], [515.08, 163.7], [448.83, 164.25]]}, {\"category\": \"car\", \"corners_3d\": [[892.92, 220.69], [958.95, 222.59], [952.9, 233.73], [873.06, 230.91], [892.92, 169.22], [958.95, 169.07], [952.9, 168.22], [873.06, 168.44]]}, {\"category\": \"car\", \"corners_3d\": [[418.82, 242.55], [460.9, 245.3], [318.53, 263.64], [273.0, 259.36], [418.82, 190.27], [460.9, 190.96], [318.53, 195.54], [273.0, 194.47]]}, {\"category\": \"car\", \"corners_3d\": [[305.11, 240.19], [339.18, 242.46], [233.58, 253.5], [198.26, 250.48], [305.11, 196.34], [339.18, 197.13], [233.58, 200.98], [198.26, 199.92]]}]\n```", - "options": null, - "id": 358 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001614", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1042.07, 261.75], [1167.96, 268.25], [1243.64, 315.12], [1054.54, 301.12], [1042.07, 151.03], [1167.96, 149.43], [1243.64, 137.92], [1054.54, 141.36]]}, {\"category\": \"van\", \"corners_3d\": [[1256.43, 200.71], [1191.85, 199.74], [1169.67, 195.78], [1223.77, 196.48], [1256.43, 132.29], [1191.85, 133.71], [1169.67, 139.47], [1223.77, 138.45]]}]\n```", - "options": null, - "id": 359 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001614", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[688.62, 251.2], [761.87, 255.3], [676.44, 284.34], [584.75, 276.96], [688.62, 186.41], [761.87, 187.12], [676.44, 192.14], [584.75, 190.87]]}, {\"category\": \"car\", \"corners_3d\": [[577.82, 246.81], [635.32, 250.45], [515.08, 273.34], [448.83, 267.33], [577.82, 166.12], [635.32, 165.78], [515.08, 163.7], [448.83, 164.25]]}, {\"category\": \"car\", \"corners_3d\": [[892.92, 220.69], [958.95, 222.59], [952.9, 233.73], [873.06, 230.91], [892.92, 169.22], [958.95, 169.07], [952.9, 168.22], [873.06, 168.44]]}, {\"category\": \"truck\", \"corners_3d\": [[1156.98, 204.97], [1086.6, 203.68], [1077.9, 197.76], [1134.02, 198.6], [1156.98, 111.73], [1086.6, 114.18], [1077.9, 125.45], [1134.02, 123.86]]}, {\"category\": \"car\", \"corners_3d\": [[418.82, 242.55], [460.9, 245.3], [318.53, 263.64], [273.0, 259.36], [418.82, 190.27], [460.9, 190.96], [318.53, 195.54], [273.0, 194.47]]}, {\"category\": \"car\", \"corners_3d\": [[305.11, 240.19], [339.18, 242.46], [233.58, 253.5], [198.26, 250.48], [305.11, 196.34], [339.18, 197.13], [233.58, 200.98], [198.26, 199.92]]}]\n```", - "options": null, - "id": 360 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001614", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1042.07, 261.75], [1167.96, 268.25], [1243.64, 315.12], [1054.54, 301.12], [1042.07, 151.03], [1167.96, 149.43], [1243.64, 137.92], [1054.54, 141.36]]}, {\"category\": \"van\", \"corners_3d\": [[1256.43, 200.71], [1191.85, 199.74], [1169.67, 195.78], [1223.77, 196.48], [1256.43, 132.29], [1191.85, 133.71], [1169.67, 139.47], [1223.77, 138.45]]}, {\"category\": \"truck\", \"corners_3d\": [[1156.98, 204.97], [1086.6, 203.68], [1077.9, 197.76], [1134.02, 198.6], [1156.98, 111.73], [1086.6, 114.18], [1077.9, 125.45], [1134.02, 123.86]]}]\n```", - "options": null, - "id": 361 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001614", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1042.07, 261.75], [1167.96, 268.25], [1243.64, 315.12], [1054.54, 301.12], [1042.07, 151.03], [1167.96, 149.43], [1243.64, 137.92], [1054.54, 141.36]]}, {\"category\": \"car\", \"corners_3d\": [[688.62, 251.2], [761.87, 255.3], [676.44, 284.34], [584.75, 276.96], [688.62, 186.41], [761.87, 187.12], [676.44, 192.14], [584.75, 190.87]]}, {\"category\": \"car\", \"corners_3d\": [[577.82, 246.81], [635.32, 250.45], [515.08, 273.34], [448.83, 267.33], [577.82, 166.12], [635.32, 165.78], [515.08, 163.7], [448.83, 164.25]]}, {\"category\": \"car\", \"corners_3d\": [[892.92, 220.69], [958.95, 222.59], [952.9, 233.73], [873.06, 230.91], [892.92, 169.22], [958.95, 169.07], [952.9, 168.22], [873.06, 168.44]]}, {\"category\": \"van\", \"corners_3d\": [[1256.43, 200.71], [1191.85, 199.74], [1169.67, 195.78], [1223.77, 196.48], [1256.43, 132.29], [1191.85, 133.71], [1169.67, 139.47], [1223.77, 138.45]]}, {\"category\": \"car\", \"corners_3d\": [[418.82, 242.55], [460.9, 245.3], [318.53, 263.64], [273.0, 259.36], [418.82, 190.27], [460.9, 190.96], [318.53, 195.54], [273.0, 194.47]]}, {\"category\": \"car\", \"corners_3d\": [[305.11, 240.19], [339.18, 242.46], [233.58, 253.5], [198.26, 250.48], [305.11, 196.34], [339.18, 197.13], [233.58, 200.98], [198.26, 199.92]]}]\n```", - "options": null, - "id": 362 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001614", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck, car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1042.07, 261.75], [1167.96, 268.25], [1243.64, 315.12], [1054.54, 301.12], [1042.07, 151.03], [1167.96, 149.43], [1243.64, 137.92], [1054.54, 141.36]]}, {\"category\": \"car\", \"corners_3d\": [[688.62, 251.2], [761.87, 255.3], [676.44, 284.34], [584.75, 276.96], [688.62, 186.41], [761.87, 187.12], [676.44, 192.14], [584.75, 190.87]]}, {\"category\": \"car\", \"corners_3d\": [[577.82, 246.81], [635.32, 250.45], [515.08, 273.34], [448.83, 267.33], [577.82, 166.12], [635.32, 165.78], [515.08, 163.7], [448.83, 164.25]]}, {\"category\": \"car\", \"corners_3d\": [[892.92, 220.69], [958.95, 222.59], [952.9, 233.73], [873.06, 230.91], [892.92, 169.22], [958.95, 169.07], [952.9, 168.22], [873.06, 168.44]]}, {\"category\": \"van\", \"corners_3d\": [[1256.43, 200.71], [1191.85, 199.74], [1169.67, 195.78], [1223.77, 196.48], [1256.43, 132.29], [1191.85, 133.71], [1169.67, 139.47], [1223.77, 138.45]]}, {\"category\": \"truck\", \"corners_3d\": [[1156.98, 204.97], [1086.6, 203.68], [1077.9, 197.76], [1134.02, 198.6], [1156.98, 111.73], [1086.6, 114.18], [1077.9, 125.45], [1134.02, 123.86]]}, {\"category\": \"car\", \"corners_3d\": [[418.82, 242.55], [460.9, 245.3], [318.53, 263.64], [273.0, 259.36], [418.82, 190.27], [460.9, 190.96], [318.53, 195.54], [273.0, 194.47]]}, {\"category\": \"car\", \"corners_3d\": [[305.11, 240.19], [339.18, 242.46], [233.58, 253.5], [198.26, 250.48], [305.11, 196.34], [339.18, 197.13], [233.58, 200.98], [198.26, 199.92]]}]\n```", - "options": null, - "id": 363 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001636", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[745.86, 204.69], [775.46, 205.19], [762.29, 208.44], [729.99, 207.83], [745.86, 180.65], [775.46, 180.77], [762.29, 181.56], [729.99, 181.42]]}, {\"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```", - "options": null, - "id": 364 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001636", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1122.67, 248.43], [1148.86, 251.81], [1059.58, 252.87], [1036.93, 249.4], [1122.67, 161.82], [1148.86, 161.33], [1059.58, 161.17], [1036.93, 161.68]]}]\n```", - "options": null, - "id": 365 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001636", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[557.8, 211.6], [558.65, 211.2], [574.32, 211.23], [573.63, 211.63], [557.8, 180.79], [558.65, 180.7], [574.32, 180.71], [573.63, 180.79]]}]\n```", - "options": null, - "id": 366 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001636", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1122.67, 248.43], [1148.86, 251.81], [1059.58, 252.87], [1036.93, 249.4], [1122.67, 161.82], [1148.86, 161.33], [1059.58, 161.17], [1036.93, 161.68]]}, {\"category\": \"car\", \"corners_3d\": [[745.86, 204.69], [775.46, 205.19], [762.29, 208.44], [729.99, 207.83], [745.86, 180.65], [775.46, 180.77], [762.29, 181.56], [729.99, 181.42]]}, {\"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```", - "options": null, - "id": 367 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001636", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[745.86, 204.69], [775.46, 205.19], [762.29, 208.44], [729.99, 207.83], [745.86, 180.65], [775.46, 180.77], [762.29, 181.56], [729.99, 181.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[557.8, 211.6], [558.65, 211.2], [574.32, 211.23], [573.63, 211.63], [557.8, 180.79], [558.65, 180.7], [574.32, 180.71], [573.63, 180.79]]}, {\"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```", - "options": null, - "id": 368 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001636", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1122.67, 248.43], [1148.86, 251.81], [1059.58, 252.87], [1036.93, 249.4], [1122.67, 161.82], [1148.86, 161.33], [1059.58, 161.17], [1036.93, 161.68]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[557.8, 211.6], [558.65, 211.2], [574.32, 211.23], [573.63, 211.63], [557.8, 180.79], [558.65, 180.7], [574.32, 180.71], [573.63, 180.79]]}]\n```", - "options": null, - "id": 369 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001636", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1122.67, 248.43], [1148.86, 251.81], [1059.58, 252.87], [1036.93, 249.4], [1122.67, 161.82], [1148.86, 161.33], [1059.58, 161.17], [1036.93, 161.68]]}, {\"category\": \"car\", \"corners_3d\": [[745.86, 204.69], [775.46, 205.19], [762.29, 208.44], [729.99, 207.83], [745.86, 180.65], [775.46, 180.77], [762.29, 181.56], [729.99, 181.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[557.8, 211.6], [558.65, 211.2], [574.32, 211.23], [573.63, 211.63], [557.8, 180.79], [558.65, 180.7], [574.32, 180.71], [573.63, 180.79]]}, {\"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```", - "options": null, - "id": 370 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001642", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[503.71, 207.57], [496.69, 209.06], [415.79, 208.73], [426.04, 207.27], [503.71, 183.48], [496.69, 183.93], [415.79, 183.83], [426.04, 183.38]]}]\n```", - "options": null, - "id": 371 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001669", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-402.42, 1036.56], [-1210.32, 1037.33], [106.83, 411.46], [329.65, 411.4], [-402.42, 268.67], [-1210.32, 268.75], [106.83, 199.32], [329.65, 199.32]]}, {\"category\": \"car\", \"corners_3d\": [[559.09, 221.44], [519.54, 221.44], [530.23, 215.64], [565.05, 215.64], [559.09, 187.62], [519.54, 187.62], [530.23, 185.85], [565.05, 185.85]]}]\n```", - "options": null, - "id": 372 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001720", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-74.57, 340.16], [-210.64, 339.86], [49.7, 287.28], [143.19, 287.42], [-74.57, 213.42], [-210.64, 213.34], [49.7, 200.6], [143.19, 200.63]]}, {\"category\": \"car\", \"corners_3d\": [[183.01, 275.35], [100.71, 275.25], [213.33, 252.88], [277.73, 252.94], [183.01, 198.72], [100.71, 198.7], [213.33, 193.05], [277.73, 193.07]]}, {\"category\": \"car\", \"corners_3d\": [[431.29, 216.72], [395.41, 216.68], [422.78, 211.41], [454.37, 211.45], [431.29, 181.92], [395.41, 181.91], [422.78, 180.83], [454.37, 180.83]]}]\n```", - "options": null, - "id": 373 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001720", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[869.67, 198.33], [880.16, 198.33], [896.21, 199.85], [885.1, 199.85], [869.67, 155.93], [880.16, 155.93], [896.21, 154.93], [885.1, 154.93]]}]\n```", - "options": null, - "id": 374 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001720", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[645.45, 183.84], [665.22, 183.81], [712.53, 187.87], [685.57, 187.93], [645.45, 155.7], [665.22, 155.75], [712.53, 149.4], [685.57, 149.31]]}]\n```", - "options": null, - "id": 375 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001720", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-74.57, 340.16], [-210.64, 339.86], [49.7, 287.28], [143.19, 287.42], [-74.57, 213.42], [-210.64, 213.34], [49.7, 200.6], [143.19, 200.63]]}, {\"category\": \"car\", \"corners_3d\": [[183.01, 275.35], [100.71, 275.25], [213.33, 252.88], [277.73, 252.94], [183.01, 198.72], [100.71, 198.7], [213.33, 193.05], [277.73, 193.07]]}, {\"category\": \"car\", \"corners_3d\": [[431.29, 216.72], [395.41, 216.68], [422.78, 211.41], [454.37, 211.45], [431.29, 181.92], [395.41, 181.91], [422.78, 180.83], [454.37, 180.83]]}, {\"category\": \"cyclist\", \"corners_3d\": [[869.67, 198.33], [880.16, 198.33], [896.21, 199.85], [885.1, 199.85], [869.67, 155.93], [880.16, 155.93], [896.21, 154.93], [885.1, 154.93]]}]\n```", - "options": null, - "id": 376 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001720", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[645.45, 183.84], [665.22, 183.81], [712.53, 187.87], [685.57, 187.93], [645.45, 155.7], [665.22, 155.75], [712.53, 149.4], [685.57, 149.31]]}, {\"category\": \"car\", \"corners_3d\": [[-74.57, 340.16], [-210.64, 339.86], [49.7, 287.28], [143.19, 287.42], [-74.57, 213.42], [-210.64, 213.34], [49.7, 200.6], [143.19, 200.63]]}, {\"category\": \"car\", \"corners_3d\": [[183.01, 275.35], [100.71, 275.25], [213.33, 252.88], [277.73, 252.94], [183.01, 198.72], [100.71, 198.7], [213.33, 193.05], [277.73, 193.07]]}, {\"category\": \"car\", \"corners_3d\": [[431.29, 216.72], [395.41, 216.68], [422.78, 211.41], [454.37, 211.45], [431.29, 181.92], [395.41, 181.91], [422.78, 180.83], [454.37, 180.83]]}]\n```", - "options": null, - "id": 377 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001720", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[645.45, 183.84], [665.22, 183.81], [712.53, 187.87], [685.57, 187.93], [645.45, 155.7], [665.22, 155.75], [712.53, 149.4], [685.57, 149.31]]}, {\"category\": \"cyclist\", \"corners_3d\": [[869.67, 198.33], [880.16, 198.33], [896.21, 199.85], [885.1, 199.85], [869.67, 155.93], [880.16, 155.93], [896.21, 154.93], [885.1, 154.93]]}]\n```", - "options": null, - "id": 378 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001720", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist, tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[645.45, 183.84], [665.22, 183.81], [712.53, 187.87], [685.57, 187.93], [645.45, 155.7], [665.22, 155.75], [712.53, 149.4], [685.57, 149.31]]}, {\"category\": \"car\", \"corners_3d\": [[-74.57, 340.16], [-210.64, 339.86], [49.7, 287.28], [143.19, 287.42], [-74.57, 213.42], [-210.64, 213.34], [49.7, 200.6], [143.19, 200.63]]}, {\"category\": \"car\", \"corners_3d\": [[183.01, 275.35], [100.71, 275.25], [213.33, 252.88], [277.73, 252.94], [183.01, 198.72], [100.71, 198.7], [213.33, 193.05], [277.73, 193.07]]}, {\"category\": \"car\", \"corners_3d\": [[431.29, 216.72], [395.41, 216.68], [422.78, 211.41], [454.37, 211.45], [431.29, 181.92], [395.41, 181.91], [422.78, 180.83], [454.37, 180.83]]}, {\"category\": \"cyclist\", \"corners_3d\": [[869.67, 198.33], [880.16, 198.33], [896.21, 199.85], [885.1, 199.85], [869.67, 155.93], [880.16, 155.93], [896.21, 154.93], [885.1, 154.93]]}]\n```", - "options": null, - "id": 379 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001769", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[679.46, 183.25], [698.75, 183.26], [727.7, 186.96], [701.55, 186.95], [679.46, 156.19], [698.75, 156.19], [727.7, 150.26], [701.55, 150.26]]}]\n```", - "options": null, - "id": 380 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001769", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-119.47, 346.82], [-270.0, 347.9], [8.56, 291.03], [109.22, 290.53], [-119.47, 184.54], [-270.0, 184.61], [8.56, 180.79], [109.22, 180.76]]}, {\"category\": \"car\", \"corners_3d\": [[152.89, 279.61], [79.28, 278.46], [201.73, 256.98], [261.3, 257.71], [152.89, 199.8], [79.28, 199.51], [201.73, 194.09], [261.3, 194.28]]}, {\"category\": \"car\", \"corners_3d\": [[283.1, 252.96], [213.19, 252.89], [292.56, 237.13], [348.74, 237.18], [283.1, 181.06], [213.19, 181.05], [292.56, 179.44], [348.74, 179.44]]}, {\"category\": \"car\", \"corners_3d\": [[409.02, 223.91], [365.62, 223.85], [398.48, 217.35], [436.38, 217.39], [409.02, 183.64], [365.62, 183.62], [398.48, 182.25], [436.38, 182.26]]}, {\"category\": \"car\", \"corners_3d\": [[445.68, 214.61], [410.85, 214.51], [435.76, 210.08], [466.93, 210.16], [445.68, 181.91], [410.85, 181.88], [435.76, 180.92], [466.93, 180.94]]}, {\"category\": \"car\", \"corners_3d\": [[517.83, 199.99], [494.87, 199.95], [506.35, 197.78], [527.48, 197.81], [517.83, 177.7], [494.87, 177.69], [506.35, 177.3], [527.48, 177.31]]}]\n```", - "options": null, - "id": 381 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001769", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (tram, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[679.46, 183.25], [698.75, 183.26], [727.7, 186.96], [701.55, 186.95], [679.46, 156.19], [698.75, 156.19], [727.7, 150.26], [701.55, 150.26]]}, {\"category\": \"car\", \"corners_3d\": [[-119.47, 346.82], [-270.0, 347.9], [8.56, 291.03], [109.22, 290.53], [-119.47, 184.54], [-270.0, 184.61], [8.56, 180.79], [109.22, 180.76]]}, {\"category\": \"car\", \"corners_3d\": [[152.89, 279.61], [79.28, 278.46], [201.73, 256.98], [261.3, 257.71], [152.89, 199.8], [79.28, 199.51], [201.73, 194.09], [261.3, 194.28]]}, {\"category\": \"car\", \"corners_3d\": [[283.1, 252.96], [213.19, 252.89], [292.56, 237.13], [348.74, 237.18], [283.1, 181.06], [213.19, 181.05], [292.56, 179.44], [348.74, 179.44]]}, {\"category\": \"car\", \"corners_3d\": [[409.02, 223.91], [365.62, 223.85], [398.48, 217.35], [436.38, 217.39], [409.02, 183.64], [365.62, 183.62], [398.48, 182.25], [436.38, 182.26]]}, {\"category\": \"car\", \"corners_3d\": [[445.68, 214.61], [410.85, 214.51], [435.76, 210.08], [466.93, 210.16], [445.68, 181.91], [410.85, 181.88], [435.76, 180.92], [466.93, 180.94]]}, {\"category\": \"car\", \"corners_3d\": [[517.83, 199.99], [494.87, 199.95], [506.35, 197.78], [527.48, 197.81], [517.83, 177.7], [494.87, 177.69], [506.35, 177.3], [527.48, 177.31]]}]\n```", - "options": null, - "id": 382 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001793", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[536.16, 232.64], [476.46, 232.59], [499.34, 222.82], [549.27, 222.85], [536.16, 185.31], [476.46, 185.3], [499.34, 183.27], [549.27, 183.27]]}]\n```", - "options": null, - "id": 383 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "001994", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[983.39, 296.19], [1005.05, 302.93], [901.77, 302.75], [885.81, 296.04], [983.39, 143.1], [1005.05, 140.93], [901.77, 140.99], [885.81, 143.16]]}]\n```", - "options": null, - "id": 384 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002032", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[595.05, 188.8], [579.23, 188.73], [589.92, 187.84], [604.91, 187.91], [595.05, 173.88], [579.23, 173.87], [589.92, 173.82], [604.91, 173.82]]}, {\"category\": \"car\", \"corners_3d\": [[721.71, 188.98], [739.2, 189.06], [738.06, 190.13], [719.42, 190.04], [721.71, 173.5], [739.2, 173.5], [738.06, 173.55], [719.42, 173.54]]}]\n```", - "options": null, - "id": 385 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002033", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[580.22, 258.61], [578.37, 248.14], [798.27, 246.73], [830.01, 256.79], [580.22, 175.11], [578.37, 174.83], [798.27, 174.79], [830.01, 175.06]]}, {\"category\": \"car\", \"corners_3d\": [[548.16, 205.74], [571.13, 205.63], [575.03, 208.77], [549.87, 208.91], [548.16, 184.19], [571.13, 184.15], [575.03, 185.24], [549.87, 185.28]]}, {\"category\": \"car\", \"corners_3d\": [[481.16, 212.37], [453.67, 212.49], [462.67, 208.83], [487.61, 208.73], [481.16, 189.18], [453.67, 189.23], [462.67, 187.72], [487.61, 187.68]]}]\n```", - "options": null, - "id": 386 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002038", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[-258.82, 353.36], [-221.49, 325.99], [204.12, 286.91], [231.72, 301.43], [-258.82, 147.17], [-221.49, 151.07], [204.12, 156.63], [231.72, 154.56]]}]\n```", - "options": null, - "id": 387 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002038", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1143.9, 242.89], [1234.05, 248.3], [1199.72, 261.37], [1098.1, 254.01], [1143.9, 174.9], [1234.05, 175.06], [1199.72, 175.44], [1098.1, 175.22]]}, {\"category\": \"car\", \"corners_3d\": [[672.23, 255.75], [623.73, 249.88], [734.01, 239.56], [782.79, 243.91], [672.23, 181.27], [623.73, 180.67], [734.01, 179.63], [782.79, 180.07]]}, {\"category\": \"car\", \"corners_3d\": [[923.63, 227.88], [877.5, 225.17], [936.05, 219.78], [979.92, 221.95], [923.63, 181.65], [877.5, 181.21], [936.05, 180.35], [979.92, 180.7]]}, {\"category\": \"car\", \"corners_3d\": [[1049.95, 214.77], [1007.35, 213.09], [1044.17, 209.71], [1084.47, 211.12], [1049.95, 181.61], [1007.35, 181.26], [1044.17, 180.55], [1084.47, 180.85]]}, {\"category\": \"car\", \"corners_3d\": [[1117.2, 210.06], [1079.15, 208.79], [1103.27, 206.53], [1139.65, 207.64], [1117.2, 180.25], [1079.15, 180.0], [1103.27, 179.55], [1139.65, 179.77]]}]\n```", - "options": null, - "id": 388 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002038", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1143.9, 242.89], [1234.05, 248.3], [1199.72, 261.37], [1098.1, 254.01], [1143.9, 174.9], [1234.05, 175.06], [1199.72, 175.44], [1098.1, 175.22]]}, {\"category\": \"van\", \"corners_3d\": [[-258.82, 353.36], [-221.49, 325.99], [204.12, 286.91], [231.72, 301.43], [-258.82, 147.17], [-221.49, 151.07], [204.12, 156.63], [231.72, 154.56]]}, {\"category\": \"car\", \"corners_3d\": [[672.23, 255.75], [623.73, 249.88], [734.01, 239.56], [782.79, 243.91], [672.23, 181.27], [623.73, 180.67], [734.01, 179.63], [782.79, 180.07]]}, {\"category\": \"car\", \"corners_3d\": [[923.63, 227.88], [877.5, 225.17], [936.05, 219.78], [979.92, 221.95], [923.63, 181.65], [877.5, 181.21], [936.05, 180.35], [979.92, 180.7]]}, {\"category\": \"car\", \"corners_3d\": [[1049.95, 214.77], [1007.35, 213.09], [1044.17, 209.71], [1084.47, 211.12], [1049.95, 181.61], [1007.35, 181.26], [1044.17, 180.55], [1084.47, 180.85]]}, {\"category\": \"car\", \"corners_3d\": [[1117.2, 210.06], [1079.15, 208.79], [1103.27, 206.53], [1139.65, 207.64], [1117.2, 180.25], [1079.15, 180.0], [1103.27, 179.55], [1139.65, 179.77]]}]\n```", - "options": null, - "id": 389 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002041", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[332.41, 323.76], [200.33, 325.4], [304.39, 281.95], [398.33, 281.11], [332.41, 178.79], [200.33, 178.86], [304.39, 177.15], [398.33, 177.11]]}, {\"category\": \"car\", \"corners_3d\": [[449.74, 268.36], [355.84, 274.82], [320.3, 253.48], [397.35, 249.39], [449.74, 190.25], [355.84, 191.43], [320.3, 187.54], [397.35, 186.8]]}, {\"category\": \"car\", \"corners_3d\": [[890.58, 181.76], [907.98, 182.02], [859.6, 182.33], [843.06, 182.05], [890.58, 161.1], [907.98, 160.75], [859.6, 160.34], [843.06, 160.71]]}, {\"category\": \"car\", \"corners_3d\": [[1037.32, 362.94], [1270.61, 383.29], [1705.24, 634.7], [1163.57, 546.83], [1037.32, 187.88], [1270.61, 189.49], [1705.24, 209.37], [1163.57, 202.42]]}, {\"category\": \"car\", \"corners_3d\": [[1043.29, 282.96], [1188.04, 295.15], [1181.7, 338.24], [993.43, 316.69], [1043.29, 169.3], [1188.04, 168.91], [1181.7, 167.51], [993.43, 168.21]]}, {\"category\": \"car\", \"corners_3d\": [[727.52, 228.03], [731.51, 223.81], [900.2, 226.51], [910.75, 231.21], [727.52, 163.04], [731.51, 163.79], [900.2, 163.31], [910.75, 162.47]]}, {\"category\": \"car\", \"corners_3d\": [[1092.37, 215.8], [1039.24, 212.94], [1116.64, 210.41], [1171.35, 212.9], [1092.37, 168.55], [1039.24, 168.83], [1116.64, 169.09], [1171.35, 168.84]]}]\n```", - "options": null, - "id": 390 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002064", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[330.33, 325.0], [188.55, 324.72], [323.96, 276.63], [420.96, 276.76], [330.33, 185.94], [188.55, 185.92], [323.96, 181.78], [420.96, 181.79]]}, {\"category\": \"car\", \"corners_3d\": [[660.05, 199.66], [691.87, 199.73], [695.43, 202.23], [660.66, 202.15], [660.05, 169.5], [691.87, 169.49], [695.43, 169.17], [660.66, 169.18]]}]\n```", - "options": null, - "id": 391 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002094", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[815.49, 230.54], [790.68, 230.6], [784.83, 228.94], [808.94, 228.88], [815.49, 172.51], [790.68, 172.51], [784.83, 172.52], [808.94, 172.52]]}]\n```", - "options": null, - "id": 392 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002094", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[34.25, 258.65], [-44.59, 258.75], [76.43, 242.66], [140.41, 242.6], [34.25, 190.86], [-44.59, 190.88], [76.43, 187.5], [140.41, 187.49]]}]\n```", - "options": null, - "id": 393 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002094", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[815.49, 230.54], [790.68, 230.6], [784.83, 228.94], [808.94, 228.88], [815.49, 172.51], [790.68, 172.51], [784.83, 172.52], [808.94, 172.52]]}, {\"category\": \"car\", \"corners_3d\": [[34.25, 258.65], [-44.59, 258.75], [76.43, 242.66], [140.41, 242.6], [34.25, 190.86], [-44.59, 190.88], [76.43, 187.5], [140.41, 187.49]]}]\n```", - "options": null, - "id": 394 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002156", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[317.94, 199.16], [314.92, 198.8], [366.28, 198.14], [370.14, 198.47], [317.94, 140.82], [314.92, 141.57], [366.28, 142.99], [370.14, 142.29]]}, {\"category\": \"cyclist\", \"corners_3d\": [[524.92, 184.48], [514.59, 184.44], [535.72, 184.28], [545.86, 184.32], [524.92, 143.42], [514.59, 143.8], [535.72, 145.21], [545.86, 144.86]]}]\n```", - "options": null, - "id": 395 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002156", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1220.24, 452.43], [1190.82, 476.53], [1013.16, 429.32], [1050.54, 412.07], [1220.24, 130.18], [1190.82, 125.71], [1013.16, 134.45], [1050.54, 137.65]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1174.17, 485.01], [1168.03, 534.64], [976.68, 485.77], [1005.64, 448.16], [1174.17, 146.96], [1168.03, 141.49], [976.68, 146.88], [1005.64, 151.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[695.4, 353.27], [657.98, 364.93], [571.24, 345.62], [609.9, 336.21], [695.4, 140.75], [657.98, 138.07], [571.24, 142.51], [609.9, 144.68]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[615.39, 366.15], [565.18, 387.24], [481.67, 371.02], [536.23, 352.96], [615.39, 130.18], [565.18, 124.46], [481.67, 128.86], [536.23, 133.76]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[637.23, 342.37], [647.45, 329.25], [748.04, 333.96], [746.64, 347.95], [637.23, 155.2], [647.45, 157.25], [748.04, 156.51], [746.64, 154.32]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[575.26, 220.03], [555.83, 220.39], [547.75, 218.66], [566.42, 218.34], [575.26, 145.28], [555.83, 144.96], [547.75, 146.5], [566.42, 146.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[549.85, 221.94], [527.19, 222.13], [526.36, 220.64], [548.21, 220.46], [549.85, 152.18], [527.19, 152.05], [526.36, 153.07], [548.21, 153.19]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[571.52, 203.96], [550.14, 204.07], [548.78, 203.47], [569.63, 203.37], [571.52, 143.03], [550.14, 142.86], [548.78, 143.81], [569.63, 143.98]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[615.36, 203.05], [599.47, 203.19], [593.13, 202.47], [608.55, 202.33], [615.36, 146.0], [599.47, 145.78], [593.13, 146.88], [608.55, 147.09]]}]\n```", - "options": null, - "id": 396 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002156", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1220.24, 452.43], [1190.82, 476.53], [1013.16, 429.32], [1050.54, 412.07], [1220.24, 130.18], [1190.82, 125.71], [1013.16, 134.45], [1050.54, 137.65]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1174.17, 485.01], [1168.03, 534.64], [976.68, 485.77], [1005.64, 448.16], [1174.17, 146.96], [1168.03, 141.49], [976.68, 146.88], [1005.64, 151.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[695.4, 353.27], [657.98, 364.93], [571.24, 345.62], [609.9, 336.21], [695.4, 140.75], [657.98, 138.07], [571.24, 142.51], [609.9, 144.68]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[615.39, 366.15], [565.18, 387.24], [481.67, 371.02], [536.23, 352.96], [615.39, 130.18], [565.18, 124.46], [481.67, 128.86], [536.23, 133.76]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[637.23, 342.37], [647.45, 329.25], [748.04, 333.96], [746.64, 347.95], [637.23, 155.2], [647.45, 157.25], [748.04, 156.51], [746.64, 154.32]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[575.26, 220.03], [555.83, 220.39], [547.75, 218.66], [566.42, 218.34], [575.26, 145.28], [555.83, 144.96], [547.75, 146.5], [566.42, 146.79]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[549.85, 221.94], [527.19, 222.13], [526.36, 220.64], [548.21, 220.46], [549.85, 152.18], [527.19, 152.05], [526.36, 153.07], [548.21, 153.19]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[571.52, 203.96], [550.14, 204.07], [548.78, 203.47], [569.63, 203.37], [571.52, 143.03], [550.14, 142.86], [548.78, 143.81], [569.63, 143.98]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[615.36, 203.05], [599.47, 203.19], [593.13, 202.47], [608.55, 202.33], [615.36, 146.0], [599.47, 145.78], [593.13, 146.88], [608.55, 147.09]]}, {\"category\": \"cyclist\", \"corners_3d\": [[317.94, 199.16], [314.92, 198.8], [366.28, 198.14], [370.14, 198.47], [317.94, 140.82], [314.92, 141.57], [366.28, 142.99], [370.14, 142.29]]}, {\"category\": \"cyclist\", \"corners_3d\": [[524.92, 184.48], [514.59, 184.44], [535.72, 184.28], [545.86, 184.32], [524.92, 143.42], [514.59, 143.8], [535.72, 145.21], [545.86, 144.86]]}]\n```", - "options": null, - "id": 397 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002240", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[280.23, 270.6], [238.29, 283.61], [17.9, 284.28], [86.11, 271.12], [280.23, 191.97], [238.29, 194.52], [17.9, 194.65], [86.11, 192.08]]}, {\"category\": \"car\", \"corners_3d\": [[191.88, 254.26], [233.02, 247.45], [365.91, 248.93], [336.29, 256.03], [191.88, 192.89], [233.02, 191.21], [365.91, 191.58], [336.29, 193.32]]}, {\"category\": \"car\", \"corners_3d\": [[1002.04, 189.1], [967.88, 188.97], [945.66, 187.34], [976.19, 187.45], [1002.04, 160.89], [967.88, 160.99], [945.66, 162.19], [976.19, 162.11]]}]\n```", - "options": null, - "id": 398 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002242", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[525.42, 235.02], [466.99, 234.98], [490.29, 225.28], [539.61, 225.31], [525.42, 187.06], [466.99, 187.05], [490.29, 184.84], [539.61, 184.84]]}]\n```", - "options": null, - "id": 399 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002244", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[569.87, 202.47], [546.35, 202.49], [549.83, 200.35], [571.65, 200.34], [569.87, 179.27], [546.35, 179.27], [549.83, 178.81], [571.65, 178.8]]}]\n```", - "options": null, - "id": 400 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002320", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (motorcyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"motorcyclist\", \"corners_3d\": [[120.53, 558.64], [-33.01, 554.45], [191.13, 428.32], [295.19, 430.19], [120.53, 172.77], [-33.01, 172.77], [191.13, 172.8], [295.19, 172.8]]}]\n```", - "options": null, - "id": 401 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002320", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[648.36, 190.24], [619.9, 190.22], [620.89, 188.62], [646.72, 188.63], [648.36, 163.89], [619.9, 163.9], [620.89, 164.73], [646.72, 164.72]]}]\n```", - "options": null, - "id": 402 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002320", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (motorcyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"motorcyclist\", \"corners_3d\": [[120.53, 558.64], [-33.01, 554.45], [191.13, 428.32], [295.19, 430.19], [120.53, 172.77], [-33.01, 172.77], [191.13, 172.8], [295.19, 172.8]]}, {\"category\": \"car\", \"corners_3d\": [[648.36, 190.24], [619.9, 190.22], [620.89, 188.62], [646.72, 188.63], [648.36, 163.89], [619.9, 163.9], [620.89, 164.73], [646.72, 164.72]]}]\n```", - "options": null, - "id": 403 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002355", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[184.01, 252.95], [127.94, 253.2], [198.25, 240.71], [245.43, 240.53], [184.01, 203.79], [127.94, 203.89], [198.25, 199.06], [245.43, 198.99]]}]\n```", - "options": null, - "id": 404 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002355", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[401.54, 231.23], [458.48, 231.19], [420.55, 246.51], [348.62, 246.57], [401.54, 159.1], [458.48, 159.11], [420.55, 155.51], [348.62, 155.49]]}]\n```", - "options": null, - "id": 405 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002355", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[401.54, 231.23], [458.48, 231.19], [420.55, 246.51], [348.62, 246.57], [401.54, 159.1], [458.48, 159.11], [420.55, 155.51], [348.62, 155.49]]}, {\"category\": \"car\", \"corners_3d\": [[184.01, 252.95], [127.94, 253.2], [198.25, 240.71], [245.43, 240.53], [184.01, 203.79], [127.94, 203.89], [198.25, 199.06], [245.43, 198.99]]}]\n```", - "options": null, - "id": 406 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002361", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[947.72, 302.81], [1093.55, 323.91], [944.01, 374.59], [784.62, 338.64], [947.72, 191.28], [1093.55, 194.27], [944.01, 201.46], [784.62, 196.36]]}, {\"category\": \"car\", \"corners_3d\": [[867.68, 282.03], [958.39, 296.79], [710.76, 330.46], [635.33, 307.35], [867.68, 188.8], [958.39, 190.95], [710.76, 195.87], [635.33, 192.49]]}, {\"category\": \"car\", \"corners_3d\": [[595.28, 252.64], [635.75, 259.53], [457.67, 274.21], [427.31, 264.92], [595.28, 190.71], [635.75, 192.25], [457.67, 195.54], [427.31, 193.46]]}, {\"category\": \"car\", \"corners_3d\": [[357.9, 257.39], [341.28, 251.08], [494.9, 242.28], [520.41, 247.21], [357.9, 180.71], [341.28, 180.13], [494.9, 179.31], [520.41, 179.77]]}]\n```", - "options": null, - "id": 407 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002410", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[574.96, 327.12], [574.12, 318.59], [682.12, 317.29], [689.56, 325.65], [574.96, 148.26], [574.12, 150.13], [682.12, 150.42], [689.56, 148.58]]}]\n```", - "options": null, - "id": 408 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002417", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[578.93, 194.62], [559.17, 194.62], [562.1, 193.34], [580.7, 193.34], [578.93, 178.43], [559.17, 178.43], [562.1, 178.1], [580.7, 178.1]]}]\n```", - "options": null, - "id": 409 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002434", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[59.11, 248.66], [-4.85, 248.6], [100.11, 235.8], [153.33, 235.84], [59.11, 194.08], [-4.85, 194.07], [100.11, 190.48], [153.33, 190.49]]}, {\"category\": \"car\", \"corners_3d\": [[280.47, 220.02], [241.14, 220.02], [275.56, 215.61], [311.2, 215.61], [280.47, 189.88], [241.14, 189.88], [275.56, 188.29], [311.2, 188.29]]}]\n```", - "options": null, - "id": 410 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002469", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[552.77, 195.42], [533.4, 195.43], [537.78, 193.98], [555.9, 193.98], [552.77, 174.96], [533.4, 174.96], [537.78, 174.82], [555.9, 174.82]]}]\n```", - "options": null, - "id": 411 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002515", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[493.42, 292.38], [376.76, 294.68], [410.88, 266.05], [500.03, 264.7], [493.42, 185.76], [376.76, 186.0], [410.88, 182.91], [500.03, 182.77]]}, {\"category\": \"car\", \"corners_3d\": [[515.3, 248.13], [454.27, 248.14], [482.19, 234.55], [532.21, 234.54], [515.3, 185.92], [454.27, 185.92], [482.19, 183.56], [532.21, 183.56]]}]\n```", - "options": null, - "id": 412 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002518", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[479.5, 243.44], [417.85, 243.57], [447.72, 231.61], [498.92, 231.53], [479.5, 182.82], [417.85, 182.84], [447.72, 181.15], [498.92, 181.14]]}]\n```", - "options": null, - "id": 413 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002550", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[356.66, 458.45], [42.88, 469.41], [268.05, 336.12], [439.06, 332.74], [356.66, 181.2], [42.88, 181.52], [268.05, 177.62], [439.06, 177.52]]}, {\"category\": \"car\", \"corners_3d\": [[491.96, 282.24], [394.58, 281.95], [454.23, 253.52], [526.29, 253.67], [491.96, 182.37], [394.58, 182.35], [454.23, 179.87], [526.29, 179.89]]}]\n```", - "options": null, - "id": 414 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002608", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[519.41, 255.0], [453.32, 256.04], [461.34, 243.69], [517.63, 242.93], [519.41, 193.93], [453.32, 194.2], [461.34, 191.03], [517.63, 190.84]]}, {\"category\": \"car\", \"corners_3d\": [[521.57, 236.35], [479.45, 236.42], [494.6, 228.05], [531.16, 228.0], [521.57, 193.56], [479.45, 193.58], [494.6, 190.85], [531.16, 190.83]]}]\n```", - "options": null, - "id": 415 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002660", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[336.23, 251.78], [302.38, 260.22], [104.92, 258.74], [157.03, 250.57], [336.23, 195.27], [302.38, 197.67], [104.92, 197.25], [157.03, 194.93]]}]\n```", - "options": null, - "id": 416 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002672", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[697.5, 206.09], [721.9, 206.34], [717.33, 209.02], [691.02, 208.72], [697.5, 183.24], [721.9, 183.32], [717.33, 184.15], [691.02, 184.06]]}, {\"category\": \"car\", \"corners_3d\": [[-12.87, 362.69], [-152.88, 357.77], [130.91, 295.95], [228.27, 298.11], [-12.87, 204.29], [-152.88, 203.48], [130.91, 193.24], [228.27, 193.6]]}, {\"category\": \"car\", \"corners_3d\": [[257.19, 292.18], [165.81, 290.25], [300.11, 260.52], [369.7, 261.58], [257.19, 171.58], [165.81, 171.6], [300.11, 171.92], [369.7, 171.91]]}, {\"category\": \"car\", \"corners_3d\": [[723.58, 241.23], [785.76, 241.41], [823.57, 258.25], [746.06, 257.97], [723.58, 182.82], [785.76, 182.85], [823.57, 185.3], [746.06, 185.26]]}, {\"category\": \"car\", \"corners_3d\": [[656.37, 209.16], [629.29, 208.83], [641.35, 206.05], [666.42, 206.33], [656.37, 183.79], [629.29, 183.69], [641.35, 182.86], [666.42, 182.94]]}, {\"category\": \"car\", \"corners_3d\": [[723.85, 230.19], [769.18, 230.55], [785.43, 241.37], [731.54, 240.86], [723.85, 169.28], [769.18, 169.26], [785.43, 168.58], [731.54, 168.61]]}]\n```", - "options": null, - "id": 417 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002750", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1110.63, 199.86], [1067.93, 199.08], [1075.33, 196.59], [1114.06, 197.23], [1110.63, 156.5], [1067.93, 156.98], [1075.33, 158.48], [1114.06, 158.09]]}]\n```", - "options": null, - "id": 418 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002750", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[867.05, 247.29], [803.42, 242.98], [872.67, 231.56], [928.94, 234.55], [867.05, 180.51], [803.42, 180.07], [872.67, 178.89], [928.94, 179.2]]}, {\"category\": \"car\", \"corners_3d\": [[1151.69, 204.03], [1109.85, 203.09], [1118.74, 200.83], [1157.62, 201.64], [1151.69, 175.03], [1109.85, 174.97], [1118.74, 174.81], [1157.62, 174.86]]}, {\"category\": \"car\", \"corners_3d\": [[945.9, 233.96], [885.38, 230.82], [930.94, 223.13], [985.18, 225.47], [945.9, 183.21], [885.38, 182.67], [930.94, 181.37], [985.18, 181.77]]}, {\"category\": \"car\", \"corners_3d\": [[1010.21, 218.11], [968.72, 216.57], [994.63, 212.42], [1032.87, 213.67], [1010.21, 180.8], [968.72, 180.53], [994.63, 179.8], [1032.87, 180.02]]}, {\"category\": \"car\", \"corners_3d\": [[1082.3, 205.78], [1044.71, 204.92], [1050.69, 203.02], [1086.15, 203.78], [1082.3, 177.55], [1044.71, 177.43], [1050.69, 177.16], [1086.15, 177.27]]}]\n```", - "options": null, - "id": 419 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002750", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[867.05, 247.29], [803.42, 242.98], [872.67, 231.56], [928.94, 234.55], [867.05, 180.51], [803.42, 180.07], [872.67, 178.89], [928.94, 179.2]]}, {\"category\": \"car\", \"corners_3d\": [[1151.69, 204.03], [1109.85, 203.09], [1118.74, 200.83], [1157.62, 201.64], [1151.69, 175.03], [1109.85, 174.97], [1118.74, 174.81], [1157.62, 174.86]]}, {\"category\": \"car\", \"corners_3d\": [[945.9, 233.96], [885.38, 230.82], [930.94, 223.13], [985.18, 225.47], [945.9, 183.21], [885.38, 182.67], [930.94, 181.37], [985.18, 181.77]]}, {\"category\": \"car\", \"corners_3d\": [[1010.21, 218.11], [968.72, 216.57], [994.63, 212.42], [1032.87, 213.67], [1010.21, 180.8], [968.72, 180.53], [994.63, 179.8], [1032.87, 180.02]]}, {\"category\": \"van\", \"corners_3d\": [[1110.63, 199.86], [1067.93, 199.08], [1075.33, 196.59], [1114.06, 197.23], [1110.63, 156.5], [1067.93, 156.98], [1075.33, 158.48], [1114.06, 158.09]]}, {\"category\": \"car\", \"corners_3d\": [[1082.3, 205.78], [1044.71, 204.92], [1050.69, 203.02], [1086.15, 203.78], [1082.3, 177.55], [1044.71, 177.43], [1050.69, 177.16], [1086.15, 177.27]]}]\n```", - "options": null, - "id": 420 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002766", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[564.73, 205.94], [533.54, 205.99], [538.63, 202.85], [566.86, 202.82], [564.73, 173.18], [533.54, 173.18], [538.63, 173.15], [566.86, 173.15]]}, {\"category\": \"car\", \"corners_3d\": [[574.96, 193.71], [558.7, 193.7], [562.15, 192.59], [577.54, 192.59], [574.96, 175.62], [558.7, 175.62], [562.15, 175.47], [577.54, 175.47]]}]\n```", - "options": null, - "id": 421 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002769", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[501.91, 300.39], [507.79, 294.39], [592.67, 294.68], [591.26, 300.71], [501.91, 158.84], [507.79, 159.92], [592.67, 159.87], [591.26, 158.78]]}]\n```", - "options": null, - "id": 422 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002770", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[192.43, 371.4], [375.66, 371.95], [-142.44, 798.74], [-710.59, 793.36], [192.43, 185.43], [375.66, 185.46], [-142.44, 212.48], [-710.59, 212.14]]}, {\"category\": \"car\", \"corners_3d\": [[495.91, 271.35], [404.99, 271.11], [454.93, 248.64], [525.12, 248.78], [495.91, 184.26], [404.99, 184.24], [454.93, 181.63], [525.12, 181.65]]}, {\"category\": \"car\", \"corners_3d\": [[581.7, 204.66], [553.33, 204.65], [558.91, 201.83], [584.76, 201.83], [581.7, 180.37], [553.33, 180.37], [558.91, 179.7], [584.76, 179.71]]}]\n```", - "options": null, - "id": 423 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002784", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[562.64, 211.17], [536.37, 211.34], [535.22, 208.28], [559.41, 208.14], [562.64, 184.23], [536.37, 184.28], [535.22, 183.37], [559.41, 183.33]]}]\n```", - "options": null, - "id": 424 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002787", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[945.08, 271.92], [820.87, 273.54], [734.29, 242.07], [821.08, 241.3], [945.08, 138.56], [820.87, 137.99], [734.29, 148.89], [821.08, 149.16]]}, {\"category\": \"car\", \"corners_3d\": [[383.94, 291.65], [280.15, 292.67], [359.6, 259.33], [434.19, 258.8], [383.94, 193.2], [280.15, 193.37], [359.6, 187.66], [434.19, 187.57]]}, {\"category\": \"car\", \"corners_3d\": [[480.65, 234.65], [421.83, 235.11], [441.01, 225.32], [490.52, 225.0], [480.65, 187.08], [421.83, 187.18], [441.01, 184.93], [490.52, 184.86]]}, {\"category\": \"car\", \"corners_3d\": [[503.27, 219.64], [455.95, 219.82], [470.59, 213.56], [511.56, 213.42], [503.27, 173.41], [455.95, 173.41], [470.59, 173.34], [511.56, 173.34]]}, {\"category\": \"car\", \"corners_3d\": [[588.56, 193.89], [613.27, 193.85], [616.81, 195.45], [590.22, 195.5], [588.56, 170.26], [613.27, 170.26], [616.81, 170.07], [590.22, 170.06]]}]\n```", - "options": null, - "id": 425 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002787", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[581.0, 193.77], [587.62, 193.61], [591.92, 193.77], [585.29, 193.93], [581.0, 165.65], [587.62, 165.7], [591.92, 165.65], [585.29, 165.6]]}]\n```", - "options": null, - "id": 426 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002787", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[945.08, 271.92], [820.87, 273.54], [734.29, 242.07], [821.08, 241.3], [945.08, 138.56], [820.87, 137.99], [734.29, 148.89], [821.08, 149.16]]}, {\"category\": \"car\", \"corners_3d\": [[383.94, 291.65], [280.15, 292.67], [359.6, 259.33], [434.19, 258.8], [383.94, 193.2], [280.15, 193.37], [359.6, 187.66], [434.19, 187.57]]}, {\"category\": \"car\", \"corners_3d\": [[480.65, 234.65], [421.83, 235.11], [441.01, 225.32], [490.52, 225.0], [480.65, 187.08], [421.83, 187.18], [441.01, 184.93], [490.52, 184.86]]}, {\"category\": \"car\", \"corners_3d\": [[503.27, 219.64], [455.95, 219.82], [470.59, 213.56], [511.56, 213.42], [503.27, 173.41], [455.95, 173.41], [470.59, 173.34], [511.56, 173.34]]}, {\"category\": \"car\", \"corners_3d\": [[588.56, 193.89], [613.27, 193.85], [616.81, 195.45], [590.22, 195.5], [588.56, 170.26], [613.27, 170.26], [616.81, 170.07], [590.22, 170.06]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[581.0, 193.77], [587.62, 193.61], [591.92, 193.77], [585.29, 193.93], [581.0, 165.65], [587.62, 165.7], [591.92, 165.65], [585.29, 165.6]]}]\n```", - "options": null, - "id": 427 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002791", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[622.29, 211.39], [652.92, 211.67], [642.33, 216.74], [607.82, 216.38], [622.29, 184.71], [652.92, 184.79], [642.33, 186.35], [607.82, 186.24]]}]\n```", - "options": null, - "id": 428 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002791", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[677.08, 216.22], [711.83, 216.61], [706.39, 223.04], [666.64, 222.52], [677.08, 175.28], [711.83, 175.3], [706.39, 175.66], [666.64, 175.63]]}]\n```", - "options": null, - "id": 429 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002791", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[604.77, 203.05], [581.18, 203.51], [532.26, 198.7], [552.83, 198.37], [604.77, 168.57], [581.18, 168.5], [532.26, 169.18], [552.83, 169.23]]}]\n```", - "options": null, - "id": 430 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002791", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[622.29, 211.39], [652.92, 211.67], [642.33, 216.74], [607.82, 216.38], [622.29, 184.71], [652.92, 184.79], [642.33, 186.35], [607.82, 186.24]]}, {\"category\": \"van\", \"corners_3d\": [[677.08, 216.22], [711.83, 216.61], [706.39, 223.04], [666.64, 222.52], [677.08, 175.28], [711.83, 175.3], [706.39, 175.66], [666.64, 175.63]]}]\n```", - "options": null, - "id": 431 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002791", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[622.29, 211.39], [652.92, 211.67], [642.33, 216.74], [607.82, 216.38], [622.29, 184.71], [652.92, 184.79], [642.33, 186.35], [607.82, 186.24]]}, {\"category\": \"tram\", \"corners_3d\": [[604.77, 203.05], [581.18, 203.51], [532.26, 198.7], [552.83, 198.37], [604.77, 168.57], [581.18, 168.5], [532.26, 169.18], [552.83, 169.23]]}]\n```", - "options": null, - "id": 432 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002791", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[677.08, 216.22], [711.83, 216.61], [706.39, 223.04], [666.64, 222.52], [677.08, 175.28], [711.83, 175.3], [706.39, 175.66], [666.64, 175.63]]}, {\"category\": \"tram\", \"corners_3d\": [[604.77, 203.05], [581.18, 203.51], [532.26, 198.7], [552.83, 198.37], [604.77, 168.57], [581.18, 168.5], [532.26, 169.18], [552.83, 169.23]]}]\n```", - "options": null, - "id": 433 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002791", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van, tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[622.29, 211.39], [652.92, 211.67], [642.33, 216.74], [607.82, 216.38], [622.29, 184.71], [652.92, 184.79], [642.33, 186.35], [607.82, 186.24]]}, {\"category\": \"van\", \"corners_3d\": [[677.08, 216.22], [711.83, 216.61], [706.39, 223.04], [666.64, 222.52], [677.08, 175.28], [711.83, 175.3], [706.39, 175.66], [666.64, 175.63]]}, {\"category\": \"tram\", \"corners_3d\": [[604.77, 203.05], [581.18, 203.51], [532.26, 198.7], [552.83, 198.37], [604.77, 168.57], [581.18, 168.5], [532.26, 169.18], [552.83, 169.23]]}]\n```", - "options": null, - "id": 434 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002815", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[750.21, 287.98], [872.55, 287.99], [1071.52, 375.3], [856.4, 375.25], [750.21, 162.1], [872.55, 162.1], [1071.52, 153.94], [856.4, 153.95]]}, {\"category\": \"car\", \"corners_3d\": [[119.18, 375.1], [-70.82, 455.83], [-857.88, 461.89], [-437.03, 378.17], [119.18, 192.08], [-70.82, 199.75], [-857.88, 200.33], [-437.03, 192.37]]}, {\"category\": \"car\", \"corners_3d\": [[-269.3, 355.89], [-94.99, 320.79], [230.37, 323.68], [130.9, 360.34], [-269.3, 197.3], [-94.99, 192.61], [230.37, 193.0], [130.9, 197.89]]}, {\"category\": \"car\", \"corners_3d\": [[-23.06, 296.69], [70.88, 278.74], [298.38, 279.48], [242.56, 297.69], [-23.06, 180.57], [70.88, 179.45], [298.38, 179.5], [242.56, 180.63]]}, {\"category\": \"car\", \"corners_3d\": [[421.7, 251.36], [396.47, 260.81], [158.92, 259.93], [209.22, 250.66], [421.7, 159.12], [396.47, 157.47], [158.92, 157.62], [209.22, 159.24]]}, {\"category\": \"car\", \"corners_3d\": [[116.62, 277.15], [177.26, 264.53], [359.35, 264.8], [323.66, 277.5], [116.62, 172.83], [177.26, 172.83], [359.35, 172.83], [323.66, 172.83]]}, {\"category\": \"car\", \"corners_3d\": [[708.5, 248.13], [788.29, 247.8], [845.89, 268.65], [744.1, 269.19], [708.5, 178.18], [788.29, 178.16], [845.89, 179.63], [744.1, 179.67]]}, {\"category\": \"car\", \"corners_3d\": [[233.57, 246.17], [268.17, 239.44], [413.23, 239.46], [393.28, 246.19], [233.57, 161.52], [268.17, 162.56], [413.23, 162.56], [393.28, 161.52]]}, {\"category\": \"car\", \"corners_3d\": [[450.81, 238.0], [428.85, 243.65], [296.4, 241.89], [328.18, 236.51], [450.81, 178.8], [428.85, 179.31], [296.4, 179.15], [328.18, 178.66]]}, {\"category\": \"car\", \"corners_3d\": [[303.66, 230.78], [328.33, 226.52], [466.25, 226.96], [452.39, 231.28], [303.66, 171.11], [328.33, 171.24], [466.25, 171.22], [452.39, 171.09]]}, {\"category\": \"car\", \"corners_3d\": [[503.82, 206.71], [495.33, 208.55], [393.21, 208.19], [406.83, 206.39], [503.82, 168.73], [495.33, 168.51], [393.21, 168.55], [406.83, 168.77]]}, {\"category\": \"car\", \"corners_3d\": [[429.63, 204.05], [438.21, 202.63], [518.12, 202.66], [513.34, 204.09], [429.63, 163.1], [438.21, 163.54], [518.12, 163.53], [513.34, 163.08]]}, {\"category\": \"car\", \"corners_3d\": [[651.89, 206.08], [687.05, 206.05], [698.35, 210.16], [658.86, 210.2], [651.89, 173.95], [687.05, 173.95], [698.35, 174.09], [658.86, 174.09]]}, {\"category\": \"car\", \"corners_3d\": [[682.87, 228.79], [741.27, 228.8], [770.12, 241.1], [698.86, 241.1], [682.87, 174.94], [741.27, 174.94], [770.12, 175.4], [698.86, 175.4]]}, {\"category\": \"car\", \"corners_3d\": [[639.37, 194.81], [668.33, 194.82], [673.68, 196.83], [642.06, 196.83], [639.37, 167.19], [668.33, 167.19], [673.68, 166.67], [642.06, 166.67]]}]\n```", - "options": null, - "id": 435 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002829", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[576.73, 296.87], [545.79, 299.64], [502.38, 282.54], [530.06, 280.46], [576.73, 169.26], [545.79, 169.18], [502.38, 169.68], [530.06, 169.74]]}]\n```", - "options": null, - "id": 436 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002829", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[443.39, 380.9], [271.73, 381.93], [393.13, 303.3], [500.06, 302.9], [443.39, 216.81], [271.73, 217.02], [393.13, 200.41], [500.06, 200.33]]}, {\"category\": \"car\", \"corners_3d\": [[800.91, 239.2], [841.7, 244.65], [702.09, 251.45], [669.28, 244.96], [800.91, 173.65], [841.7, 173.71], [702.09, 173.8], [669.28, 173.72]]}]\n```", - "options": null, - "id": 437 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002829", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[443.39, 380.9], [271.73, 381.93], [393.13, 303.3], [500.06, 302.9], [443.39, 216.81], [271.73, 217.02], [393.13, 200.41], [500.06, 200.33]]}, {\"category\": \"cyclist\", \"corners_3d\": [[576.73, 296.87], [545.79, 299.64], [502.38, 282.54], [530.06, 280.46], [576.73, 169.26], [545.79, 169.18], [502.38, 169.68], [530.06, 169.74]]}, {\"category\": \"car\", \"corners_3d\": [[800.91, 239.2], [841.7, 244.65], [702.09, 251.45], [669.28, 244.96], [800.91, 173.65], [841.7, 173.71], [702.09, 173.8], [669.28, 173.72]]}]\n```", - "options": null, - "id": 438 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002835", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[570.7, 200.79], [546.35, 200.77], [552.15, 198.67], [574.67, 198.69], [570.7, 180.37], [546.35, 180.36], [552.15, 179.8], [574.67, 179.8]]}]\n```", - "options": null, - "id": 439 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002836", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[900.78, 204.11], [899.78, 202.92], [971.14, 203.87], [975.09, 205.14], [900.78, 176.87], [899.78, 176.72], [971.14, 176.84], [975.09, 177.01]]}, {\"category\": \"car\", \"corners_3d\": [[362.94, 207.61], [340.05, 207.6], [360.71, 205.0], [381.88, 205.01], [362.94, 190.1], [340.05, 190.1], [360.71, 188.81], [381.88, 188.81]]}]\n```", - "options": null, - "id": 440 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002836", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (bus). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[592.52, 190.92], [615.59, 190.92], [617.83, 194.11], [590.69, 194.11], [592.52, 165.92], [615.59, 165.92], [617.83, 164.7], [590.69, 164.7]]}]\n```", - "options": null, - "id": 441 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002836", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, bus). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[592.52, 190.92], [615.59, 190.92], [617.83, 194.11], [590.69, 194.11], [592.52, 165.92], [615.59, 165.92], [617.83, 164.7], [590.69, 164.7]]}, {\"category\": \"car\", \"corners_3d\": [[900.78, 204.11], [899.78, 202.92], [971.14, 203.87], [975.09, 205.14], [900.78, 176.87], [899.78, 176.72], [971.14, 176.84], [975.09, 177.01]]}, {\"category\": \"car\", \"corners_3d\": [[362.94, 207.61], [340.05, 207.6], [360.71, 205.0], [381.88, 205.01], [362.94, 190.1], [340.05, 190.1], [360.71, 188.81], [381.88, 188.81]]}]\n```", - "options": null, - "id": 442 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002879", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[543.49, 240.78], [472.58, 240.98], [496.89, 226.55], [552.76, 226.42], [543.49, 177.6], [472.58, 177.61], [496.89, 176.6], [552.76, 176.6]]}, {\"category\": \"car\", \"corners_3d\": [[572.85, 210.38], [526.6, 210.12], [548.67, 205.04], [588.71, 205.24], [572.85, 166.37], [526.6, 166.42], [548.67, 167.29], [588.71, 167.26]]}, {\"category\": \"car\", \"corners_3d\": [[646.49, 194.04], [664.11, 194.64], [630.5, 195.52], [613.16, 194.86], [646.49, 168.25], [664.11, 168.12], [630.5, 167.93], [613.16, 168.07]]}, {\"category\": \"car\", \"corners_3d\": [[-571.54, 398.14], [-284.51, 341.25], [193.93, 334.85], [74.84, 386.83], [-571.54, 184.43], [-284.51, 181.51], [193.93, 181.18], [74.84, 183.85]]}]\n```", - "options": null, - "id": 443 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002879", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[440.94, 312.28], [276.69, 312.63], [385.2, 265.97], [494.53, 265.82], [440.94, 147.95], [276.69, 147.89], [385.2, 156.22], [494.53, 156.25]]}, {\"category\": \"van\", \"corners_3d\": [[730.75, 276.75], [831.89, 276.62], [941.56, 326.31], [792.09, 326.61], [730.75, 166.61], [831.89, 166.62], [941.56, 163.63], [792.09, 163.61]]}, {\"category\": \"van\", \"corners_3d\": [[559.02, 221.84], [504.93, 221.84], [521.68, 213.96], [567.06, 213.96], [559.02, 162.98], [504.93, 162.98], [521.68, 164.57], [567.06, 164.57]]}]\n```", - "options": null, - "id": 444 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002879", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[440.94, 312.28], [276.69, 312.63], [385.2, 265.97], [494.53, 265.82], [440.94, 147.95], [276.69, 147.89], [385.2, 156.22], [494.53, 156.25]]}, {\"category\": \"van\", \"corners_3d\": [[730.75, 276.75], [831.89, 276.62], [941.56, 326.31], [792.09, 326.61], [730.75, 166.61], [831.89, 166.62], [941.56, 163.63], [792.09, 163.61]]}, {\"category\": \"car\", \"corners_3d\": [[543.49, 240.78], [472.58, 240.98], [496.89, 226.55], [552.76, 226.42], [543.49, 177.6], [472.58, 177.61], [496.89, 176.6], [552.76, 176.6]]}, {\"category\": \"van\", \"corners_3d\": [[559.02, 221.84], [504.93, 221.84], [521.68, 213.96], [567.06, 213.96], [559.02, 162.98], [504.93, 162.98], [521.68, 164.57], [567.06, 164.57]]}, {\"category\": \"car\", \"corners_3d\": [[572.85, 210.38], [526.6, 210.12], [548.67, 205.04], [588.71, 205.24], [572.85, 166.37], [526.6, 166.42], [548.67, 167.29], [588.71, 167.26]]}, {\"category\": \"car\", \"corners_3d\": [[646.49, 194.04], [664.11, 194.64], [630.5, 195.52], [613.16, 194.86], [646.49, 168.25], [664.11, 168.12], [630.5, 167.93], [613.16, 168.07]]}, {\"category\": \"car\", \"corners_3d\": [[-571.54, 398.14], [-284.51, 341.25], [193.93, 334.85], [74.84, 386.83], [-571.54, 184.43], [-284.51, 181.51], [193.93, 181.18], [74.84, 183.85]]}]\n```", - "options": null, - "id": 445 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002910", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[648.48, 232.4], [702.8, 232.0], [743.58, 247.28], [675.45, 247.92], [648.48, 181.98], [702.8, 181.91], [743.58, 184.26], [675.45, 184.35]]}, {\"category\": \"car\", \"corners_3d\": [[798.75, 304.85], [935.78, 302.29], [1220.48, 394.87], [991.88, 402.52], [798.75, 165.87], [935.78, 166.01], [1220.48, 161.11], [991.88, 160.71]]}]\n```", - "options": null, - "id": 446 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002941", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[555.58, 215.8], [521.17, 216.05], [521.25, 211.66], [552.18, 211.46], [555.58, 180.6], [521.17, 180.64], [521.25, 179.85], [552.18, 179.82]]}, {\"category\": \"car\", \"corners_3d\": [[524.24, 202.77], [500.35, 203.02], [493.97, 200.71], [516.09, 200.5], [524.24, 182.22], [500.35, 182.3], [493.97, 181.57], [516.09, 181.5]]}]\n```", - "options": null, - "id": 447 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002958", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[745.33, 270.83], [846.22, 270.7], [999.5, 332.34], [835.24, 332.68], [745.33, 180.52], [846.22, 180.51], [999.5, 185.33], [835.24, 185.35]]}]\n```", - "options": null, - "id": 448 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002959", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[691.13, 243.47], [767.98, 243.47], [836.37, 274.07], [726.22, 274.06], [691.13, 151.11], [767.98, 151.11], [836.37, 141.69], [726.22, 141.69]]}]\n```", - "options": null, - "id": 449 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002959", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[933.44, 415.23], [1148.44, 415.29], [2585.57, 1062.51], [1795.82, 1061.73], [933.44, 207.24], [1148.44, 207.25], [2585.57, 299.07], [1795.82, 298.96]]}, {\"category\": \"car\", \"corners_3d\": [[765.17, 293.89], [880.32, 293.71], [1039.47, 363.05], [858.47, 363.49], [765.17, 185.9], [880.32, 185.88], [1039.47, 193.35], [858.47, 193.4]]}, {\"category\": \"car\", \"corners_3d\": [[524.83, 236.03], [467.08, 236.38], [482.54, 225.63], [530.49, 225.38], [524.83, 178.17], [467.08, 178.2], [482.54, 177.29], [530.49, 177.27]]}, {\"category\": \"car\", \"corners_3d\": [[668.01, 217.29], [707.89, 217.27], [723.29, 223.79], [677.57, 223.82], [668.01, 181.38], [707.89, 181.38], [723.29, 182.63], [677.57, 182.63]]}, {\"category\": \"car\", \"corners_3d\": [[552.59, 208.82], [516.26, 208.88], [524.44, 204.73], [556.58, 204.69], [552.59, 179.45], [516.26, 179.46], [524.44, 178.7], [556.58, 178.69]]}, {\"category\": \"car\", \"corners_3d\": [[568.43, 202.34], [539.21, 202.39], [542.59, 199.95], [569.4, 199.91], [568.43, 174.84], [539.21, 174.84], [542.59, 174.68], [569.4, 174.67]]}, {\"category\": \"car\", \"corners_3d\": [[679.73, 204.64], [650.27, 204.66], [645.62, 201.61], [672.26, 201.6], [679.73, 177.67], [650.27, 177.67], [645.62, 177.21], [672.26, 177.21]]}]\n```", - "options": null, - "id": 450 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "002959", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[933.44, 415.23], [1148.44, 415.29], [2585.57, 1062.51], [1795.82, 1061.73], [933.44, 207.24], [1148.44, 207.25], [2585.57, 299.07], [1795.82, 298.96]]}, {\"category\": \"car\", \"corners_3d\": [[765.17, 293.89], [880.32, 293.71], [1039.47, 363.05], [858.47, 363.49], [765.17, 185.9], [880.32, 185.88], [1039.47, 193.35], [858.47, 193.4]]}, {\"category\": \"van\", \"corners_3d\": [[691.13, 243.47], [767.98, 243.47], [836.37, 274.07], [726.22, 274.06], [691.13, 151.11], [767.98, 151.11], [836.37, 141.69], [726.22, 141.69]]}, {\"category\": \"car\", \"corners_3d\": [[524.83, 236.03], [467.08, 236.38], [482.54, 225.63], [530.49, 225.38], [524.83, 178.17], [467.08, 178.2], [482.54, 177.29], [530.49, 177.27]]}, {\"category\": \"car\", \"corners_3d\": [[668.01, 217.29], [707.89, 217.27], [723.29, 223.79], [677.57, 223.82], [668.01, 181.38], [707.89, 181.38], [723.29, 182.63], [677.57, 182.63]]}, {\"category\": \"car\", \"corners_3d\": [[552.59, 208.82], [516.26, 208.88], [524.44, 204.73], [556.58, 204.69], [552.59, 179.45], [516.26, 179.46], [524.44, 178.7], [556.58, 178.69]]}, {\"category\": \"car\", \"corners_3d\": [[568.43, 202.34], [539.21, 202.39], [542.59, 199.95], [569.4, 199.91], [568.43, 174.84], [539.21, 174.84], [542.59, 174.68], [569.4, 174.67]]}, {\"category\": \"car\", \"corners_3d\": [[679.73, 204.64], [650.27, 204.66], [645.62, 201.61], [672.26, 201.6], [679.73, 177.67], [650.27, 177.67], [645.62, 177.21], [672.26, 177.21]]}]\n```", - "options": null, - "id": 451 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003015", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[420.66, 208.9], [453.56, 208.52], [455.88, 212.68], [419.14, 213.15], [420.66, 173.55], [453.56, 173.55], [455.88, 173.63], [419.14, 173.64]]}, {\"category\": \"van\", \"corners_3d\": [[307.05, 260.68], [233.57, 262.83], [267.97, 245.96], [327.27, 244.53], [307.05, 191.58], [233.57, 192.04], [267.97, 188.44], [327.27, 188.14]]}]\n```", - "options": null, - "id": 452 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003015", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[516.56, 264.92], [557.28, 263.64], [575.07, 275.94], [529.03, 277.59], [516.56, 177.36], [557.28, 177.3], [575.07, 177.9], [529.03, 177.99]]}]\n```", - "options": null, - "id": 453 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003015", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[791.71, 272.02], [901.28, 277.79], [894.44, 328.99], [736.2, 316.56], [791.71, 181.25], [901.28, 181.74], [894.44, 186.07], [736.2, 185.02]]}, {\"category\": \"car\", \"corners_3d\": [[597.12, 253.32], [666.68, 251.27], [744.57, 275.86], [655.15, 279.42], [597.12, 185.4], [666.68, 185.08], [744.57, 188.91], [655.15, 189.47]]}, {\"category\": \"car\", \"corners_3d\": [[255.14, 287.95], [153.88, 291.08], [224.23, 263.47], [300.89, 261.62], [255.14, 206.5], [153.88, 207.41], [224.23, 199.34], [300.89, 198.8]]}, {\"category\": \"car\", \"corners_3d\": [[539.46, 234.34], [593.66, 233.19], [624.07, 244.13], [560.51, 245.74], [539.46, 182.0], [593.66, 181.83], [624.07, 183.46], [560.51, 183.7]]}, {\"category\": \"car\", \"corners_3d\": [[514.5, 223.18], [559.71, 222.25], [586.76, 229.98], [534.91, 231.24], [514.5, 182.48], [559.71, 182.31], [586.76, 183.79], [534.91, 184.03]]}, {\"category\": \"car\", \"corners_3d\": [[332.35, 237.47], [278.96, 238.76], [294.53, 229.8], [340.53, 228.84], [332.35, 193.37], [278.96, 193.77], [294.53, 190.93], [340.53, 190.63]]}, {\"category\": \"car\", \"corners_3d\": [[492.08, 219.01], [525.86, 218.42], [540.05, 223.67], [502.54, 224.41], [492.08, 188.16], [525.86, 187.96], [540.05, 189.7], [502.54, 189.95]]}]\n```", - "options": null, - "id": 454 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003015", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[420.66, 208.9], [453.56, 208.52], [455.88, 212.68], [419.14, 213.15], [420.66, 173.55], [453.56, 173.55], [455.88, 173.63], [419.14, 173.64]]}, {\"category\": \"cyclist\", \"corners_3d\": [[516.56, 264.92], [557.28, 263.64], [575.07, 275.94], [529.03, 277.59], [516.56, 177.36], [557.28, 177.3], [575.07, 177.9], [529.03, 177.99]]}, {\"category\": \"van\", \"corners_3d\": [[307.05, 260.68], [233.57, 262.83], [267.97, 245.96], [327.27, 244.53], [307.05, 191.58], [233.57, 192.04], [267.97, 188.44], [327.27, 188.14]]}]\n```", - "options": null, - "id": 455 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003015", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[420.66, 208.9], [453.56, 208.52], [455.88, 212.68], [419.14, 213.15], [420.66, 173.55], [453.56, 173.55], [455.88, 173.63], [419.14, 173.64]]}, {\"category\": \"car\", \"corners_3d\": [[791.71, 272.02], [901.28, 277.79], [894.44, 328.99], [736.2, 316.56], [791.71, 181.25], [901.28, 181.74], [894.44, 186.07], [736.2, 185.02]]}, {\"category\": \"car\", \"corners_3d\": [[597.12, 253.32], [666.68, 251.27], [744.57, 275.86], [655.15, 279.42], [597.12, 185.4], [666.68, 185.08], [744.57, 188.91], [655.15, 189.47]]}, {\"category\": \"car\", \"corners_3d\": [[255.14, 287.95], [153.88, 291.08], [224.23, 263.47], [300.89, 261.62], [255.14, 206.5], [153.88, 207.41], [224.23, 199.34], [300.89, 198.8]]}, {\"category\": \"van\", \"corners_3d\": [[307.05, 260.68], [233.57, 262.83], [267.97, 245.96], [327.27, 244.53], [307.05, 191.58], [233.57, 192.04], [267.97, 188.44], [327.27, 188.14]]}, {\"category\": \"car\", \"corners_3d\": [[539.46, 234.34], [593.66, 233.19], [624.07, 244.13], [560.51, 245.74], [539.46, 182.0], [593.66, 181.83], [624.07, 183.46], [560.51, 183.7]]}, {\"category\": \"car\", \"corners_3d\": [[514.5, 223.18], [559.71, 222.25], [586.76, 229.98], [534.91, 231.24], [514.5, 182.48], [559.71, 182.31], [586.76, 183.79], [534.91, 184.03]]}, {\"category\": \"car\", \"corners_3d\": [[332.35, 237.47], [278.96, 238.76], [294.53, 229.8], [340.53, 228.84], [332.35, 193.37], [278.96, 193.77], [294.53, 190.93], [340.53, 190.63]]}, {\"category\": \"car\", \"corners_3d\": [[492.08, 219.01], [525.86, 218.42], [540.05, 223.67], [502.54, 224.41], [492.08, 188.16], [525.86, 187.96], [540.05, 189.7], [502.54, 189.95]]}]\n```", - "options": null, - "id": 456 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003015", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[516.56, 264.92], [557.28, 263.64], [575.07, 275.94], [529.03, 277.59], [516.56, 177.36], [557.28, 177.3], [575.07, 177.9], [529.03, 177.99]]}, {\"category\": \"car\", \"corners_3d\": [[791.71, 272.02], [901.28, 277.79], [894.44, 328.99], [736.2, 316.56], [791.71, 181.25], [901.28, 181.74], [894.44, 186.07], [736.2, 185.02]]}, {\"category\": \"car\", \"corners_3d\": [[597.12, 253.32], [666.68, 251.27], [744.57, 275.86], [655.15, 279.42], [597.12, 185.4], [666.68, 185.08], [744.57, 188.91], [655.15, 189.47]]}, {\"category\": \"car\", \"corners_3d\": [[255.14, 287.95], [153.88, 291.08], [224.23, 263.47], [300.89, 261.62], [255.14, 206.5], [153.88, 207.41], [224.23, 199.34], [300.89, 198.8]]}, {\"category\": \"car\", \"corners_3d\": [[539.46, 234.34], [593.66, 233.19], [624.07, 244.13], [560.51, 245.74], [539.46, 182.0], [593.66, 181.83], [624.07, 183.46], [560.51, 183.7]]}, {\"category\": \"car\", \"corners_3d\": [[514.5, 223.18], [559.71, 222.25], [586.76, 229.98], [534.91, 231.24], [514.5, 182.48], [559.71, 182.31], [586.76, 183.79], [534.91, 184.03]]}, {\"category\": \"car\", \"corners_3d\": [[332.35, 237.47], [278.96, 238.76], [294.53, 229.8], [340.53, 228.84], [332.35, 193.37], [278.96, 193.77], [294.53, 190.93], [340.53, 190.63]]}, {\"category\": \"car\", \"corners_3d\": [[492.08, 219.01], [525.86, 218.42], [540.05, 223.67], [502.54, 224.41], [492.08, 188.16], [525.86, 187.96], [540.05, 189.7], [502.54, 189.95]]}]\n```", - "options": null, - "id": 457 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003015", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[420.66, 208.9], [453.56, 208.52], [455.88, 212.68], [419.14, 213.15], [420.66, 173.55], [453.56, 173.55], [455.88, 173.63], [419.14, 173.64]]}, {\"category\": \"cyclist\", \"corners_3d\": [[516.56, 264.92], [557.28, 263.64], [575.07, 275.94], [529.03, 277.59], [516.56, 177.36], [557.28, 177.3], [575.07, 177.9], [529.03, 177.99]]}, {\"category\": \"car\", \"corners_3d\": [[791.71, 272.02], [901.28, 277.79], [894.44, 328.99], [736.2, 316.56], [791.71, 181.25], [901.28, 181.74], [894.44, 186.07], [736.2, 185.02]]}, {\"category\": \"car\", \"corners_3d\": [[597.12, 253.32], [666.68, 251.27], [744.57, 275.86], [655.15, 279.42], [597.12, 185.4], [666.68, 185.08], [744.57, 188.91], [655.15, 189.47]]}, {\"category\": \"car\", \"corners_3d\": [[255.14, 287.95], [153.88, 291.08], [224.23, 263.47], [300.89, 261.62], [255.14, 206.5], [153.88, 207.41], [224.23, 199.34], [300.89, 198.8]]}, {\"category\": \"van\", \"corners_3d\": [[307.05, 260.68], [233.57, 262.83], [267.97, 245.96], [327.27, 244.53], [307.05, 191.58], [233.57, 192.04], [267.97, 188.44], [327.27, 188.14]]}, {\"category\": \"car\", \"corners_3d\": [[539.46, 234.34], [593.66, 233.19], [624.07, 244.13], [560.51, 245.74], [539.46, 182.0], [593.66, 181.83], [624.07, 183.46], [560.51, 183.7]]}, {\"category\": \"car\", \"corners_3d\": [[514.5, 223.18], [559.71, 222.25], [586.76, 229.98], [534.91, 231.24], [514.5, 182.48], [559.71, 182.31], [586.76, 183.79], [534.91, 184.03]]}, {\"category\": \"car\", \"corners_3d\": [[332.35, 237.47], [278.96, 238.76], [294.53, 229.8], [340.53, 228.84], [332.35, 193.37], [278.96, 193.77], [294.53, 190.93], [340.53, 190.63]]}, {\"category\": \"car\", \"corners_3d\": [[492.08, 219.01], [525.86, 218.42], [540.05, 223.67], [502.54, 224.41], [492.08, 188.16], [525.86, 187.96], [540.05, 189.7], [502.54, 189.95]]}]\n```", - "options": null, - "id": 458 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003049", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[415.25, 229.81], [397.14, 234.15], [254.55, 233.41], [282.44, 229.17], [415.25, 188.38], [397.14, 189.56], [254.55, 189.36], [282.44, 188.2]]}]\n```", - "options": null, - "id": 459 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003062", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[597.04, 229.89], [619.21, 229.8], [622.84, 234.38], [598.89, 234.48], [597.04, 204.85], [619.21, 204.81], [622.84, 207.37], [598.89, 207.43]]}]\n```", - "options": null, - "id": 460 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003062", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[514.72, 258.59], [435.35, 258.79], [468.72, 240.78], [531.42, 240.66], [514.72, 192.7], [435.35, 192.74], [468.72, 188.57], [531.42, 188.55]]}]\n```", - "options": null, - "id": 461 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003062", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[514.72, 258.59], [435.35, 258.79], [468.72, 240.78], [531.42, 240.66], [514.72, 192.7], [435.35, 192.74], [468.72, 188.57], [531.42, 188.55]]}, {\"category\": \"van\", \"corners_3d\": [[597.04, 229.89], [619.21, 229.8], [622.84, 234.38], [598.89, 234.48], [597.04, 204.85], [619.21, 204.81], [622.84, 207.37], [598.89, 207.43]]}]\n```", - "options": null, - "id": 462 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003076", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[431.4, 212.67], [398.0, 212.98], [406.28, 209.17], [436.47, 208.92], [431.4, 183.8], [398.0, 183.88], [406.28, 182.84], [436.47, 182.77]]}]\n```", - "options": null, - "id": 463 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003103", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[495.03, 267.0], [406.54, 266.9], [455.84, 244.78], [523.55, 244.85], [495.03, 174.47], [406.54, 174.47], [455.84, 174.09], [523.55, 174.09]]}, {\"category\": \"car\", \"corners_3d\": [[900.31, 457.4], [1173.09, 457.49], [1973.57, 862.21], [1312.62, 861.71], [900.31, 184.79], [1173.09, 184.8], [1973.57, 201.78], [1312.62, 201.76]]}, {\"category\": \"car\", \"corners_3d\": [[750.08, 300.77], [874.38, 302.09], [990.63, 370.11], [800.12, 367.04], [750.08, 195.81], [874.38, 196.05], [990.63, 208.25], [800.12, 207.7]]}, {\"category\": \"car\", \"corners_3d\": [[513.64, 223.09], [561.31, 223.29], [545.25, 232.08], [489.4, 231.8], [513.64, 182.89], [561.31, 182.93], [545.25, 184.69], [489.4, 184.63]]}, {\"category\": \"car\", \"corners_3d\": [[675.1, 206.03], [644.75, 206.03], [641.56, 203.08], [669.21, 203.07], [675.1, 177.94], [644.75, 177.94], [641.56, 177.48], [669.21, 177.48]]}, {\"category\": \"car\", \"corners_3d\": [[-211.97, 305.83], [-75.43, 283.76], [249.76, 283.84], [177.88, 305.95], [-211.97, 165.78], [-75.43, 166.95], [249.76, 166.95], [177.88, 165.77]]}]\n```", - "options": null, - "id": 464 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003103", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[549.85, 306.32], [508.89, 306.25], [520.64, 291.65], [557.12, 291.7], [549.85, 171.34], [508.89, 171.34], [520.64, 171.5], [557.12, 171.5]]}]\n```", - "options": null, - "id": 465 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003103", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[589.22, 193.23], [566.07, 193.16], [574.2, 191.81], [595.83, 191.88], [589.22, 169.66], [566.07, 169.68], [574.2, 169.89], [595.83, 169.88]]}]\n```", - "options": null, - "id": 466 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003103", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[495.03, 267.0], [406.54, 266.9], [455.84, 244.78], [523.55, 244.85], [495.03, 174.47], [406.54, 174.47], [455.84, 174.09], [523.55, 174.09]]}, {\"category\": \"car\", \"corners_3d\": [[900.31, 457.4], [1173.09, 457.49], [1973.57, 862.21], [1312.62, 861.71], [900.31, 184.79], [1173.09, 184.8], [1973.57, 201.78], [1312.62, 201.76]]}, {\"category\": \"cyclist\", \"corners_3d\": [[549.85, 306.32], [508.89, 306.25], [520.64, 291.65], [557.12, 291.7], [549.85, 171.34], [508.89, 171.34], [520.64, 171.5], [557.12, 171.5]]}, {\"category\": \"car\", \"corners_3d\": [[750.08, 300.77], [874.38, 302.09], [990.63, 370.11], [800.12, 367.04], [750.08, 195.81], [874.38, 196.05], [990.63, 208.25], [800.12, 207.7]]}, {\"category\": \"car\", \"corners_3d\": [[513.64, 223.09], [561.31, 223.29], [545.25, 232.08], [489.4, 231.8], [513.64, 182.89], [561.31, 182.93], [545.25, 184.69], [489.4, 184.63]]}, {\"category\": \"car\", \"corners_3d\": [[675.1, 206.03], [644.75, 206.03], [641.56, 203.08], [669.21, 203.07], [675.1, 177.94], [644.75, 177.94], [641.56, 177.48], [669.21, 177.48]]}, {\"category\": \"car\", \"corners_3d\": [[-211.97, 305.83], [-75.43, 283.76], [249.76, 283.84], [177.88, 305.95], [-211.97, 165.78], [-75.43, 166.95], [249.76, 166.95], [177.88, 165.77]]}]\n```", - "options": null, - "id": 467 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003103", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[495.03, 267.0], [406.54, 266.9], [455.84, 244.78], [523.55, 244.85], [495.03, 174.47], [406.54, 174.47], [455.84, 174.09], [523.55, 174.09]]}, {\"category\": \"car\", \"corners_3d\": [[900.31, 457.4], [1173.09, 457.49], [1973.57, 862.21], [1312.62, 861.71], [900.31, 184.79], [1173.09, 184.8], [1973.57, 201.78], [1312.62, 201.76]]}, {\"category\": \"car\", \"corners_3d\": [[750.08, 300.77], [874.38, 302.09], [990.63, 370.11], [800.12, 367.04], [750.08, 195.81], [874.38, 196.05], [990.63, 208.25], [800.12, 207.7]]}, {\"category\": \"car\", \"corners_3d\": [[513.64, 223.09], [561.31, 223.29], [545.25, 232.08], [489.4, 231.8], [513.64, 182.89], [561.31, 182.93], [545.25, 184.69], [489.4, 184.63]]}, {\"category\": \"car\", \"corners_3d\": [[675.1, 206.03], [644.75, 206.03], [641.56, 203.08], [669.21, 203.07], [675.1, 177.94], [644.75, 177.94], [641.56, 177.48], [669.21, 177.48]]}, {\"category\": \"van\", \"corners_3d\": [[589.22, 193.23], [566.07, 193.16], [574.2, 191.81], [595.83, 191.88], [589.22, 169.66], [566.07, 169.68], [574.2, 169.89], [595.83, 169.88]]}, {\"category\": \"car\", \"corners_3d\": [[-211.97, 305.83], [-75.43, 283.76], [249.76, 283.84], [177.88, 305.95], [-211.97, 165.78], [-75.43, 166.95], [249.76, 166.95], [177.88, 165.77]]}]\n```", - "options": null, - "id": 468 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003103", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[549.85, 306.32], [508.89, 306.25], [520.64, 291.65], [557.12, 291.7], [549.85, 171.34], [508.89, 171.34], [520.64, 171.5], [557.12, 171.5]]}, {\"category\": \"van\", \"corners_3d\": [[589.22, 193.23], [566.07, 193.16], [574.2, 191.81], [595.83, 191.88], [589.22, 169.66], [566.07, 169.68], [574.2, 169.89], [595.83, 169.88]]}]\n```", - "options": null, - "id": 469 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003103", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[495.03, 267.0], [406.54, 266.9], [455.84, 244.78], [523.55, 244.85], [495.03, 174.47], [406.54, 174.47], [455.84, 174.09], [523.55, 174.09]]}, {\"category\": \"car\", \"corners_3d\": [[900.31, 457.4], [1173.09, 457.49], [1973.57, 862.21], [1312.62, 861.71], [900.31, 184.79], [1173.09, 184.8], [1973.57, 201.78], [1312.62, 201.76]]}, {\"category\": \"cyclist\", \"corners_3d\": [[549.85, 306.32], [508.89, 306.25], [520.64, 291.65], [557.12, 291.7], [549.85, 171.34], [508.89, 171.34], [520.64, 171.5], [557.12, 171.5]]}, {\"category\": \"car\", \"corners_3d\": [[750.08, 300.77], [874.38, 302.09], [990.63, 370.11], [800.12, 367.04], [750.08, 195.81], [874.38, 196.05], [990.63, 208.25], [800.12, 207.7]]}, {\"category\": \"car\", \"corners_3d\": [[513.64, 223.09], [561.31, 223.29], [545.25, 232.08], [489.4, 231.8], [513.64, 182.89], [561.31, 182.93], [545.25, 184.69], [489.4, 184.63]]}, {\"category\": \"car\", \"corners_3d\": [[675.1, 206.03], [644.75, 206.03], [641.56, 203.08], [669.21, 203.07], [675.1, 177.94], [644.75, 177.94], [641.56, 177.48], [669.21, 177.48]]}, {\"category\": \"van\", \"corners_3d\": [[589.22, 193.23], [566.07, 193.16], [574.2, 191.81], [595.83, 191.88], [589.22, 169.66], [566.07, 169.68], [574.2, 169.89], [595.83, 169.88]]}, {\"category\": \"car\", \"corners_3d\": [[-211.97, 305.83], [-75.43, 283.76], [249.76, 283.84], [177.88, 305.95], [-211.97, 165.78], [-75.43, 166.95], [249.76, 166.95], [177.88, 165.77]]}]\n```", - "options": null, - "id": 470 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003111", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[669.22, 236.2], [727.18, 235.75], [778.49, 253.46], [704.52, 254.21], [669.22, 182.23], [727.18, 182.16], [778.49, 184.78], [704.52, 184.89]]}, {\"category\": \"car\", \"corners_3d\": [[857.76, 319.9], [1009.81, 317.04], [1413.72, 442.34], [1139.79, 452.51], [857.76, 165.08], [1009.81, 165.23], [1413.72, 158.6], [1139.79, 158.07]]}]\n```", - "options": null, - "id": 471 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003117", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[400.89, 227.15], [365.59, 228.05], [363.36, 223.71], [395.96, 222.94], [400.89, 195.6], [365.59, 195.98], [363.36, 194.16], [395.96, 193.84]]}, {\"category\": \"car\", \"corners_3d\": [[266.84, 479.8], [-44.43, 488.89], [268.26, 324.96], [415.95, 322.82], [266.84, 226.26], [-44.43, 227.84], [268.26, 199.32], [415.95, 198.95]]}]\n```", - "options": null, - "id": 472 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003153", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[546.44, 229.33], [498.58, 229.07], [519.34, 221.82], [561.09, 222.02], [546.44, 186.54], [498.58, 186.47], [519.34, 184.72], [561.09, 184.77]]}]\n```", - "options": null, - "id": 473 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003158", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[289.69, 241.45], [272.56, 241.37], [296.09, 237.04], [312.17, 237.11], [289.69, 189.43], [272.56, 189.41], [296.09, 188.36], [312.17, 188.38]]}]\n```", - "options": null, - "id": 474 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003158", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[494.43, 256.5], [426.33, 256.34], [465.27, 239.85], [519.95, 239.94], [494.43, 191.63], [426.33, 191.59], [465.27, 187.89], [519.95, 187.91]]}]\n```", - "options": null, - "id": 475 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003158", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[494.43, 256.5], [426.33, 256.34], [465.27, 239.85], [519.95, 239.94], [494.43, 191.63], [426.33, 191.59], [465.27, 187.89], [519.95, 187.91]]}, {\"category\": \"cyclist\", \"corners_3d\": [[289.69, 241.45], [272.56, 241.37], [296.09, 237.04], [312.17, 237.11], [289.69, 189.43], [272.56, 189.41], [296.09, 188.36], [312.17, 188.38]]}]\n```", - "options": null, - "id": 476 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003202", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[134.25, 287.84], [77.86, 287.66], [105.61, 281.82], [159.17, 281.98], [134.25, 176.41], [77.86, 176.4], [105.61, 176.22], [159.17, 176.23]]}]\n```", - "options": null, - "id": 477 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003233", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[516.66, 191.61], [527.19, 192.13], [332.19, 193.24], [331.18, 192.6], [516.66, 141.31], [527.19, 139.22], [332.19, 134.82], [331.18, 137.35]]}]\n```", - "options": null, - "id": 478 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003233", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[775.55, 290.55], [903.51, 289.98], [1111.0, 359.42], [902.2, 360.96], [775.55, 151.08], [903.51, 151.24], [1111.0, 131.86], [902.2, 131.43]]}]\n```", - "options": null, - "id": 479 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003233", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (person_sitting). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[438.49, 275.55], [484.0, 275.5], [476.03, 282.33], [427.2, 282.39], [438.49, 189.86], [484.0, 189.85], [476.03, 190.46], [427.2, 190.46]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[397.06, 274.66], [441.09, 274.67], [429.33, 281.52], [382.06, 281.52], [397.06, 190.48], [441.09, 190.48], [429.33, 191.14], [382.06, 191.14]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[451.29, 271.71], [414.23, 271.62], [424.87, 266.82], [459.96, 266.9], [451.29, 190.26], [414.23, 190.25], [424.87, 189.78], [459.96, 189.79]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[475.13, 253.99], [444.78, 253.96], [451.59, 250.92], [480.66, 250.94], [475.13, 184.1], [444.78, 184.1], [451.59, 183.99], [480.66, 183.99]]}]\n```", - "options": null, - "id": 480 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003233", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[278.49, 279.19], [322.28, 278.09], [315.18, 283.24], [268.93, 284.47], [278.49, 161.34], [322.28, 161.57], [315.18, 160.51], [268.93, 160.25]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[300.68, 275.2], [265.45, 275.06], [280.85, 271.02], [314.57, 271.14], [300.68, 159.09], [265.45, 159.12], [280.85, 160.09], [314.57, 160.06]]}]\n```", - "options": null, - "id": 481 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003233", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (tram, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[775.55, 290.55], [903.51, 289.98], [1111.0, 359.42], [902.2, 360.96], [775.55, 151.08], [903.51, 151.24], [1111.0, 131.86], [902.2, 131.43]]}, {\"category\": \"tram\", \"corners_3d\": [[516.66, 191.61], [527.19, 192.13], [332.19, 193.24], [331.18, 192.6], [516.66, 141.31], [527.19, 139.22], [332.19, 134.82], [331.18, 137.35]]}]\n```", - "options": null, - "id": 482 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003233", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (tram, person_sitting). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[516.66, 191.61], [527.19, 192.13], [332.19, 193.24], [331.18, 192.6], [516.66, 141.31], [527.19, 139.22], [332.19, 134.82], [331.18, 137.35]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[438.49, 275.55], [484.0, 275.5], [476.03, 282.33], [427.2, 282.39], [438.49, 189.86], [484.0, 189.85], [476.03, 190.46], [427.2, 190.46]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[397.06, 274.66], [441.09, 274.67], [429.33, 281.52], [382.06, 281.52], [397.06, 190.48], [441.09, 190.48], [429.33, 191.14], [382.06, 191.14]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[451.29, 271.71], [414.23, 271.62], [424.87, 266.82], [459.96, 266.9], [451.29, 190.26], [414.23, 190.25], [424.87, 189.78], [459.96, 189.79]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[475.13, 253.99], [444.78, 253.96], [451.59, 250.92], [480.66, 250.94], [475.13, 184.1], [444.78, 184.1], [451.59, 183.99], [480.66, 183.99]]}]\n```", - "options": null, - "id": 483 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003233", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (tram, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[516.66, 191.61], [527.19, 192.13], [332.19, 193.24], [331.18, 192.6], [516.66, 141.31], [527.19, 139.22], [332.19, 134.82], [331.18, 137.35]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[278.49, 279.19], [322.28, 278.09], [315.18, 283.24], [268.93, 284.47], [278.49, 161.34], [322.28, 161.57], [315.18, 160.51], [268.93, 160.25]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[300.68, 275.2], [265.45, 275.06], [280.85, 271.02], [314.57, 271.14], [300.68, 159.09], [265.45, 159.12], [280.85, 160.09], [314.57, 160.06]]}]\n```", - "options": null, - "id": 484 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003233", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, person_sitting). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[775.55, 290.55], [903.51, 289.98], [1111.0, 359.42], [902.2, 360.96], [775.55, 151.08], [903.51, 151.24], [1111.0, 131.86], [902.2, 131.43]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[438.49, 275.55], [484.0, 275.5], [476.03, 282.33], [427.2, 282.39], [438.49, 189.86], [484.0, 189.85], [476.03, 190.46], [427.2, 190.46]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[397.06, 274.66], [441.09, 274.67], [429.33, 281.52], [382.06, 281.52], [397.06, 190.48], [441.09, 190.48], [429.33, 191.14], [382.06, 191.14]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[451.29, 271.71], [414.23, 271.62], [424.87, 266.82], [459.96, 266.9], [451.29, 190.26], [414.23, 190.25], [424.87, 189.78], [459.96, 189.79]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[475.13, 253.99], [444.78, 253.96], [451.59, 250.92], [480.66, 250.94], [475.13, 184.1], [444.78, 184.1], [451.59, 183.99], [480.66, 183.99]]}]\n```", - "options": null, - "id": 485 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003233", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[775.55, 290.55], [903.51, 289.98], [1111.0, 359.42], [902.2, 360.96], [775.55, 151.08], [903.51, 151.24], [1111.0, 131.86], [902.2, 131.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[278.49, 279.19], [322.28, 278.09], [315.18, 283.24], [268.93, 284.47], [278.49, 161.34], [322.28, 161.57], [315.18, 160.51], [268.93, 160.25]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[300.68, 275.2], [265.45, 275.06], [280.85, 271.02], [314.57, 271.14], [300.68, 159.09], [265.45, 159.12], [280.85, 160.09], [314.57, 160.06]]}]\n```", - "options": null, - "id": 486 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003233", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (person_sitting, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"person_sitting\", \"corners_3d\": [[438.49, 275.55], [484.0, 275.5], [476.03, 282.33], [427.2, 282.39], [438.49, 189.86], [484.0, 189.85], [476.03, 190.46], [427.2, 190.46]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[397.06, 274.66], [441.09, 274.67], [429.33, 281.52], [382.06, 281.52], [397.06, 190.48], [441.09, 190.48], [429.33, 191.14], [382.06, 191.14]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[451.29, 271.71], [414.23, 271.62], [424.87, 266.82], [459.96, 266.9], [451.29, 190.26], [414.23, 190.25], [424.87, 189.78], [459.96, 189.79]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[475.13, 253.99], [444.78, 253.96], [451.59, 250.92], [480.66, 250.94], [475.13, 184.1], [444.78, 184.1], [451.59, 183.99], [480.66, 183.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[278.49, 279.19], [322.28, 278.09], [315.18, 283.24], [268.93, 284.47], [278.49, 161.34], [322.28, 161.57], [315.18, 160.51], [268.93, 160.25]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[300.68, 275.2], [265.45, 275.06], [280.85, 271.02], [314.57, 271.14], [300.68, 159.09], [265.45, 159.12], [280.85, 160.09], [314.57, 160.06]]}]\n```", - "options": null, - "id": 487 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003233", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (tram, van, person_sitting). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[775.55, 290.55], [903.51, 289.98], [1111.0, 359.42], [902.2, 360.96], [775.55, 151.08], [903.51, 151.24], [1111.0, 131.86], [902.2, 131.43]]}, {\"category\": \"tram\", \"corners_3d\": [[516.66, 191.61], [527.19, 192.13], [332.19, 193.24], [331.18, 192.6], [516.66, 141.31], [527.19, 139.22], [332.19, 134.82], [331.18, 137.35]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[438.49, 275.55], [484.0, 275.5], [476.03, 282.33], [427.2, 282.39], [438.49, 189.86], [484.0, 189.85], [476.03, 190.46], [427.2, 190.46]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[397.06, 274.66], [441.09, 274.67], [429.33, 281.52], [382.06, 281.52], [397.06, 190.48], [441.09, 190.48], [429.33, 191.14], [382.06, 191.14]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[451.29, 271.71], [414.23, 271.62], [424.87, 266.82], [459.96, 266.9], [451.29, 190.26], [414.23, 190.25], [424.87, 189.78], [459.96, 189.79]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[475.13, 253.99], [444.78, 253.96], [451.59, 250.92], [480.66, 250.94], [475.13, 184.1], [444.78, 184.1], [451.59, 183.99], [480.66, 183.99]]}]\n```", - "options": null, - "id": 488 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003233", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (tram, van, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[775.55, 290.55], [903.51, 289.98], [1111.0, 359.42], [902.2, 360.96], [775.55, 151.08], [903.51, 151.24], [1111.0, 131.86], [902.2, 131.43]]}, {\"category\": \"tram\", \"corners_3d\": [[516.66, 191.61], [527.19, 192.13], [332.19, 193.24], [331.18, 192.6], [516.66, 141.31], [527.19, 139.22], [332.19, 134.82], [331.18, 137.35]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[278.49, 279.19], [322.28, 278.09], [315.18, 283.24], [268.93, 284.47], [278.49, 161.34], [322.28, 161.57], [315.18, 160.51], [268.93, 160.25]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[300.68, 275.2], [265.45, 275.06], [280.85, 271.02], [314.57, 271.14], [300.68, 159.09], [265.45, 159.12], [280.85, 160.09], [314.57, 160.06]]}]\n```", - "options": null, - "id": 489 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003233", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (tram, person_sitting, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[516.66, 191.61], [527.19, 192.13], [332.19, 193.24], [331.18, 192.6], [516.66, 141.31], [527.19, 139.22], [332.19, 134.82], [331.18, 137.35]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[438.49, 275.55], [484.0, 275.5], [476.03, 282.33], [427.2, 282.39], [438.49, 189.86], [484.0, 189.85], [476.03, 190.46], [427.2, 190.46]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[397.06, 274.66], [441.09, 274.67], [429.33, 281.52], [382.06, 281.52], [397.06, 190.48], [441.09, 190.48], [429.33, 191.14], [382.06, 191.14]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[451.29, 271.71], [414.23, 271.62], [424.87, 266.82], [459.96, 266.9], [451.29, 190.26], [414.23, 190.25], [424.87, 189.78], [459.96, 189.79]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[475.13, 253.99], [444.78, 253.96], [451.59, 250.92], [480.66, 250.94], [475.13, 184.1], [444.78, 184.1], [451.59, 183.99], [480.66, 183.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[278.49, 279.19], [322.28, 278.09], [315.18, 283.24], [268.93, 284.47], [278.49, 161.34], [322.28, 161.57], [315.18, 160.51], [268.93, 160.25]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[300.68, 275.2], [265.45, 275.06], [280.85, 271.02], [314.57, 271.14], [300.68, 159.09], [265.45, 159.12], [280.85, 160.09], [314.57, 160.06]]}]\n```", - "options": null, - "id": 490 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003233", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, person_sitting, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[775.55, 290.55], [903.51, 289.98], [1111.0, 359.42], [902.2, 360.96], [775.55, 151.08], [903.51, 151.24], [1111.0, 131.86], [902.2, 131.43]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[438.49, 275.55], [484.0, 275.5], [476.03, 282.33], [427.2, 282.39], [438.49, 189.86], [484.0, 189.85], [476.03, 190.46], [427.2, 190.46]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[397.06, 274.66], [441.09, 274.67], [429.33, 281.52], [382.06, 281.52], [397.06, 190.48], [441.09, 190.48], [429.33, 191.14], [382.06, 191.14]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[451.29, 271.71], [414.23, 271.62], [424.87, 266.82], [459.96, 266.9], [451.29, 190.26], [414.23, 190.25], [424.87, 189.78], [459.96, 189.79]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[475.13, 253.99], [444.78, 253.96], [451.59, 250.92], [480.66, 250.94], [475.13, 184.1], [444.78, 184.1], [451.59, 183.99], [480.66, 183.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[278.49, 279.19], [322.28, 278.09], [315.18, 283.24], [268.93, 284.47], [278.49, 161.34], [322.28, 161.57], [315.18, 160.51], [268.93, 160.25]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[300.68, 275.2], [265.45, 275.06], [280.85, 271.02], [314.57, 271.14], [300.68, 159.09], [265.45, 159.12], [280.85, 160.09], [314.57, 160.06]]}]\n```", - "options": null, - "id": 491 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003233", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (tram, van, person_sitting, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[775.55, 290.55], [903.51, 289.98], [1111.0, 359.42], [902.2, 360.96], [775.55, 151.08], [903.51, 151.24], [1111.0, 131.86], [902.2, 131.43]]}, {\"category\": \"tram\", \"corners_3d\": [[516.66, 191.61], [527.19, 192.13], [332.19, 193.24], [331.18, 192.6], [516.66, 141.31], [527.19, 139.22], [332.19, 134.82], [331.18, 137.35]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[438.49, 275.55], [484.0, 275.5], [476.03, 282.33], [427.2, 282.39], [438.49, 189.86], [484.0, 189.85], [476.03, 190.46], [427.2, 190.46]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[397.06, 274.66], [441.09, 274.67], [429.33, 281.52], [382.06, 281.52], [397.06, 190.48], [441.09, 190.48], [429.33, 191.14], [382.06, 191.14]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[451.29, 271.71], [414.23, 271.62], [424.87, 266.82], [459.96, 266.9], [451.29, 190.26], [414.23, 190.25], [424.87, 189.78], [459.96, 189.79]]}, {\"category\": \"person_sitting\", \"corners_3d\": [[475.13, 253.99], [444.78, 253.96], [451.59, 250.92], [480.66, 250.94], [475.13, 184.1], [444.78, 184.1], [451.59, 183.99], [480.66, 183.99]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[278.49, 279.19], [322.28, 278.09], [315.18, 283.24], [268.93, 284.47], [278.49, 161.34], [322.28, 161.57], [315.18, 160.51], [268.93, 160.25]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[300.68, 275.2], [265.45, 275.06], [280.85, 271.02], [314.57, 271.14], [300.68, 159.09], [265.45, 159.12], [280.85, 160.09], [314.57, 160.06]]}]\n```", - "options": null, - "id": 492 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003257", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[793.05, 361.29], [831.46, 353.62], [870.03, 363.82], [831.08, 372.4], [793.05, 170.54], [831.46, 170.64], [870.03, 170.51], [831.08, 170.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[411.51, 269.12], [435.44, 267.15], [446.31, 269.98], [421.88, 272.07], [411.51, 187.42], [435.44, 187.13], [446.31, 187.55], [421.88, 187.87]]}]\n```", - "options": null, - "id": 493 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003257", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[804.78, 322.15], [954.81, 326.39], [1127.12, 435.26], [867.65, 423.12], [804.78, 186.73], [954.81, 187.13], [1127.12, 197.25], [867.65, 196.12]]}, {\"category\": \"car\", \"corners_3d\": [[764.43, 220.23], [807.63, 220.65], [825.67, 230.4], [773.56, 229.8], [764.43, 178.84], [807.63, 178.89], [825.67, 180.12], [773.56, 180.04]]}, {\"category\": \"car\", \"corners_3d\": [[322.41, 264.42], [360.18, 256.86], [484.32, 260.57], [456.45, 268.84], [322.41, 191.25], [360.18, 189.73], [484.32, 190.47], [456.45, 192.13]]}]\n```", - "options": null, - "id": 494 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003257", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[753.01, 204.99], [784.09, 205.22], [789.79, 208.9], [755.16, 208.61], [753.01, 169.97], [784.09, 169.95], [789.79, 169.62], [755.16, 169.65]]}]\n```", - "options": null, - "id": 495 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003257", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[804.78, 322.15], [954.81, 326.39], [1127.12, 435.26], [867.65, 423.12], [804.78, 186.73], [954.81, 187.13], [1127.12, 197.25], [867.65, 196.12]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[793.05, 361.29], [831.46, 353.62], [870.03, 363.82], [831.08, 372.4], [793.05, 170.54], [831.46, 170.64], [870.03, 170.51], [831.08, 170.41]]}, {\"category\": \"car\", \"corners_3d\": [[764.43, 220.23], [807.63, 220.65], [825.67, 230.4], [773.56, 229.8], [764.43, 178.84], [807.63, 178.89], [825.67, 180.12], [773.56, 180.04]]}, {\"category\": \"car\", \"corners_3d\": [[322.41, 264.42], [360.18, 256.86], [484.32, 260.57], [456.45, 268.84], [322.41, 191.25], [360.18, 189.73], [484.32, 190.47], [456.45, 192.13]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[411.51, 269.12], [435.44, 267.15], [446.31, 269.98], [421.88, 272.07], [411.51, 187.42], [435.44, 187.13], [446.31, 187.55], [421.88, 187.87]]}]\n```", - "options": null, - "id": 496 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003257", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[793.05, 361.29], [831.46, 353.62], [870.03, 363.82], [831.08, 372.4], [793.05, 170.54], [831.46, 170.64], [870.03, 170.51], [831.08, 170.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[411.51, 269.12], [435.44, 267.15], [446.31, 269.98], [421.88, 272.07], [411.51, 187.42], [435.44, 187.13], [446.31, 187.55], [421.88, 187.87]]}, {\"category\": \"van\", \"corners_3d\": [[753.01, 204.99], [784.09, 205.22], [789.79, 208.9], [755.16, 208.61], [753.01, 169.97], [784.09, 169.95], [789.79, 169.62], [755.16, 169.65]]}]\n```", - "options": null, - "id": 497 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003257", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[804.78, 322.15], [954.81, 326.39], [1127.12, 435.26], [867.65, 423.12], [804.78, 186.73], [954.81, 187.13], [1127.12, 197.25], [867.65, 196.12]]}, {\"category\": \"car\", \"corners_3d\": [[764.43, 220.23], [807.63, 220.65], [825.67, 230.4], [773.56, 229.8], [764.43, 178.84], [807.63, 178.89], [825.67, 180.12], [773.56, 180.04]]}, {\"category\": \"car\", \"corners_3d\": [[322.41, 264.42], [360.18, 256.86], [484.32, 260.57], [456.45, 268.84], [322.41, 191.25], [360.18, 189.73], [484.32, 190.47], [456.45, 192.13]]}, {\"category\": \"van\", \"corners_3d\": [[753.01, 204.99], [784.09, 205.22], [789.79, 208.9], [755.16, 208.61], [753.01, 169.97], [784.09, 169.95], [789.79, 169.62], [755.16, 169.65]]}]\n```", - "options": null, - "id": 498 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003257", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[804.78, 322.15], [954.81, 326.39], [1127.12, 435.26], [867.65, 423.12], [804.78, 186.73], [954.81, 187.13], [1127.12, 197.25], [867.65, 196.12]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[793.05, 361.29], [831.46, 353.62], [870.03, 363.82], [831.08, 372.4], [793.05, 170.54], [831.46, 170.64], [870.03, 170.51], [831.08, 170.41]]}, {\"category\": \"car\", \"corners_3d\": [[764.43, 220.23], [807.63, 220.65], [825.67, 230.4], [773.56, 229.8], [764.43, 178.84], [807.63, 178.89], [825.67, 180.12], [773.56, 180.04]]}, {\"category\": \"car\", \"corners_3d\": [[322.41, 264.42], [360.18, 256.86], [484.32, 260.57], [456.45, 268.84], [322.41, 191.25], [360.18, 189.73], [484.32, 190.47], [456.45, 192.13]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[411.51, 269.12], [435.44, 267.15], [446.31, 269.98], [421.88, 272.07], [411.51, 187.42], [435.44, 187.13], [446.31, 187.55], [421.88, 187.87]]}, {\"category\": \"van\", \"corners_3d\": [[753.01, 204.99], [784.09, 205.22], [789.79, 208.9], [755.16, 208.61], [753.01, 169.97], [784.09, 169.95], [789.79, 169.62], [755.16, 169.65]]}]\n```", - "options": null, - "id": 499 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003271", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[399.17, 318.53], [275.96, 318.55], [375.48, 275.01], [461.87, 275.0], [399.17, 181.68], [275.96, 181.68], [375.48, 179.04], [461.87, 179.04]]}, {\"category\": \"car\", \"corners_3d\": [[522.5, 238.31], [466.54, 238.47], [484.15, 228.9], [531.94, 228.79], [522.5, 183.75], [466.54, 183.78], [484.15, 182.19], [531.94, 182.17]]}, {\"category\": \"car\", \"corners_3d\": [[561.4, 210.21], [530.21, 209.79], [554.44, 206.39], [582.98, 206.73], [561.4, 177.09], [530.21, 177.05], [554.44, 176.66], [582.98, 176.7]]}]\n```", - "options": null, - "id": 500 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003366", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[23.55, 256.02], [-34.05, 256.03], [70.32, 242.53], [118.57, 242.53], [23.55, 203.21], [-34.05, 203.21], [70.32, 198.28], [118.57, 198.28]]}, {\"category\": \"car\", \"corners_3d\": [[273.2, 230.7], [231.89, 230.7], [268.09, 225.15], [305.43, 225.14], [273.2, 194.71], [231.89, 194.71], [268.09, 192.61], [305.43, 192.61]]}]\n```", - "options": null, - "id": 501 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003422", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-99.76, 278.1], [-186.17, 278.11], [-41.68, 258.99], [29.01, 258.98], [-99.76, 202.85], [-186.17, 202.85], [-41.68, 197.4], [29.01, 197.4]]}, {\"category\": \"car\", \"corners_3d\": [[348.16, 212.81], [319.51, 212.76], [343.8, 209.65], [370.23, 209.69], [348.16, 190.38], [319.51, 190.36], [343.8, 189.0], [370.23, 189.01]]}, {\"category\": \"car\", \"corners_3d\": [[484.04, 196.89], [466.22, 196.9], [473.15, 195.67], [490.06, 195.66], [484.04, 181.0], [466.22, 181.0], [473.15, 180.58], [490.06, 180.58]]}]\n```", - "options": null, - "id": 502 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003422", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[700.39, 197.16], [708.29, 197.16], [711.6, 197.98], [703.44, 197.98], [700.39, 170.06], [708.29, 170.06], [711.6, 169.97], [703.44, 169.97]]}]\n```", - "options": null, - "id": 503 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003422", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[528.71, 195.83], [502.83, 195.82], [526.2, 191.08], [546.75, 191.09], [528.71, 164.97], [502.83, 164.97], [526.2, 166.6], [546.75, 166.6]]}]\n```", - "options": null, - "id": 504 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003422", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-99.76, 278.1], [-186.17, 278.11], [-41.68, 258.99], [29.01, 258.98], [-99.76, 202.85], [-186.17, 202.85], [-41.68, 197.4], [29.01, 197.4]]}, {\"category\": \"car\", \"corners_3d\": [[348.16, 212.81], [319.51, 212.76], [343.8, 209.65], [370.23, 209.69], [348.16, 190.38], [319.51, 190.36], [343.8, 189.0], [370.23, 189.01]]}, {\"category\": \"car\", \"corners_3d\": [[484.04, 196.89], [466.22, 196.9], [473.15, 195.67], [490.06, 195.66], [484.04, 181.0], [466.22, 181.0], [473.15, 180.58], [490.06, 180.58]]}, {\"category\": \"cyclist\", \"corners_3d\": [[700.39, 197.16], [708.29, 197.16], [711.6, 197.98], [703.44, 197.98], [700.39, 170.06], [708.29, 170.06], [711.6, 169.97], [703.44, 169.97]]}]\n```", - "options": null, - "id": 505 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003422", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-99.76, 278.1], [-186.17, 278.11], [-41.68, 258.99], [29.01, 258.98], [-99.76, 202.85], [-186.17, 202.85], [-41.68, 197.4], [29.01, 197.4]]}, {\"category\": \"car\", \"corners_3d\": [[348.16, 212.81], [319.51, 212.76], [343.8, 209.65], [370.23, 209.69], [348.16, 190.38], [319.51, 190.36], [343.8, 189.0], [370.23, 189.01]]}, {\"category\": \"car\", \"corners_3d\": [[484.04, 196.89], [466.22, 196.9], [473.15, 195.67], [490.06, 195.66], [484.04, 181.0], [466.22, 181.0], [473.15, 180.58], [490.06, 180.58]]}, {\"category\": \"tram\", \"corners_3d\": [[528.71, 195.83], [502.83, 195.82], [526.2, 191.08], [546.75, 191.09], [528.71, 164.97], [502.83, 164.97], [526.2, 166.6], [546.75, 166.6]]}]\n```", - "options": null, - "id": 506 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003422", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"tram\", \"corners_3d\": [[528.71, 195.83], [502.83, 195.82], [526.2, 191.08], [546.75, 191.09], [528.71, 164.97], [502.83, 164.97], [526.2, 166.6], [546.75, 166.6]]}, {\"category\": \"cyclist\", \"corners_3d\": [[700.39, 197.16], [708.29, 197.16], [711.6, 197.98], [703.44, 197.98], [700.39, 170.06], [708.29, 170.06], [711.6, 169.97], [703.44, 169.97]]}]\n```", - "options": null, - "id": 507 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003422", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist, tram). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-99.76, 278.1], [-186.17, 278.11], [-41.68, 258.99], [29.01, 258.98], [-99.76, 202.85], [-186.17, 202.85], [-41.68, 197.4], [29.01, 197.4]]}, {\"category\": \"car\", \"corners_3d\": [[348.16, 212.81], [319.51, 212.76], [343.8, 209.65], [370.23, 209.69], [348.16, 190.38], [319.51, 190.36], [343.8, 189.0], [370.23, 189.01]]}, {\"category\": \"car\", \"corners_3d\": [[484.04, 196.89], [466.22, 196.9], [473.15, 195.67], [490.06, 195.66], [484.04, 181.0], [466.22, 181.0], [473.15, 180.58], [490.06, 180.58]]}, {\"category\": \"tram\", \"corners_3d\": [[528.71, 195.83], [502.83, 195.82], [526.2, 191.08], [546.75, 191.09], [528.71, 164.97], [502.83, 164.97], [526.2, 166.6], [546.75, 166.6]]}, {\"category\": \"cyclist\", \"corners_3d\": [[700.39, 197.16], [708.29, 197.16], [711.6, 197.98], [703.44, 197.98], [700.39, 170.06], [708.29, 170.06], [711.6, 169.97], [703.44, 169.97]]}]\n```", - "options": null, - "id": 508 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003432", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[253.94, 372.58], [67.17, 371.57], [279.32, 295.78], [395.3, 296.16], [253.94, 176.22], [67.17, 176.2], [279.32, 174.92], [395.3, 174.93]]}, {\"category\": \"van\", \"corners_3d\": [[542.84, 213.04], [502.46, 213.02], [516.14, 208.19], [551.67, 208.21], [542.84, 168.95], [502.46, 168.95], [516.14, 169.42], [551.67, 169.42]]}]\n```", - "options": null, - "id": 509 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003432", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[524.71, 226.17], [479.34, 226.17], [497.2, 218.83], [536.31, 218.83], [524.71, 183.57], [479.34, 183.57], [497.2, 182.09], [536.31, 182.09]]}, {\"category\": \"car\", \"corners_3d\": [[576.99, 195.36], [550.34, 195.37], [553.96, 193.53], [578.43, 193.51], [576.99, 173.15], [550.34, 173.15], [553.96, 173.13], [578.43, 173.13]]}]\n```", - "options": null, - "id": 510 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003432", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[253.94, 372.58], [67.17, 371.57], [279.32, 295.78], [395.3, 296.16], [253.94, 176.22], [67.17, 176.2], [279.32, 174.92], [395.3, 174.93]]}, {\"category\": \"car\", \"corners_3d\": [[524.71, 226.17], [479.34, 226.17], [497.2, 218.83], [536.31, 218.83], [524.71, 183.57], [479.34, 183.57], [497.2, 182.09], [536.31, 182.09]]}, {\"category\": \"van\", \"corners_3d\": [[542.84, 213.04], [502.46, 213.02], [516.14, 208.19], [551.67, 208.21], [542.84, 168.95], [502.46, 168.95], [516.14, 169.42], [551.67, 169.42]]}, {\"category\": \"car\", \"corners_3d\": [[576.99, 195.36], [550.34, 195.37], [553.96, 193.53], [578.43, 193.51], [576.99, 173.15], [550.34, 173.15], [553.96, 173.13], [578.43, 173.13]]}]\n```", - "options": null, - "id": 511 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003439", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-242.16, 879.37], [-977.21, 913.88], [103.07, 397.45], [317.56, 394.17], [-242.16, 197.08], [-977.21, 198.26], [103.07, 180.55], [317.56, 180.44]]}, {\"category\": \"car\", \"corners_3d\": [[388.47, 213.76], [359.86, 214.44], [353.62, 211.83], [380.56, 211.23], [388.47, 188.09], [359.86, 188.35], [353.62, 187.38], [380.56, 187.15]]}, {\"category\": \"car\", \"corners_3d\": [[464.53, 264.71], [378.4, 266.01], [411.61, 244.46], [477.68, 243.69], [464.53, 192.49], [378.4, 192.77], [411.61, 188.16], [477.68, 188.0]]}]\n```", - "options": null, - "id": 512 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003446", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[554.47, 191.11], [539.42, 191.14], [540.74, 190.22], [555.04, 190.2], [554.47, 175.25], [539.42, 175.26], [540.74, 175.14], [555.04, 175.13]]}]\n```", - "options": null, - "id": 513 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003480", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[313.34, 432.22], [141.37, 446.17], [219.18, 371.28], [341.37, 364.0], [313.34, 158.3], [141.37, 157.01], [219.18, 163.95], [341.37, 164.62]]}, {\"category\": \"cyclist\", \"corners_3d\": [[558.59, 226.82], [534.19, 226.87], [537.83, 223.12], [560.21, 223.09], [558.59, 164.49], [534.19, 164.48], [537.83, 165.88], [560.21, 165.9]]}]\n```", - "options": null, - "id": 514 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003480", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[820.73, 261.87], [873.26, 261.45], [910.66, 270.72], [852.23, 271.23], [820.73, 146.46], [873.26, 146.64], [910.66, 142.6], [852.23, 142.38]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[573.4, 244.89], [615.37, 246.48], [597.31, 250.54], [553.25, 248.74], [573.4, 159.81], [615.37, 159.26], [597.31, 157.88], [553.25, 158.49]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[399.2, 252.79], [407.98, 249.42], [450.27, 249.34], [443.6, 252.7], [399.2, 160.66], [407.98, 161.65], [450.27, 161.68], [443.6, 160.69]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[539.97, 241.36], [503.81, 241.15], [511.37, 238.07], [545.68, 238.26], [539.97, 158.12], [503.81, 158.2], [511.37, 159.4], [545.68, 159.33]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[770.75, 236.99], [732.48, 237.02], [723.27, 233.37], [759.02, 233.34], [770.75, 144.57], [732.48, 144.55], [723.27, 146.98], [759.02, 147.0]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.43, 207.54], [572.28, 207.5], [572.73, 208.31], [553.29, 208.35], [553.43, 160.82], [572.28, 160.86], [572.73, 160.22], [553.29, 160.18]]}]\n```", - "options": null, - "id": 515 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003480", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[313.34, 432.22], [141.37, 446.17], [219.18, 371.28], [341.37, 364.0], [313.34, 158.3], [141.37, 157.01], [219.18, 163.95], [341.37, 164.62]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[820.73, 261.87], [873.26, 261.45], [910.66, 270.72], [852.23, 271.23], [820.73, 146.46], [873.26, 146.64], [910.66, 142.6], [852.23, 142.38]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[573.4, 244.89], [615.37, 246.48], [597.31, 250.54], [553.25, 248.74], [573.4, 159.81], [615.37, 159.26], [597.31, 157.88], [553.25, 158.49]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[399.2, 252.79], [407.98, 249.42], [450.27, 249.34], [443.6, 252.7], [399.2, 160.66], [407.98, 161.65], [450.27, 161.68], [443.6, 160.69]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[539.97, 241.36], [503.81, 241.15], [511.37, 238.07], [545.68, 238.26], [539.97, 158.12], [503.81, 158.2], [511.37, 159.4], [545.68, 159.33]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[770.75, 236.99], [732.48, 237.02], [723.27, 233.37], [759.02, 233.34], [770.75, 144.57], [732.48, 144.55], [723.27, 146.98], [759.02, 147.0]]}, {\"category\": \"cyclist\", \"corners_3d\": [[558.59, 226.82], [534.19, 226.87], [537.83, 223.12], [560.21, 223.09], [558.59, 164.49], [534.19, 164.48], [537.83, 165.88], [560.21, 165.9]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.43, 207.54], [572.28, 207.5], [572.73, 208.31], [553.29, 208.35], [553.43, 160.82], [572.28, 160.86], [572.73, 160.22], [553.29, 160.18]]}]\n```", - "options": null, - "id": 516 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003481", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[652.45, 314.83], [644.37, 307.56], [728.16, 305.39], [740.8, 312.42], [652.45, 154.9], [644.37, 156.29], [728.16, 156.7], [740.8, 155.36]]}]\n```", - "options": null, - "id": 517 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003549", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[582.53, 211.43], [618.88, 211.41], [620.92, 216.35], [579.92, 216.37], [582.53, 177.08], [618.88, 177.08], [620.92, 177.62], [579.92, 177.62]]}, {\"category\": \"car\", \"corners_3d\": [[598.82, 204.01], [628.41, 204.01], [630.55, 207.66], [597.49, 207.66], [598.82, 173.02], [628.41, 173.02], [630.55, 173.04], [597.49, 173.04]]}, {\"category\": \"car\", \"corners_3d\": [[917.62, 379.63], [1123.71, 383.14], [1925.81, 741.84], [1348.42, 716.89], [917.62, 183.02], [1123.71, 183.19], [1925.81, 200.83], [1348.42, 199.61]]}, {\"category\": \"car\", \"corners_3d\": [[687.01, 234.02], [746.72, 234.02], [783.24, 250.38], [707.56, 250.37], [687.01, 169.89], [746.72, 169.89], [783.24, 169.1], [707.56, 169.1]]}, {\"category\": \"car\", \"corners_3d\": [[-104.77, 361.63], [-275.18, 361.66], [43.48, 293.61], [152.44, 293.6], [-104.77, 205.83], [-275.18, 205.84], [43.48, 193.95], [152.44, 193.95]]}, {\"category\": \"car\", \"corners_3d\": [[670.21, 217.29], [708.96, 217.25], [724.42, 223.31], [680.39, 223.37], [670.21, 176.57], [708.96, 176.57], [724.42, 177.07], [680.39, 177.08]]}, {\"category\": \"car\", \"corners_3d\": [[645.9, 200.46], [671.53, 200.46], [676.77, 202.81], [648.95, 202.81], [645.9, 174.16], [671.53, 174.16], [676.77, 174.27], [648.95, 174.27]]}]\n```", - "options": null, - "id": 518 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003593", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[432.2, 205.81], [439.2, 204.5], [503.92, 204.5], [499.6, 205.81], [432.2, 178.68], [439.2, 178.44], [503.92, 178.44], [499.6, 178.67]]}]\n```", - "options": null, - "id": 519 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003596", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[181.43, 265.74], [239.36, 269.36], [47.58, 298.17], [-17.57, 292.13], [181.43, 198.91], [239.36, 199.92], [47.58, 208.0], [-17.57, 206.31]]}]\n```", - "options": null, - "id": 520 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003604", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[505.6, 249.45], [430.89, 249.06], [471.12, 234.75], [531.91, 235.01], [505.6, 182.64], [430.89, 182.59], [471.12, 180.76], [531.91, 180.79]]}, {\"category\": \"car\", \"corners_3d\": [[542.2, 193.39], [564.5, 193.37], [563.31, 194.81], [539.43, 194.84], [542.2, 170.84], [564.5, 170.84], [563.31, 170.7], [539.43, 170.7]]}]\n```", - "options": null, - "id": 521 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003617", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[153.39, 297.74], [178.05, 291.89], [260.17, 292.72], [239.66, 298.66], [153.39, 156.39], [178.05, 157.59], [260.17, 157.42], [239.66, 156.2]]}]\n```", - "options": null, - "id": 522 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003651", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[647.64, 227.47], [704.41, 227.85], [710.6, 239.53], [641.83, 238.96], [647.64, 174.21], [704.41, 174.22], [710.6, 174.51], [641.83, 174.49]]}, {\"category\": \"car\", \"corners_3d\": [[1080.62, 291.38], [1265.17, 310.97], [1189.19, 380.63], [945.22, 339.26], [1080.62, 172.82], [1265.17, 172.82], [1189.19, 172.8], [945.22, 172.81]]}, {\"category\": \"car\", \"corners_3d\": [[532.67, 219.78], [491.11, 219.31], [520.17, 213.73], [556.94, 214.09], [532.67, 171.47], [491.11, 171.49], [520.17, 171.65], [556.94, 171.64]]}, {\"category\": \"car\", \"corners_3d\": [[191.98, 263.63], [123.01, 261.99], [254.29, 242.8], [310.1, 243.81], [191.98, 187.48], [123.01, 187.22], [254.29, 184.12], [310.1, 184.29]]}, {\"category\": \"car\", \"corners_3d\": [[443.87, 221.74], [404.19, 221.01], [457.39, 214.43], [492.25, 214.97], [443.87, 184.21], [404.19, 184.04], [457.39, 182.51], [492.25, 182.64]]}, {\"category\": \"car\", \"corners_3d\": [[752.76, 193.28], [774.51, 193.82], [737.66, 195.09], [715.65, 194.48], [752.76, 169.08], [774.51, 168.98], [737.66, 168.74], [715.65, 168.85]]}, {\"category\": \"car\", \"corners_3d\": [[-62.85, 272.53], [-128.71, 270.94], [76.83, 247.13], [129.02, 248.04], [-62.85, 196.45], [-128.71, 196.07], [76.83, 190.43], [129.02, 190.65]]}]\n```", - "options": null, - "id": 523 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003696", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-651.77, 444.28], [-348.47, 378.19], [68.84, 375.7], [-97.46, 439.94], [-651.77, 231.62], [-348.47, 217.31], [68.84, 216.77], [-97.46, 230.68]]}]\n```", - "options": null, - "id": 524 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003696", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[482.22, 211.58], [490.37, 210.71], [512.89, 211.11], [505.19, 212.0], [482.22, 163.81], [490.37, 164.01], [512.89, 163.92], [505.19, 163.71]]}]\n```", - "options": null, - "id": 525 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003696", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[709.2, 221.38], [729.97, 221.98], [708.81, 226.91], [686.26, 226.18], [709.2, 163.74], [729.97, 163.63], [708.81, 162.7], [686.26, 162.84]]}]\n```", - "options": null, - "id": 526 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003696", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-651.77, 444.28], [-348.47, 378.19], [68.84, 375.7], [-97.46, 439.94], [-651.77, 231.62], [-348.47, 217.31], [68.84, 216.77], [-97.46, 230.68]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[482.22, 211.58], [490.37, 210.71], [512.89, 211.11], [505.19, 212.0], [482.22, 163.81], [490.37, 164.01], [512.89, 163.92], [505.19, 163.71]]}]\n```", - "options": null, - "id": 527 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003696", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[709.2, 221.38], [729.97, 221.98], [708.81, 226.91], [686.26, 226.18], [709.2, 163.74], [729.97, 163.63], [708.81, 162.7], [686.26, 162.84]]}, {\"category\": \"car\", \"corners_3d\": [[-651.77, 444.28], [-348.47, 378.19], [68.84, 375.7], [-97.46, 439.94], [-651.77, 231.62], [-348.47, 217.31], [68.84, 216.77], [-97.46, 230.68]]}]\n```", - "options": null, - "id": 528 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003696", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[709.2, 221.38], [729.97, 221.98], [708.81, 226.91], [686.26, 226.18], [709.2, 163.74], [729.97, 163.63], [708.81, 162.7], [686.26, 162.84]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[482.22, 211.58], [490.37, 210.71], [512.89, 211.11], [505.19, 212.0], [482.22, 163.81], [490.37, 164.01], [512.89, 163.92], [505.19, 163.71]]}]\n```", - "options": null, - "id": 529 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003696", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, pedestrian, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[709.2, 221.38], [729.97, 221.98], [708.81, 226.91], [686.26, 226.18], [709.2, 163.74], [729.97, 163.63], [708.81, 162.7], [686.26, 162.84]]}, {\"category\": \"car\", \"corners_3d\": [[-651.77, 444.28], [-348.47, 378.19], [68.84, 375.7], [-97.46, 439.94], [-651.77, 231.62], [-348.47, 217.31], [68.84, 216.77], [-97.46, 230.68]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[482.22, 211.58], [490.37, 210.71], [512.89, 211.11], [505.19, 212.0], [482.22, 163.81], [490.37, 164.01], [512.89, 163.92], [505.19, 163.71]]}]\n```", - "options": null, - "id": 530 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003747", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[953.27, 234.55], [956.78, 230.38], [1124.61, 236.19], [1134.2, 241.27], [953.27, 179.93], [956.78, 179.46], [1124.61, 180.12], [1134.2, 180.71]]}, {\"category\": \"car\", \"corners_3d\": [[765.95, 224.36], [788.51, 226.15], [728.85, 229.78], [706.94, 227.74], [765.95, 198.6], [788.51, 199.5], [728.85, 201.31], [706.94, 200.29]]}]\n```", - "options": null, - "id": 531 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003747", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[627.84, 222.28], [662.79, 222.18], [673.64, 228.75], [634.05, 228.87], [627.84, 182.73], [662.79, 182.72], [673.64, 184.03], [634.05, 184.05]]}, {\"category\": \"van\", \"corners_3d\": [[551.81, 202.36], [571.66, 202.26], [575.4, 204.23], [554.23, 204.34], [551.81, 177.82], [571.66, 177.8], [575.4, 178.13], [554.23, 178.15]]}]\n```", - "options": null, - "id": 532 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003747", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[953.27, 234.55], [956.78, 230.38], [1124.61, 236.19], [1134.2, 241.27], [953.27, 179.93], [956.78, 179.46], [1124.61, 180.12], [1134.2, 180.71]]}, {\"category\": \"van\", \"corners_3d\": [[627.84, 222.28], [662.79, 222.18], [673.64, 228.75], [634.05, 228.87], [627.84, 182.73], [662.79, 182.72], [673.64, 184.03], [634.05, 184.05]]}, {\"category\": \"car\", \"corners_3d\": [[765.95, 224.36], [788.51, 226.15], [728.85, 229.78], [706.94, 227.74], [765.95, 198.6], [788.51, 199.5], [728.85, 201.31], [706.94, 200.29]]}, {\"category\": \"van\", \"corners_3d\": [[551.81, 202.36], [571.66, 202.26], [575.4, 204.23], [554.23, 204.34], [551.81, 177.82], [571.66, 177.8], [575.4, 178.13], [554.23, 178.15]]}]\n```", - "options": null, - "id": 533 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003748", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[351.69, 204.45], [321.62, 204.49], [344.63, 201.75], [372.07, 201.72], [351.69, 181.44], [321.62, 181.45], [344.63, 180.7], [372.07, 180.69]]}]\n```", - "options": null, - "id": 534 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003767", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[10.25, 234.08], [-37.01, 234.04], [55.11, 225.41], [95.76, 225.44], [10.25, 176.16], [-37.01, 176.16], [55.11, 175.69], [95.76, 175.69]]}, {\"category\": \"car\", \"corners_3d\": [[573.06, 197.68], [553.58, 197.69], [556.84, 196.01], [575.0, 196.0], [573.06, 178.33], [553.58, 178.33], [556.84, 177.96], [575.0, 177.96]]}, {\"category\": \"car\", \"corners_3d\": [[385.98, 197.4], [404.18, 197.4], [394.4, 198.56], [375.34, 198.56], [385.98, 180.89], [404.18, 180.89], [394.4, 181.27], [375.34, 181.27]]}, {\"category\": \"car\", \"corners_3d\": [[587.82, 209.36], [620.53, 209.36], [622.01, 214.58], [584.63, 214.58], [587.82, 176.86], [620.53, 176.86], [622.01, 177.43], [584.63, 177.43]]}]\n```", - "options": null, - "id": 535 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003767", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[543.55, 220.88], [483.65, 220.92], [511.02, 209.79], [557.03, 209.76], [543.55, 137.32], [483.65, 137.29], [511.02, 145.53], [557.03, 145.55]]}]\n```", - "options": null, - "id": 536 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003767", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[10.25, 234.08], [-37.01, 234.04], [55.11, 225.41], [95.76, 225.44], [10.25, 176.16], [-37.01, 176.16], [55.11, 175.69], [95.76, 175.69]]}, {\"category\": \"truck\", \"corners_3d\": [[543.55, 220.88], [483.65, 220.92], [511.02, 209.79], [557.03, 209.76], [543.55, 137.32], [483.65, 137.29], [511.02, 145.53], [557.03, 145.55]]}, {\"category\": \"car\", \"corners_3d\": [[573.06, 197.68], [553.58, 197.69], [556.84, 196.01], [575.0, 196.0], [573.06, 178.33], [553.58, 178.33], [556.84, 177.96], [575.0, 177.96]]}, {\"category\": \"car\", \"corners_3d\": [[385.98, 197.4], [404.18, 197.4], [394.4, 198.56], [375.34, 198.56], [385.98, 180.89], [404.18, 180.89], [394.4, 181.27], [375.34, 181.27]]}, {\"category\": \"car\", \"corners_3d\": [[587.82, 209.36], [620.53, 209.36], [622.01, 214.58], [584.63, 214.58], [587.82, 176.86], [620.53, 176.86], [622.01, 177.43], [584.63, 177.43]]}]\n```", - "options": null, - "id": 537 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003769", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (bus). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[683.21, 225.55], [765.41, 225.61], [935.07, 286.06], [758.47, 285.77], [683.21, 136.67], [765.41, 136.62], [935.07, 95.11], [758.47, 95.31]]}]\n```", - "options": null, - "id": 538 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003769", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[240.56, 232.29], [193.81, 232.3], [248.13, 224.52], [288.76, 224.52], [240.56, 194.13], [193.81, 194.13], [248.13, 191.35], [288.76, 191.35]]}]\n```", - "options": null, - "id": 539 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003769", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (bus, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"bus\", \"corners_3d\": [[683.21, 225.55], [765.41, 225.61], [935.07, 286.06], [758.47, 285.77], [683.21, 136.67], [765.41, 136.62], [935.07, 95.11], [758.47, 95.31]]}, {\"category\": \"car\", \"corners_3d\": [[240.56, 232.29], [193.81, 232.3], [248.13, 224.52], [288.76, 224.52], [240.56, 194.13], [193.81, 194.13], [248.13, 191.35], [288.76, 191.35]]}]\n```", - "options": null, - "id": 540 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003808", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[693.7, 189.3], [712.75, 189.51], [694.39, 190.7], [674.22, 190.47], [693.7, 163.32], [712.75, 163.2], [694.39, 162.51], [674.22, 162.64]]}]\n```", - "options": null, - "id": 541 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003808", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[175.8, 362.57], [373.06, 361.08], [-14.3, 699.55], [-582.94, 711.38], [175.8, 210.54], [373.06, 210.24], [-14.3, 277.47], [-582.94, 279.82]]}, {\"category\": \"car\", \"corners_3d\": [[665.06, 230.28], [720.0, 230.11], [757.43, 245.55], [687.76, 245.83], [665.06, 177.45], [720.0, 177.44], [757.43, 178.67], [687.76, 178.7]]}, {\"category\": \"car\", \"corners_3d\": [[422.29, 254.48], [495.31, 254.16], [463.56, 284.19], [363.25, 284.79], [422.29, 187.08], [495.31, 187.02], [463.56, 192.25], [363.25, 192.36]]}, {\"category\": \"car\", \"corners_3d\": [[771.33, 327.3], [917.38, 326.38], [1244.32, 478.98], [955.26, 482.66], [771.33, 179.33], [917.38, 179.3], [1244.32, 185.7], [955.26, 185.85]]}, {\"category\": \"car\", \"corners_3d\": [[635.44, 204.85], [667.55, 204.95], [668.26, 208.32], [632.79, 208.2], [635.44, 177.92], [667.55, 177.94], [668.26, 178.47], [632.79, 178.45]]}]\n```", - "options": null, - "id": 542 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003808", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[175.8, 362.57], [373.06, 361.08], [-14.3, 699.55], [-582.94, 711.38], [175.8, 210.54], [373.06, 210.24], [-14.3, 277.47], [-582.94, 279.82]]}, {\"category\": \"car\", \"corners_3d\": [[665.06, 230.28], [720.0, 230.11], [757.43, 245.55], [687.76, 245.83], [665.06, 177.45], [720.0, 177.44], [757.43, 178.67], [687.76, 178.7]]}, {\"category\": \"car\", \"corners_3d\": [[422.29, 254.48], [495.31, 254.16], [463.56, 284.19], [363.25, 284.79], [422.29, 187.08], [495.31, 187.02], [463.56, 192.25], [363.25, 192.36]]}, {\"category\": \"car\", \"corners_3d\": [[771.33, 327.3], [917.38, 326.38], [1244.32, 478.98], [955.26, 482.66], [771.33, 179.33], [917.38, 179.3], [1244.32, 185.7], [955.26, 185.85]]}, {\"category\": \"car\", \"corners_3d\": [[635.44, 204.85], [667.55, 204.95], [668.26, 208.32], [632.79, 208.2], [635.44, 177.92], [667.55, 177.94], [668.26, 178.47], [632.79, 178.45]]}, {\"category\": \"van\", \"corners_3d\": [[693.7, 189.3], [712.75, 189.51], [694.39, 190.7], [674.22, 190.47], [693.7, 163.32], [712.75, 163.2], [694.39, 162.51], [674.22, 162.64]]}]\n```", - "options": null, - "id": 543 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003904", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[151.15, 418.4], [365.51, 415.6], [-77.49, 914.06], [-763.78, 940.72], [151.15, 201.03], [365.51, 200.71], [-77.49, 257.92], [-763.78, 260.97]]}, {\"category\": \"car\", \"corners_3d\": [[421.39, 365.61], [243.89, 366.12], [363.13, 301.73], [481.38, 301.5], [421.39, 209.64], [243.89, 209.74], [363.13, 197.45], [481.38, 197.41]]}, {\"category\": \"car\", \"corners_3d\": [[753.99, 306.71], [868.54, 305.44], [1023.26, 373.41], [851.38, 376.35], [753.99, 189.47], [868.54, 189.31], [1023.26, 197.75], [851.38, 198.11]]}, {\"category\": \"car\", \"corners_3d\": [[688.99, 259.4], [766.2, 259.31], [821.73, 288.72], [718.29, 288.87], [688.99, 187.83], [766.2, 187.81], [821.73, 192.9], [718.29, 192.93]]}, {\"category\": \"car\", \"corners_3d\": [[548.49, 238.97], [497.6, 238.78], [517.59, 229.39], [561.26, 229.52], [548.49, 187.39], [497.6, 187.35], [517.59, 185.28], [561.26, 185.31]]}, {\"category\": \"car\", \"corners_3d\": [[578.08, 225.63], [533.93, 225.56], [545.98, 218.46], [584.2, 218.51], [578.08, 174.71], [533.93, 174.71], [545.98, 174.46], [584.2, 174.46]]}, {\"category\": \"car\", \"corners_3d\": [[600.65, 216.83], [563.97, 216.65], [575.71, 211.65], [608.23, 211.79], [600.65, 176.79], [563.97, 176.77], [575.71, 176.32], [608.23, 176.33]]}, {\"category\": \"car\", \"corners_3d\": [[639.31, 205.91], [615.92, 205.61], [630.64, 203.2], [652.43, 203.46], [639.31, 183.42], [615.92, 183.32], [630.64, 182.56], [652.43, 182.64]]}]\n```", - "options": null, - "id": 544 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003928", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[543.06, 235.92], [490.7, 235.92], [508.72, 226.32], [553.1, 226.31], [543.06, 191.13], [490.7, 191.14], [508.72, 188.35], [553.1, 188.35]]}, {\"category\": \"car\", \"corners_3d\": [[588.94, 190.73], [569.14, 190.74], [570.72, 189.63], [589.29, 189.62], [588.94, 169.89], [569.14, 169.89], [570.72, 170.07], [589.29, 170.07]]}]\n```", - "options": null, - "id": 545 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003974", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[847.14, 321.53], [825.73, 309.84], [895.08, 309.57], [922.7, 321.2], [847.14, 145.04], [825.73, 147.98], [895.08, 148.05], [922.7, 145.12]]}]\n```", - "options": null, - "id": 546 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003978", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-472.53, 416.72], [-244.32, 362.71], [117.82, 356.41], [-2.07, 406.43], [-472.53, 236.33], [-244.32, 222.27], [117.82, 220.63], [-2.07, 233.65]]}, {\"category\": \"car\", \"corners_3d\": [[257.67, 251.57], [283.43, 245.27], [419.88, 244.71], [406.11, 250.91], [257.67, 194.05], [283.43, 192.35], [419.88, 192.2], [406.11, 193.87]]}]\n```", - "options": null, - "id": 547 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003978", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[612.55, 225.22], [630.66, 225.18], [634.55, 230.01], [614.78, 230.05], [612.55, 176.01], [630.66, 176.01], [634.55, 176.3], [614.78, 176.3]]}]\n```", - "options": null, - "id": 548 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003978", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[612.55, 225.22], [630.66, 225.18], [634.55, 230.01], [614.78, 230.05], [612.55, 176.01], [630.66, 176.01], [634.55, 176.3], [614.78, 176.3]]}, {\"category\": \"car\", \"corners_3d\": [[-472.53, 416.72], [-244.32, 362.71], [117.82, 356.41], [-2.07, 406.43], [-472.53, 236.33], [-244.32, 222.27], [117.82, 220.63], [-2.07, 233.65]]}, {\"category\": \"car\", \"corners_3d\": [[257.67, 251.57], [283.43, 245.27], [419.88, 244.71], [406.11, 250.91], [257.67, 194.05], [283.43, 192.35], [419.88, 192.2], [406.11, 193.87]]}]\n```", - "options": null, - "id": 549 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "003997", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[882.92, 288.94], [908.9, 297.51], [819.03, 298.59], [799.47, 289.86], [882.92, 145.47], [908.9, 142.7], [819.03, 142.36], [799.47, 145.18]]}]\n```", - "options": null, - "id": 550 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004034", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[373.56, 320.82], [235.49, 320.84], [354.34, 273.75], [448.47, 273.74], [373.56, 170.33], [235.49, 170.33], [354.34, 171.13], [448.47, 171.13]]}]\n```", - "options": null, - "id": 551 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004034", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-55.19, 590.88], [-424.95, 599.31], [134.35, 361.99], [295.21, 360.31], [-55.19, 262.53], [-424.95, 264.34], [134.35, 213.43], [295.21, 213.07]]}, {\"category\": \"car\", \"corners_3d\": [[540.94, 228.44], [490.87, 228.4], [508.89, 220.42], [551.76, 220.44], [540.94, 186.6], [490.87, 186.59], [508.89, 184.62], [551.76, 184.62]]}]\n```", - "options": null, - "id": 552 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004034", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-55.19, 590.88], [-424.95, 599.31], [134.35, 361.99], [295.21, 360.31], [-55.19, 262.53], [-424.95, 264.34], [134.35, 213.43], [295.21, 213.07]]}, {\"category\": \"van\", \"corners_3d\": [[373.56, 320.82], [235.49, 320.84], [354.34, 273.75], [448.47, 273.74], [373.56, 170.33], [235.49, 170.33], [354.34, 171.13], [448.47, 171.13]]}, {\"category\": \"car\", \"corners_3d\": [[540.94, 228.44], [490.87, 228.4], [508.89, 220.42], [551.76, 220.44], [540.94, 186.6], [490.87, 186.59], [508.89, 184.62], [551.76, 184.62]]}]\n```", - "options": null, - "id": 553 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004043", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[490.51, 240.61], [435.28, 241.38], [445.08, 230.99], [491.92, 230.44], [490.51, 185.07], [435.28, 185.21], [445.08, 183.34], [491.92, 183.24]]}, {\"category\": \"car\", \"corners_3d\": [[488.04, 215.02], [453.58, 215.62], [446.46, 211.13], [477.44, 210.65], [488.04, 185.78], [453.58, 185.96], [446.46, 184.59], [477.44, 184.44]]}]\n```", - "options": null, - "id": 554 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004060", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1082.52, 300.88], [1206.82, 327.88], [815.4, 343.54], [754.66, 311.37], [1082.52, 191.01], [1206.82, 194.84], [815.4, 197.06], [754.66, 192.5]]}]\n```", - "options": null, - "id": 555 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004080", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[771.03, 212.71], [814.19, 213.18], [822.66, 219.09], [773.15, 218.48], [771.03, 173.35], [814.19, 173.36], [822.66, 173.43], [773.15, 173.42]]}]\n```", - "options": null, - "id": 556 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004160", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[740.36, 286.57], [841.95, 285.94], [1006.82, 357.45], [841.9, 359.14], [740.36, 178.89], [841.95, 178.85], [1006.82, 182.65], [841.9, 182.74]]}, {\"category\": \"car\", \"corners_3d\": [[683.68, 248.8], [751.87, 248.45], [808.27, 272.43], [718.67, 273.04], [683.68, 180.3], [751.87, 180.27], [808.27, 182.62], [718.67, 182.68]]}, {\"category\": \"car\", \"corners_3d\": [[402.57, 210.44], [411.84, 208.87], [475.77, 208.94], [469.26, 210.51], [402.57, 178.25], [411.84, 178.02], [475.77, 178.03], [469.26, 178.26]]}, {\"category\": \"car\", \"corners_3d\": [[701.94, 195.46], [676.38, 195.5], [667.21, 193.63], [690.67, 193.59], [701.94, 168.66], [676.38, 168.65], [667.21, 169.0], [690.67, 169.0]]}, {\"category\": \"car\", \"corners_3d\": [[584.57, 197.94], [562.92, 197.97], [564.77, 196.07], [584.78, 196.05], [584.57, 176.91], [562.92, 176.91], [564.77, 176.61], [584.78, 176.6]]}]\n```", - "options": null, - "id": 557 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004183", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[537.81, 248.1], [480.66, 248.16], [500.49, 235.83], [548.28, 235.78], [537.81, 199.3], [480.66, 199.32], [500.49, 194.99], [548.28, 194.97]]}, {\"category\": \"car\", \"corners_3d\": [[588.16, 197.42], [567.8, 197.44], [569.51, 195.87], [588.57, 195.85], [588.16, 176.0], [567.8, 176.01], [569.51, 175.8], [588.57, 175.8]]}]\n```", - "options": null, - "id": 558 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004258", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[467.0, 229.92], [423.69, 230.65], [426.08, 224.86], [465.08, 224.26], [467.0, 190.35], [423.69, 190.58], [426.08, 188.8], [465.08, 188.62]]}]\n```", - "options": null, - "id": 559 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004268", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[404.97, 194.12], [399.42, 193.89], [443.49, 193.38], [449.55, 193.59], [404.97, 139.46], [399.42, 140.15], [443.49, 141.69], [449.55, 141.05]]}]\n```", - "options": null, - "id": 560 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004268", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[971.17, 402.29], [938.9, 420.61], [815.75, 390.88], [852.61, 376.68], [971.17, 137.65], [938.9, 134.11], [815.75, 139.86], [852.61, 142.6]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[928.89, 423.56], [888.45, 454.2], [765.34, 422.38], [813.9, 398.13], [928.89, 146.17], [888.45, 141.84], [765.34, 146.34], [813.9, 149.76]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[548.74, 326.23], [514.78, 335.96], [438.61, 324.28], [474.6, 315.92], [548.74, 141.69], [514.78, 139.1], [438.61, 142.21], [474.6, 144.44]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[470.01, 336.57], [415.18, 352.45], [348.47, 341.48], [405.91, 327.48], [470.01, 129.37], [415.18, 124.17], [348.47, 127.76], [405.91, 132.35]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[850.83, 372.96], [882.25, 359.67], [1011.05, 382.96], [986.1, 400.1], [850.83, 156.52], [882.25, 158.18], [1011.05, 155.27], [986.1, 153.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.58, 212.66], [536.27, 213.0], [527.12, 211.76], [543.88, 211.45], [553.58, 144.98], [536.27, 144.62], [527.12, 145.98], [543.88, 146.32]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[545.95, 197.81], [525.29, 197.92], [522.92, 197.35], [542.93, 197.25], [545.95, 133.92], [525.29, 133.64], [522.92, 135.17], [542.93, 135.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[566.32, 199.89], [546.52, 199.96], [545.52, 199.51], [564.86, 199.43], [566.32, 143.55], [546.52, 143.4], [545.52, 144.27], [564.86, 144.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[600.27, 198.97], [585.32, 199.08], [579.56, 198.53], [594.11, 198.42], [600.27, 145.43], [585.32, 145.21], [579.56, 146.26], [594.11, 146.47]]}]\n```", - "options": null, - "id": 561 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004268", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[971.17, 402.29], [938.9, 420.61], [815.75, 390.88], [852.61, 376.68], [971.17, 137.65], [938.9, 134.11], [815.75, 139.86], [852.61, 142.6]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[928.89, 423.56], [888.45, 454.2], [765.34, 422.38], [813.9, 398.13], [928.89, 146.17], [888.45, 141.84], [765.34, 146.34], [813.9, 149.76]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[548.74, 326.23], [514.78, 335.96], [438.61, 324.28], [474.6, 315.92], [548.74, 141.69], [514.78, 139.1], [438.61, 142.21], [474.6, 144.44]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[470.01, 336.57], [415.18, 352.45], [348.47, 341.48], [405.91, 327.48], [470.01, 129.37], [415.18, 124.17], [348.47, 127.76], [405.91, 132.35]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[850.83, 372.96], [882.25, 359.67], [1011.05, 382.96], [986.1, 400.1], [850.83, 156.52], [882.25, 158.18], [1011.05, 155.27], [986.1, 153.14]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[553.58, 212.66], [536.27, 213.0], [527.12, 211.76], [543.88, 211.45], [553.58, 144.98], [536.27, 144.62], [527.12, 145.98], [543.88, 146.32]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[545.95, 197.81], [525.29, 197.92], [522.92, 197.35], [542.93, 197.25], [545.95, 133.92], [525.29, 133.64], [522.92, 135.17], [542.93, 135.43]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[566.32, 199.89], [546.52, 199.96], [545.52, 199.51], [564.86, 199.43], [566.32, 143.55], [546.52, 143.4], [545.52, 144.27], [564.86, 144.41]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[600.27, 198.97], [585.32, 199.08], [579.56, 198.53], [594.11, 198.42], [600.27, 145.43], [585.32, 145.21], [579.56, 146.26], [594.11, 146.47]]}, {\"category\": \"cyclist\", \"corners_3d\": [[404.97, 194.12], [399.42, 193.89], [443.49, 193.38], [449.55, 193.59], [404.97, 139.46], [399.42, 140.15], [443.49, 141.69], [449.55, 141.05]]}]\n```", - "options": null, - "id": 562 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004272", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[280.96, 244.57], [214.74, 246.03], [250.84, 233.89], [305.66, 232.87], [280.96, 183.09], [214.74, 183.3], [250.84, 181.56], [305.66, 181.42]]}]\n```", - "options": null, - "id": 563 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004272", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[929.16, 283.58], [1021.04, 302.89], [758.09, 319.19], [699.7, 295.19], [929.16, 186.59], [1021.04, 188.98], [758.09, 191.0], [699.7, 188.03]]}, {\"category\": \"car\", \"corners_3d\": [[817.08, 262.05], [887.91, 276.45], [629.22, 290.03], [590.49, 271.93], [817.08, 166.92], [887.91, 165.97], [629.22, 165.06], [590.49, 166.27]]}, {\"category\": \"car\", \"corners_3d\": [[496.13, 218.29], [458.13, 219.07], [446.26, 213.37], [479.82, 212.77], [496.13, 186.2], [458.13, 186.43], [446.26, 184.75], [479.82, 184.58]]}, {\"category\": \"car\", \"corners_3d\": [[763.9, 251.14], [801.52, 259.11], [607.84, 266.61], [586.57, 257.27], [763.9, 186.66], [801.52, 188.06], [607.84, 189.38], [586.57, 187.74]]}]\n```", - "options": null, - "id": 564 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004272", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[929.16, 283.58], [1021.04, 302.89], [758.09, 319.19], [699.7, 295.19], [929.16, 186.59], [1021.04, 188.98], [758.09, 191.0], [699.7, 188.03]]}, {\"category\": \"car\", \"corners_3d\": [[817.08, 262.05], [887.91, 276.45], [629.22, 290.03], [590.49, 271.93], [817.08, 166.92], [887.91, 165.97], [629.22, 165.06], [590.49, 166.27]]}, {\"category\": \"van\", \"corners_3d\": [[280.96, 244.57], [214.74, 246.03], [250.84, 233.89], [305.66, 232.87], [280.96, 183.09], [214.74, 183.3], [250.84, 181.56], [305.66, 181.42]]}, {\"category\": \"car\", \"corners_3d\": [[496.13, 218.29], [458.13, 219.07], [446.26, 213.37], [479.82, 212.77], [496.13, 186.2], [458.13, 186.43], [446.26, 184.75], [479.82, 184.58]]}, {\"category\": \"car\", \"corners_3d\": [[763.9, 251.14], [801.52, 259.11], [607.84, 266.61], [586.57, 257.27], [763.9, 186.66], [801.52, 188.06], [607.84, 189.38], [586.57, 187.74]]}]\n```", - "options": null, - "id": 565 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004296", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[573.91, 220.35], [617.71, 220.3], [620.62, 226.58], [571.02, 226.65], [573.91, 177.86], [617.71, 177.85], [620.62, 178.52], [571.02, 178.52]]}, {\"category\": \"car\", \"corners_3d\": [[51.52, 498.96], [-262.0, 501.86], [175.36, 333.87], [327.59, 333.17], [51.52, 229.91], [-262.0, 230.42], [175.36, 201.03], [327.59, 200.91]]}]\n```", - "options": null, - "id": 566 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004296", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[854.35, 190.42], [886.18, 190.42], [921.2, 192.65], [885.33, 192.65], [854.35, 158.32], [886.18, 158.32], [921.2, 156.48], [885.33, 156.48]]}]\n```", - "options": null, - "id": 567 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004296", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[784.78, 188.49], [803.15, 188.5], [816.08, 189.63], [796.36, 189.62], [784.78, 171.09], [803.15, 171.09], [816.08, 170.96], [796.36, 170.96]]}, {\"category\": \"van\", \"corners_3d\": [[814.88, 189.11], [840.03, 189.13], [859.35, 190.64], [831.85, 190.62], [814.88, 166.39], [840.03, 166.38], [859.35, 165.78], [831.85, 165.79]]}]\n```", - "options": null, - "id": 568 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004296", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[573.91, 220.35], [617.71, 220.3], [620.62, 226.58], [571.02, 226.65], [573.91, 177.86], [617.71, 177.85], [620.62, 178.52], [571.02, 178.52]]}, {\"category\": \"car\", \"corners_3d\": [[51.52, 498.96], [-262.0, 501.86], [175.36, 333.87], [327.59, 333.17], [51.52, 229.91], [-262.0, 230.42], [175.36, 201.03], [327.59, 200.91]]}, {\"category\": \"truck\", \"corners_3d\": [[854.35, 190.42], [886.18, 190.42], [921.2, 192.65], [885.33, 192.65], [854.35, 158.32], [886.18, 158.32], [921.2, 156.48], [885.33, 156.48]]}]\n```", - "options": null, - "id": 569 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004296", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[573.91, 220.35], [617.71, 220.3], [620.62, 226.58], [571.02, 226.65], [573.91, 177.86], [617.71, 177.85], [620.62, 178.52], [571.02, 178.52]]}, {\"category\": \"car\", \"corners_3d\": [[51.52, 498.96], [-262.0, 501.86], [175.36, 333.87], [327.59, 333.17], [51.52, 229.91], [-262.0, 230.42], [175.36, 201.03], [327.59, 200.91]]}, {\"category\": \"van\", \"corners_3d\": [[784.78, 188.49], [803.15, 188.5], [816.08, 189.63], [796.36, 189.62], [784.78, 171.09], [803.15, 171.09], [816.08, 170.96], [796.36, 170.96]]}, {\"category\": \"van\", \"corners_3d\": [[814.88, 189.11], [840.03, 189.13], [859.35, 190.64], [831.85, 190.62], [814.88, 166.39], [840.03, 166.38], [859.35, 165.78], [831.85, 165.79]]}]\n```", - "options": null, - "id": 570 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004296", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[784.78, 188.49], [803.15, 188.5], [816.08, 189.63], [796.36, 189.62], [784.78, 171.09], [803.15, 171.09], [816.08, 170.96], [796.36, 170.96]]}, {\"category\": \"van\", \"corners_3d\": [[814.88, 189.11], [840.03, 189.13], [859.35, 190.64], [831.85, 190.62], [814.88, 166.39], [840.03, 166.38], [859.35, 165.78], [831.85, 165.79]]}, {\"category\": \"truck\", \"corners_3d\": [[854.35, 190.42], [886.18, 190.42], [921.2, 192.65], [885.33, 192.65], [854.35, 158.32], [886.18, 158.32], [921.2, 156.48], [885.33, 156.48]]}]\n```", - "options": null, - "id": 571 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004296", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, truck, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[573.91, 220.35], [617.71, 220.3], [620.62, 226.58], [571.02, 226.65], [573.91, 177.86], [617.71, 177.85], [620.62, 178.52], [571.02, 178.52]]}, {\"category\": \"car\", \"corners_3d\": [[51.52, 498.96], [-262.0, 501.86], [175.36, 333.87], [327.59, 333.17], [51.52, 229.91], [-262.0, 230.42], [175.36, 201.03], [327.59, 200.91]]}, {\"category\": \"van\", \"corners_3d\": [[784.78, 188.49], [803.15, 188.5], [816.08, 189.63], [796.36, 189.62], [784.78, 171.09], [803.15, 171.09], [816.08, 170.96], [796.36, 170.96]]}, {\"category\": \"van\", \"corners_3d\": [[814.88, 189.11], [840.03, 189.13], [859.35, 190.64], [831.85, 190.62], [814.88, 166.39], [840.03, 166.38], [859.35, 165.78], [831.85, 165.79]]}, {\"category\": \"truck\", \"corners_3d\": [[854.35, 190.42], [886.18, 190.42], [921.2, 192.65], [885.33, 192.65], [854.35, 158.32], [886.18, 158.32], [921.2, 156.48], [885.33, 156.48]]}]\n```", - "options": null, - "id": 572 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004376", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[667.56, 199.89], [690.41, 200.0], [689.35, 202.06], [664.77, 201.93], [667.56, 178.42], [690.41, 178.45], [689.35, 178.87], [664.77, 178.85]]}, {\"category\": \"car\", \"corners_3d\": [[-933.0, 1395.59], [-2055.58, 1357.49], [50.04, 426.4], [298.52, 428.1], [-933.0, 288.38], [-2055.58, 284.78], [50.04, 196.81], [298.52, 196.97]]}, {\"category\": \"car\", \"corners_3d\": [[943.0, 405.88], [1157.14, 407.98], [2709.48, 1102.73], [1830.94, 1070.72], [943.0, 119.77], [1157.14, 119.29], [2709.48, -38.98], [1830.94, -31.69]]}, {\"category\": \"car\", \"corners_3d\": [[736.61, 236.18], [791.46, 236.37], [824.56, 250.16], [757.72, 249.87], [736.61, 171.11], [791.46, 171.11], [824.56, 170.73], [757.72, 170.73]]}, {\"category\": \"car\", \"corners_3d\": [[764.37, 253.15], [840.52, 254.23], [884.13, 280.32], [783.24, 278.44], [764.37, 182.29], [840.52, 182.41], [884.13, 185.48], [783.24, 185.26]]}, {\"category\": \"car\", \"corners_3d\": [[718.05, 216.9], [756.67, 217.07], [770.12, 223.27], [726.05, 223.05], [718.05, 176.85], [756.67, 176.87], [770.12, 177.43], [726.05, 177.41]]}, {\"category\": \"car\", \"corners_3d\": [[712.15, 205.47], [738.8, 205.62], [742.64, 208.64], [713.53, 208.47], [712.15, 181.43], [738.8, 181.47], [742.64, 182.27], [713.53, 182.22]]}]\n```", - "options": null, - "id": 573 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004425", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[789.74, 259.04], [846.98, 258.98], [868.96, 265.7], [806.76, 265.77], [789.74, 157.44], [846.98, 157.46], [868.96, 155.37], [806.76, 155.35]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[695.56, 213.85], [727.7, 213.88], [730.57, 214.7], [697.6, 214.67], [695.56, 153.8], [727.7, 153.77], [730.57, 153.06], [697.6, 153.09]]}]\n```", - "options": null, - "id": 574 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004425", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[593.17, 181.82], [565.62, 181.82], [569.76, 181.79], [594.96, 181.79], [593.17, 157.36], [565.62, 157.37], [569.76, 159.43], [594.96, 159.42]]}]\n```", - "options": null, - "id": 575 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004425", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[614.81, 189.28], [656.06, 189.29], [664.03, 190.81], [614.71, 190.8], [614.81, 138.08], [656.06, 138.03], [664.03, 129.53], [614.71, 129.6]]}]\n```", - "options": null, - "id": 576 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004425", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[456.02, 278.91], [397.7, 278.36], [435.01, 264.05], [484.85, 264.45], [456.02, 158.26], [397.7, 158.39], [435.01, 161.81], [484.85, 161.71]]}, {\"category\": \"cyclist\", \"corners_3d\": [[340.84, 321.48], [276.74, 321.24], [347.53, 291.92], [398.26, 292.08], [340.84, 152.95], [276.74, 153.0], [347.53, 158.98], [398.26, 158.95]]}]\n```", - "options": null, - "id": 577 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004425", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[789.74, 259.04], [846.98, 258.98], [868.96, 265.7], [806.76, 265.77], [789.74, 157.44], [846.98, 157.46], [868.96, 155.37], [806.76, 155.35]]}, {\"category\": \"car\", \"corners_3d\": [[593.17, 181.82], [565.62, 181.82], [569.76, 181.79], [594.96, 181.79], [593.17, 157.36], [565.62, 157.37], [569.76, 159.43], [594.96, 159.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[695.56, 213.85], [727.7, 213.88], [730.57, 214.7], [697.6, 214.67], [695.56, 153.8], [727.7, 153.77], [730.57, 153.06], [697.6, 153.09]]}]\n```", - "options": null, - "id": 578 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004425", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[789.74, 259.04], [846.98, 258.98], [868.96, 265.7], [806.76, 265.77], [789.74, 157.44], [846.98, 157.46], [868.96, 155.37], [806.76, 155.35]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[695.56, 213.85], [727.7, 213.88], [730.57, 214.7], [697.6, 214.67], [695.56, 153.8], [727.7, 153.77], [730.57, 153.06], [697.6, 153.09]]}, {\"category\": \"van\", \"corners_3d\": [[614.81, 189.28], [656.06, 189.29], [664.03, 190.81], [614.71, 190.8], [614.81, 138.08], [656.06, 138.03], [664.03, 129.53], [614.71, 129.6]]}]\n```", - "options": null, - "id": 579 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004425", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[789.74, 259.04], [846.98, 258.98], [868.96, 265.7], [806.76, 265.77], [789.74, 157.44], [846.98, 157.46], [868.96, 155.37], [806.76, 155.35]]}, {\"category\": \"cyclist\", \"corners_3d\": [[456.02, 278.91], [397.7, 278.36], [435.01, 264.05], [484.85, 264.45], [456.02, 158.26], [397.7, 158.39], [435.01, 161.81], [484.85, 161.71]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[695.56, 213.85], [727.7, 213.88], [730.57, 214.7], [697.6, 214.67], [695.56, 153.8], [727.7, 153.77], [730.57, 153.06], [697.6, 153.09]]}, {\"category\": \"cyclist\", \"corners_3d\": [[340.84, 321.48], [276.74, 321.24], [347.53, 291.92], [398.26, 292.08], [340.84, 152.95], [276.74, 153.0], [347.53, 158.98], [398.26, 158.95]]}]\n```", - "options": null, - "id": 580 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004425", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[593.17, 181.82], [565.62, 181.82], [569.76, 181.79], [594.96, 181.79], [593.17, 157.36], [565.62, 157.37], [569.76, 159.43], [594.96, 159.42]]}, {\"category\": \"van\", \"corners_3d\": [[614.81, 189.28], [656.06, 189.29], [664.03, 190.81], [614.71, 190.8], [614.81, 138.08], [656.06, 138.03], [664.03, 129.53], [614.71, 129.6]]}]\n```", - "options": null, - "id": 581 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004425", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[456.02, 278.91], [397.7, 278.36], [435.01, 264.05], [484.85, 264.45], [456.02, 158.26], [397.7, 158.39], [435.01, 161.81], [484.85, 161.71]]}, {\"category\": \"car\", \"corners_3d\": [[593.17, 181.82], [565.62, 181.82], [569.76, 181.79], [594.96, 181.79], [593.17, 157.36], [565.62, 157.37], [569.76, 159.43], [594.96, 159.42]]}, {\"category\": \"cyclist\", \"corners_3d\": [[340.84, 321.48], [276.74, 321.24], [347.53, 291.92], [398.26, 292.08], [340.84, 152.95], [276.74, 153.0], [347.53, 158.98], [398.26, 158.95]]}]\n```", - "options": null, - "id": 582 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004425", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[456.02, 278.91], [397.7, 278.36], [435.01, 264.05], [484.85, 264.45], [456.02, 158.26], [397.7, 158.39], [435.01, 161.81], [484.85, 161.71]]}, {\"category\": \"cyclist\", \"corners_3d\": [[340.84, 321.48], [276.74, 321.24], [347.53, 291.92], [398.26, 292.08], [340.84, 152.95], [276.74, 153.0], [347.53, 158.98], [398.26, 158.95]]}, {\"category\": \"van\", \"corners_3d\": [[614.81, 189.28], [656.06, 189.29], [664.03, 190.81], [614.71, 190.8], [614.81, 138.08], [656.06, 138.03], [664.03, 129.53], [614.71, 129.6]]}]\n```", - "options": null, - "id": 583 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004425", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[789.74, 259.04], [846.98, 258.98], [868.96, 265.7], [806.76, 265.77], [789.74, 157.44], [846.98, 157.46], [868.96, 155.37], [806.76, 155.35]]}, {\"category\": \"car\", \"corners_3d\": [[593.17, 181.82], [565.62, 181.82], [569.76, 181.79], [594.96, 181.79], [593.17, 157.36], [565.62, 157.37], [569.76, 159.43], [594.96, 159.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[695.56, 213.85], [727.7, 213.88], [730.57, 214.7], [697.6, 214.67], [695.56, 153.8], [727.7, 153.77], [730.57, 153.06], [697.6, 153.09]]}, {\"category\": \"van\", \"corners_3d\": [[614.81, 189.28], [656.06, 189.29], [664.03, 190.81], [614.71, 190.8], [614.81, 138.08], [656.06, 138.03], [664.03, 129.53], [614.71, 129.6]]}]\n```", - "options": null, - "id": 584 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004425", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[789.74, 259.04], [846.98, 258.98], [868.96, 265.7], [806.76, 265.77], [789.74, 157.44], [846.98, 157.46], [868.96, 155.37], [806.76, 155.35]]}, {\"category\": \"cyclist\", \"corners_3d\": [[456.02, 278.91], [397.7, 278.36], [435.01, 264.05], [484.85, 264.45], [456.02, 158.26], [397.7, 158.39], [435.01, 161.81], [484.85, 161.71]]}, {\"category\": \"car\", \"corners_3d\": [[593.17, 181.82], [565.62, 181.82], [569.76, 181.79], [594.96, 181.79], [593.17, 157.36], [565.62, 157.37], [569.76, 159.43], [594.96, 159.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[695.56, 213.85], [727.7, 213.88], [730.57, 214.7], [697.6, 214.67], [695.56, 153.8], [727.7, 153.77], [730.57, 153.06], [697.6, 153.09]]}, {\"category\": \"cyclist\", \"corners_3d\": [[340.84, 321.48], [276.74, 321.24], [347.53, 291.92], [398.26, 292.08], [340.84, 152.95], [276.74, 153.0], [347.53, 158.98], [398.26, 158.95]]}]\n```", - "options": null, - "id": 585 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004425", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, van, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[789.74, 259.04], [846.98, 258.98], [868.96, 265.7], [806.76, 265.77], [789.74, 157.44], [846.98, 157.46], [868.96, 155.37], [806.76, 155.35]]}, {\"category\": \"cyclist\", \"corners_3d\": [[456.02, 278.91], [397.7, 278.36], [435.01, 264.05], [484.85, 264.45], [456.02, 158.26], [397.7, 158.39], [435.01, 161.81], [484.85, 161.71]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[695.56, 213.85], [727.7, 213.88], [730.57, 214.7], [697.6, 214.67], [695.56, 153.8], [727.7, 153.77], [730.57, 153.06], [697.6, 153.09]]}, {\"category\": \"cyclist\", \"corners_3d\": [[340.84, 321.48], [276.74, 321.24], [347.53, 291.92], [398.26, 292.08], [340.84, 152.95], [276.74, 153.0], [347.53, 158.98], [398.26, 158.95]]}, {\"category\": \"van\", \"corners_3d\": [[614.81, 189.28], [656.06, 189.29], [664.03, 190.81], [614.71, 190.8], [614.81, 138.08], [656.06, 138.03], [664.03, 129.53], [614.71, 129.6]]}]\n```", - "options": null, - "id": 586 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004425", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[456.02, 278.91], [397.7, 278.36], [435.01, 264.05], [484.85, 264.45], [456.02, 158.26], [397.7, 158.39], [435.01, 161.81], [484.85, 161.71]]}, {\"category\": \"car\", \"corners_3d\": [[593.17, 181.82], [565.62, 181.82], [569.76, 181.79], [594.96, 181.79], [593.17, 157.36], [565.62, 157.37], [569.76, 159.43], [594.96, 159.42]]}, {\"category\": \"cyclist\", \"corners_3d\": [[340.84, 321.48], [276.74, 321.24], [347.53, 291.92], [398.26, 292.08], [340.84, 152.95], [276.74, 153.0], [347.53, 158.98], [398.26, 158.95]]}, {\"category\": \"van\", \"corners_3d\": [[614.81, 189.28], [656.06, 189.29], [664.03, 190.81], [614.71, 190.8], [614.81, 138.08], [656.06, 138.03], [664.03, 129.53], [614.71, 129.6]]}]\n```", - "options": null, - "id": 587 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004425", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car, van, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[789.74, 259.04], [846.98, 258.98], [868.96, 265.7], [806.76, 265.77], [789.74, 157.44], [846.98, 157.46], [868.96, 155.37], [806.76, 155.35]]}, {\"category\": \"cyclist\", \"corners_3d\": [[456.02, 278.91], [397.7, 278.36], [435.01, 264.05], [484.85, 264.45], [456.02, 158.26], [397.7, 158.39], [435.01, 161.81], [484.85, 161.71]]}, {\"category\": \"car\", \"corners_3d\": [[593.17, 181.82], [565.62, 181.82], [569.76, 181.79], [594.96, 181.79], [593.17, 157.36], [565.62, 157.37], [569.76, 159.43], [594.96, 159.42]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[695.56, 213.85], [727.7, 213.88], [730.57, 214.7], [697.6, 214.67], [695.56, 153.8], [727.7, 153.77], [730.57, 153.06], [697.6, 153.09]]}, {\"category\": \"cyclist\", \"corners_3d\": [[340.84, 321.48], [276.74, 321.24], [347.53, 291.92], [398.26, 292.08], [340.84, 152.95], [276.74, 153.0], [347.53, 158.98], [398.26, 158.95]]}, {\"category\": \"van\", \"corners_3d\": [[614.81, 189.28], [656.06, 189.29], [664.03, 190.81], [614.71, 190.8], [614.81, 138.08], [656.06, 138.03], [664.03, 129.53], [614.71, 129.6]]}]\n```", - "options": null, - "id": 588 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004444", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[772.7, 311.9], [896.37, 311.2], [1170.89, 434.56], [938.54, 437.08], [772.7, 181.01], [896.37, 180.96], [1170.89, 188.2], [938.54, 188.34]]}, {\"category\": \"car\", \"corners_3d\": [[696.47, 258.88], [773.36, 258.52], [844.93, 290.44], [739.64, 291.12], [696.47, 181.74], [773.36, 181.7], [844.93, 185.0], [739.64, 185.07]]}, {\"category\": \"car\", \"corners_3d\": [[393.93, 213.14], [404.08, 211.37], [471.35, 211.45], [464.28, 213.22], [393.93, 179.17], [404.08, 178.9], [471.35, 178.91], [464.28, 179.19]]}, {\"category\": \"car\", \"corners_3d\": [[707.71, 196.53], [681.13, 196.58], [671.21, 194.54], [695.54, 194.5], [707.71, 168.64], [681.13, 168.63], [671.21, 169.0], [695.54, 169.0]]}, {\"category\": \"car\", \"corners_3d\": [[586.41, 200.33], [563.88, 200.36], [565.72, 198.2], [586.48, 198.18], [586.41, 178.44], [563.88, 178.44], [565.72, 178.0], [586.48, 178.0]]}]\n```", - "options": null, - "id": 589 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004509", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[409.27, 299.62], [389.54, 308.02], [318.03, 306.4], [342.18, 298.21], [409.27, 160.62], [389.54, 159.22], [318.03, 159.49], [342.18, 160.86]]}]\n```", - "options": null, - "id": 590 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004575", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[161.86, 230.55], [201.34, 230.29], [151.68, 238.58], [106.19, 238.94], [161.86, 199.04], [201.34, 198.91], [151.68, 202.68], [106.19, 202.84]]}, {\"category\": \"car\", \"corners_3d\": [[100.85, 232.04], [140.44, 231.85], [78.64, 240.4], [33.06, 240.65], [100.85, 198.09], [140.44, 198.01], [78.64, 201.65], [33.06, 201.76]]}]\n```", - "options": null, - "id": 591 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004619", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1166.07, 384.0], [1126.78, 392.19], [1018.85, 365.24], [1056.96, 358.97], [1166.07, 139.5], [1126.78, 137.85], [1018.85, 143.28], [1056.96, 144.54]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[678.57, 279.95], [631.91, 281.25], [619.85, 273.86], [663.28, 272.75], [678.57, 131.72], [631.91, 131.09], [619.85, 134.71], [663.28, 135.26]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[733.01, 276.45], [689.48, 277.63], [677.85, 272.54], [719.26, 271.48], [733.01, 147.09], [689.48, 146.68], [677.85, 148.45], [719.26, 148.82]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[828.86, 258.98], [802.18, 260.16], [775.99, 255.51], [801.51, 254.46], [828.86, 150.32], [802.18, 149.86], [775.99, 151.65], [801.51, 152.05]]}]\n```", - "options": null, - "id": 592 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004760", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[433.3, 288.06], [319.59, 289.5], [382.09, 258.01], [464.82, 257.24], [433.3, 174.19], [319.59, 174.2], [382.09, 173.84], [464.82, 173.83]]}, {\"category\": \"car\", \"corners_3d\": [[484.7, 225.84], [434.03, 226.72], [433.79, 218.74], [477.06, 218.1], [484.7, 183.27], [434.03, 183.44], [433.79, 181.87], [477.06, 181.75]]}]\n```", - "options": null, - "id": 593 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004823", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[690.49, 197.98], [713.98, 198.14], [711.28, 200.09], [686.0, 199.91], [690.49, 175.99], [713.98, 176.01], [711.28, 176.25], [686.0, 176.23]]}, {\"category\": \"car\", \"corners_3d\": [[837.76, 337.01], [1002.11, 336.66], [1443.78, 517.9], [1098.76, 519.44], [837.76, 180.73], [1002.11, 180.71], [1443.78, 189.41], [1098.76, 189.49]]}, {\"category\": \"car\", \"corners_3d\": [[551.29, 251.6], [478.88, 250.89], [515.23, 236.36], [574.33, 236.83], [551.29, 183.9], [478.88, 183.8], [515.23, 181.76], [574.33, 181.83]]}, {\"category\": \"car\", \"corners_3d\": [[778.8, 274.84], [870.34, 275.74], [968.36, 323.85], [833.31, 321.92], [778.8, 152.55], [870.34, 152.37], [968.36, 142.79], [833.31, 143.18]]}]\n```", - "options": null, - "id": 594 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004873", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[713.17, 256.55], [793.63, 256.46], [878.56, 293.7], [762.34, 293.88], [713.17, 184.58], [793.63, 184.56], [878.56, 189.78], [762.34, 189.8]]}]\n```", - "options": null, - "id": 595 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004886", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[549.23, 232.95], [492.13, 232.57], [522.23, 222.28], [569.6, 222.54], [549.23, 183.2], [492.13, 183.14], [522.23, 181.37], [569.6, 181.41]]}]\n```", - "options": null, - "id": 596 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004890", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[616.98, 211.88], [576.13, 211.9], [579.65, 206.54], [614.89, 206.52], [616.98, 175.67], [576.13, 175.67], [579.65, 175.28], [614.89, 175.28]]}]\n```", - "options": null, - "id": 597 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004903", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[543.98, 359.21], [618.53, 359.23], [621.95, 403.32], [529.02, 403.3], [543.98, 147.05], [618.53, 147.04], [621.95, 138.79], [529.02, 138.79]]}]\n```", - "options": null, - "id": 598 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004986", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[966.65, 299.02], [991.3, 307.08], [895.16, 307.08], [876.63, 299.02], [966.65, 158.27], [991.3, 156.76], [895.16, 156.76], [876.63, 158.27]]}]\n```", - "options": null, - "id": 599 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004996", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[796.6, 285.13], [833.31, 289.71], [764.92, 303.41], [727.03, 297.64], [796.6, 175.12], [833.31, 174.88], [764.92, 174.18], [727.03, 174.48]]}]\n```", - "options": null, - "id": 600 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004996", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[243.8, 259.73], [256.62, 261.58], [215.65, 265.28], [203.24, 263.26], [243.8, 169.43], [256.62, 169.18], [215.65, 168.66], [203.24, 168.94]]}]\n```", - "options": null, - "id": 601 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "004996", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[796.6, 285.13], [833.31, 289.71], [764.92, 303.41], [727.03, 297.64], [796.6, 175.12], [833.31, 174.88], [764.92, 174.18], [727.03, 174.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[243.8, 259.73], [256.62, 261.58], [215.65, 265.28], [203.24, 263.26], [243.8, 169.43], [256.62, 169.18], [215.65, 168.66], [203.24, 168.94]]}]\n```", - "options": null, - "id": 602 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005036", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[320.75, 422.34], [34.04, 425.36], [290.84, 308.16], [443.68, 307.29], [320.75, 140.85], [34.04, 140.46], [290.84, 155.49], [443.68, 155.61]]}, {\"category\": \"van\", \"corners_3d\": [[774.2, 320.73], [917.35, 320.16], [1191.0, 445.49], [927.14, 447.44], [774.2, 163.97], [917.35, 164.0], [1191.0, 156.47], [927.14, 156.35]]}, {\"category\": \"van\", \"corners_3d\": [[547.26, 236.77], [483.21, 236.83], [505.14, 225.0], [557.33, 224.96], [547.26, 167.2], [483.21, 167.19], [505.14, 168.24], [557.33, 168.24]]}]\n```", - "options": null, - "id": 603 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005036", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[521.12, 265.4], [431.01, 265.87], [468.9, 242.23], [536.05, 241.97], [521.12, 185.43], [431.01, 185.5], [468.9, 182.28], [536.05, 182.25]]}, {\"category\": \"car\", \"corners_3d\": [[566.14, 222.2], [513.09, 221.91], [536.61, 214.41], [581.62, 214.61], [566.14, 171.96], [513.09, 171.96], [536.61, 172.1], [581.62, 172.09]]}, {\"category\": \"car\", \"corners_3d\": [[657.11, 200.43], [679.66, 201.09], [652.69, 202.58], [629.66, 201.84], [657.11, 173.37], [679.66, 173.38], [652.69, 173.4], [629.66, 173.39]]}]\n```", - "options": null, - "id": 604 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005036", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[320.75, 422.34], [34.04, 425.36], [290.84, 308.16], [443.68, 307.29], [320.75, 140.85], [34.04, 140.46], [290.84, 155.49], [443.68, 155.61]]}, {\"category\": \"van\", \"corners_3d\": [[774.2, 320.73], [917.35, 320.16], [1191.0, 445.49], [927.14, 447.44], [774.2, 163.97], [917.35, 164.0], [1191.0, 156.47], [927.14, 156.35]]}, {\"category\": \"car\", \"corners_3d\": [[521.12, 265.4], [431.01, 265.87], [468.9, 242.23], [536.05, 241.97], [521.12, 185.43], [431.01, 185.5], [468.9, 182.28], [536.05, 182.25]]}, {\"category\": \"van\", \"corners_3d\": [[547.26, 236.77], [483.21, 236.83], [505.14, 225.0], [557.33, 224.96], [547.26, 167.2], [483.21, 167.19], [505.14, 168.24], [557.33, 168.24]]}, {\"category\": \"car\", \"corners_3d\": [[566.14, 222.2], [513.09, 221.91], [536.61, 214.41], [581.62, 214.61], [566.14, 171.96], [513.09, 171.96], [536.61, 172.1], [581.62, 172.09]]}, {\"category\": \"car\", \"corners_3d\": [[657.11, 200.43], [679.66, 201.09], [652.69, 202.58], [629.66, 201.84], [657.11, 173.37], [679.66, 173.38], [652.69, 173.4], [629.66, 173.39]]}]\n```", - "options": null, - "id": 605 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005048", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[805.65, 224.95], [863.43, 225.91], [878.9, 236.61], [809.38, 235.23], [805.65, 173.17], [863.43, 173.18], [878.9, 173.25], [809.38, 173.24]]}, {\"category\": \"car\", \"corners_3d\": [[344.85, 275.98], [263.78, 271.95], [409.15, 246.6], [473.23, 248.81], [344.85, 191.31], [263.78, 190.59], [409.15, 186.06], [473.23, 186.45]]}, {\"category\": \"car\", \"corners_3d\": [[1340.33, 202.35], [1368.7, 204.05], [1245.3, 202.74], [1224.53, 201.17], [1340.33, 163.18], [1368.7, 162.62], [1245.3, 163.05], [1224.53, 163.56]]}]\n```", - "options": null, - "id": 606 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005144", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[654.55, 215.24], [700.43, 215.25], [720.32, 224.58], [664.33, 224.58], [654.55, 160.1], [700.43, 160.1], [720.32, 157.29], [664.33, 157.3]]}]\n```", - "options": null, - "id": 607 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005144", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[701.73, 245.16], [764.77, 245.17], [806.72, 264.78], [726.58, 264.77], [701.73, 184.15], [764.77, 184.15], [806.72, 187.22], [726.58, 187.22]]}, {\"category\": \"car\", \"corners_3d\": [[677.53, 228.51], [729.89, 228.43], [756.55, 239.48], [693.81, 239.59], [677.53, 179.41], [729.89, 179.4], [756.55, 180.7], [693.81, 180.72]]}, {\"category\": \"car\", \"corners_3d\": [[555.16, 214.06], [520.61, 214.2], [524.73, 209.69], [555.52, 209.58], [555.16, 179.26], [520.61, 179.28], [524.73, 178.58], [555.52, 178.56]]}]\n```", - "options": null, - "id": 608 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005144", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[701.73, 245.16], [764.77, 245.17], [806.72, 264.78], [726.58, 264.77], [701.73, 184.15], [764.77, 184.15], [806.72, 187.22], [726.58, 187.22]]}, {\"category\": \"car\", \"corners_3d\": [[677.53, 228.51], [729.89, 228.43], [756.55, 239.48], [693.81, 239.59], [677.53, 179.41], [729.89, 179.4], [756.55, 180.7], [693.81, 180.72]]}, {\"category\": \"van\", \"corners_3d\": [[654.55, 215.24], [700.43, 215.25], [720.32, 224.58], [664.33, 224.58], [654.55, 160.1], [700.43, 160.1], [720.32, 157.29], [664.33, 157.3]]}, {\"category\": \"car\", \"corners_3d\": [[555.16, 214.06], [520.61, 214.2], [524.73, 209.69], [555.52, 209.58], [555.16, 179.26], [520.61, 179.28], [524.73, 178.58], [555.52, 178.56]]}]\n```", - "options": null, - "id": 609 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005309", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[246.9, 225.23], [240.23, 226.05], [230.47, 225.98], [237.28, 225.17], [246.9, 184.95], [240.23, 185.14], [230.47, 185.13], [237.28, 184.94]]}]\n```", - "options": null, - "id": 610 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005309", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[594.62, 201.3], [623.07, 201.32], [622.93, 204.0], [591.8, 203.98], [594.62, 173.19], [623.07, 173.19], [622.93, 173.22], [591.8, 173.22]]}]\n```", - "options": null, - "id": 611 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005309", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[594.62, 201.3], [623.07, 201.32], [622.93, 204.0], [591.8, 203.98], [594.62, 173.19], [623.07, 173.19], [622.93, 173.22], [591.8, 173.22]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[246.9, 225.23], [240.23, 226.05], [230.47, 225.98], [237.28, 225.17], [246.9, 184.95], [240.23, 185.14], [230.47, 185.13], [237.28, 184.94]]}]\n```", - "options": null, - "id": 612 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005345", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1077.59, 396.36], [868.89, 397.72], [749.31, 299.71], [867.77, 299.27], [1077.59, 208.52], [868.89, 208.74], [749.31, 193.1], [867.77, 193.03]]}, {\"category\": \"car\", \"corners_3d\": [[330.95, 367.3], [145.22, 368.34], [307.48, 297.75], [425.81, 297.33], [330.95, 208.58], [145.22, 208.77], [307.48, 195.8], [425.81, 195.72]]}, {\"category\": \"car\", \"corners_3d\": [[693.38, 250.44], [767.08, 250.36], [818.04, 274.42], [721.49, 274.55], [693.38, 185.93], [767.08, 185.91], [818.04, 189.97], [721.49, 189.99]]}, {\"category\": \"car\", \"corners_3d\": [[755.69, 243.5], [685.99, 243.58], [670.31, 230.41], [727.06, 230.36], [755.69, 181.56], [685.99, 181.57], [670.31, 179.95], [727.06, 179.94]]}, {\"category\": \"car\", \"corners_3d\": [[714.29, 210.52], [700.9, 208.58], [788.39, 207.62], [806.02, 209.46], [714.29, 172.36], [700.9, 172.38], [788.39, 172.4], [806.02, 172.37]]}]\n```", - "options": null, - "id": 613 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005346", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[449.31, 308.49], [324.37, 308.03], [411.0, 268.87], [499.86, 269.1], [449.31, 188.57], [324.37, 188.51], [411.0, 183.98], [499.86, 184.0]]}, {\"category\": \"car\", \"corners_3d\": [[573.97, 208.84], [542.85, 208.84], [549.23, 205.37], [577.35, 205.37], [573.97, 182.22], [542.85, 182.22], [549.23, 181.32], [577.35, 181.32]]}]\n```", - "options": null, - "id": 614 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005397", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[464.95, 216.01], [453.83, 218.47], [346.77, 217.98], [363.49, 215.58], [464.95, 184.45], [453.83, 185.11], [346.77, 184.98], [363.49, 184.33]]}]\n```", - "options": null, - "id": 615 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005444", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.86, 718.86), (cx, cy) = (607.19, 185.22). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[716.84, 206.11], [754.76, 206.28], [759.62, 208.95], [716.9, 208.74], [716.84, 172.3], [754.76, 172.2], [759.62, 170.55], [716.9, 170.68]]}, {\"category\": \"car\", \"corners_3d\": [[701.08, 194.81], [722.16, 194.85], [724.05, 195.5], [701.53, 195.46], [701.08, 176.26], [722.16, 176.23], [724.05, 175.61], [701.53, 175.65]]}, {\"category\": \"car\", \"corners_3d\": [[691.74, 189.48], [707.11, 189.48], [713.28, 189.7], [697.14, 189.7], [691.74, 176.37], [707.11, 176.38], [713.28, 175.93], [697.14, 175.92]]}]\n```", - "options": null, - "id": 616 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005528", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[788.59, 290.39], [809.98, 286.07], [949.45, 303.9], [930.74, 309.73], [788.59, 143.86], [809.98, 144.92], [949.45, 140.52], [930.74, 139.09]]}, {\"category\": \"cyclist\", \"corners_3d\": [[760.21, 264.3], [770.07, 260.88], [871.86, 267.21], [865.51, 271.14], [760.21, 149.96], [770.07, 150.82], [871.86, 149.23], [865.51, 148.25]]}]\n```", - "options": null, - "id": 617 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005528", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[795.99, 284.03], [787.89, 279.57], [817.35, 279.48], [826.67, 283.93], [795.99, 160.47], [787.89, 160.97], [817.35, 160.98], [826.67, 160.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[753.88, 267.19], [736.38, 264.44], [752.54, 262.98], [770.23, 265.65], [753.88, 170.91], [736.38, 170.96], [752.54, 170.99], [770.23, 170.94]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[208.06, 296.55], [191.74, 301.67], [168.03, 301.71], [185.29, 296.59], [208.06, 175.84], [191.74, 175.97], [168.03, 175.97], [185.29, 175.85]]}]\n```", - "options": null, - "id": 618 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005528", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[600.74, 199.75], [625.69, 199.73], [628.13, 201.92], [601.15, 201.94], [600.74, 175.09], [625.69, 175.09], [628.13, 175.27], [601.15, 175.27]]}, {\"category\": \"car\", \"corners_3d\": [[163.59, 271.98], [84.02, 271.87], [191.01, 251.97], [254.66, 252.03], [163.59, 205.88], [84.02, 205.85], [191.01, 199.22], [254.66, 199.24]]}, {\"category\": \"car\", \"corners_3d\": [[436.12, 213.32], [407.77, 213.26], [429.66, 209.42], [455.34, 209.47], [436.12, 187.19], [407.77, 187.17], [429.66, 185.81], [455.34, 185.83]]}]\n```", - "options": null, - "id": 619 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005528", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[788.59, 290.39], [809.98, 286.07], [949.45, 303.9], [930.74, 309.73], [788.59, 143.86], [809.98, 144.92], [949.45, 140.52], [930.74, 139.09]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[795.99, 284.03], [787.89, 279.57], [817.35, 279.48], [826.67, 283.93], [795.99, 160.47], [787.89, 160.97], [817.35, 160.98], [826.67, 160.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[753.88, 267.19], [736.38, 264.44], [752.54, 262.98], [770.23, 265.65], [753.88, 170.91], [736.38, 170.96], [752.54, 170.99], [770.23, 170.94]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[208.06, 296.55], [191.74, 301.67], [168.03, 301.71], [185.29, 296.59], [208.06, 175.84], [191.74, 175.97], [168.03, 175.97], [185.29, 175.85]]}, {\"category\": \"cyclist\", \"corners_3d\": [[760.21, 264.3], [770.07, 260.88], [871.86, 267.21], [865.51, 271.14], [760.21, 149.96], [770.07, 150.82], [871.86, 149.23], [865.51, 148.25]]}]\n```", - "options": null, - "id": 620 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005528", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[600.74, 199.75], [625.69, 199.73], [628.13, 201.92], [601.15, 201.94], [600.74, 175.09], [625.69, 175.09], [628.13, 175.27], [601.15, 175.27]]}, {\"category\": \"cyclist\", \"corners_3d\": [[788.59, 290.39], [809.98, 286.07], [949.45, 303.9], [930.74, 309.73], [788.59, 143.86], [809.98, 144.92], [949.45, 140.52], [930.74, 139.09]]}, {\"category\": \"car\", \"corners_3d\": [[163.59, 271.98], [84.02, 271.87], [191.01, 251.97], [254.66, 252.03], [163.59, 205.88], [84.02, 205.85], [191.01, 199.22], [254.66, 199.24]]}, {\"category\": \"car\", \"corners_3d\": [[436.12, 213.32], [407.77, 213.26], [429.66, 209.42], [455.34, 209.47], [436.12, 187.19], [407.77, 187.17], [429.66, 185.81], [455.34, 185.83]]}, {\"category\": \"cyclist\", \"corners_3d\": [[760.21, 264.3], [770.07, 260.88], [871.86, 267.21], [865.51, 271.14], [760.21, 149.96], [770.07, 150.82], [871.86, 149.23], [865.51, 148.25]]}]\n```", - "options": null, - "id": 621 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005528", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[600.74, 199.75], [625.69, 199.73], [628.13, 201.92], [601.15, 201.94], [600.74, 175.09], [625.69, 175.09], [628.13, 175.27], [601.15, 175.27]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[795.99, 284.03], [787.89, 279.57], [817.35, 279.48], [826.67, 283.93], [795.99, 160.47], [787.89, 160.97], [817.35, 160.98], [826.67, 160.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[753.88, 267.19], [736.38, 264.44], [752.54, 262.98], [770.23, 265.65], [753.88, 170.91], [736.38, 170.96], [752.54, 170.99], [770.23, 170.94]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[208.06, 296.55], [191.74, 301.67], [168.03, 301.71], [185.29, 296.59], [208.06, 175.84], [191.74, 175.97], [168.03, 175.97], [185.29, 175.85]]}, {\"category\": \"car\", \"corners_3d\": [[163.59, 271.98], [84.02, 271.87], [191.01, 251.97], [254.66, 252.03], [163.59, 205.88], [84.02, 205.85], [191.01, 199.22], [254.66, 199.24]]}, {\"category\": \"car\", \"corners_3d\": [[436.12, 213.32], [407.77, 213.26], [429.66, 209.42], [455.34, 209.47], [436.12, 187.19], [407.77, 187.17], [429.66, 185.81], [455.34, 185.83]]}]\n```", - "options": null, - "id": 622 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005528", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, pedestrian, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[600.74, 199.75], [625.69, 199.73], [628.13, 201.92], [601.15, 201.94], [600.74, 175.09], [625.69, 175.09], [628.13, 175.27], [601.15, 175.27]]}, {\"category\": \"cyclist\", \"corners_3d\": [[788.59, 290.39], [809.98, 286.07], [949.45, 303.9], [930.74, 309.73], [788.59, 143.86], [809.98, 144.92], [949.45, 140.52], [930.74, 139.09]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[795.99, 284.03], [787.89, 279.57], [817.35, 279.48], [826.67, 283.93], [795.99, 160.47], [787.89, 160.97], [817.35, 160.98], [826.67, 160.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[753.88, 267.19], [736.38, 264.44], [752.54, 262.98], [770.23, 265.65], [753.88, 170.91], [736.38, 170.96], [752.54, 170.99], [770.23, 170.94]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[208.06, 296.55], [191.74, 301.67], [168.03, 301.71], [185.29, 296.59], [208.06, 175.84], [191.74, 175.97], [168.03, 175.97], [185.29, 175.85]]}, {\"category\": \"car\", \"corners_3d\": [[163.59, 271.98], [84.02, 271.87], [191.01, 251.97], [254.66, 252.03], [163.59, 205.88], [84.02, 205.85], [191.01, 199.22], [254.66, 199.24]]}, {\"category\": \"car\", \"corners_3d\": [[436.12, 213.32], [407.77, 213.26], [429.66, 209.42], [455.34, 209.47], [436.12, 187.19], [407.77, 187.17], [429.66, 185.81], [455.34, 185.83]]}, {\"category\": \"cyclist\", \"corners_3d\": [[760.21, 264.3], [770.07, 260.88], [871.86, 267.21], [865.51, 271.14], [760.21, 149.96], [770.07, 150.82], [871.86, 149.23], [865.51, 148.25]]}]\n```", - "options": null, - "id": 623 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005577", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-746.29, 1065.6], [-1582.06, 1045.0], [21.64, 410.82], [255.99, 412.33], [-746.29, 282.33], [-1582.06, 279.8], [21.64, 202.03], [255.99, 202.22]]}, {\"category\": \"car\", \"corners_3d\": [[485.58, 214.35], [517.65, 214.21], [514.05, 218.49], [478.63, 218.67], [485.58, 182.09], [517.65, 182.06], [514.05, 183.02], [478.63, 183.05]]}, {\"category\": \"car\", \"corners_3d\": [[520.55, 210.37], [488.72, 210.42], [497.87, 206.94], [526.74, 206.9], [520.55, 180.96], [488.72, 180.97], [497.87, 180.21], [526.74, 180.2]]}, {\"category\": \"car\", \"corners_3d\": [[289.18, 223.72], [299.3, 220.61], [421.38, 219.02], [419.3, 221.93], [289.18, 174.51], [299.3, 174.41], [421.38, 174.36], [419.3, 174.45]]}]\n```", - "options": null, - "id": 624 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005672", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[448.31, 286.07], [329.1, 287.19], [398.72, 253.54], [482.6, 252.98], [448.31, 159.1], [329.1, 158.97], [398.72, 163.05], [482.6, 163.12]]}]\n```", - "options": null, - "id": 625 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005672", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[288.63, 425.05], [10.03, 432.66], [283.54, 301.34], [419.36, 299.45], [288.63, 190.47], [10.03, 191.0], [283.54, 181.83], [419.36, 181.7]]}, {\"category\": \"car\", \"corners_3d\": [[495.63, 240.32], [411.9, 240.01], [462.08, 225.1], [527.33, 225.28], [495.63, 161.2], [411.9, 161.26], [462.08, 163.83], [527.33, 163.8]]}, {\"category\": \"car\", \"corners_3d\": [[663.02, 192.62], [691.81, 192.92], [677.43, 194.32], [646.91, 193.97], [663.02, 162.77], [691.81, 162.62], [677.43, 161.9], [646.91, 162.08]]}]\n```", - "options": null, - "id": 626 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005672", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[288.63, 425.05], [10.03, 432.66], [283.54, 301.34], [419.36, 299.45], [288.63, 190.47], [10.03, 191.0], [283.54, 181.83], [419.36, 181.7]]}, {\"category\": \"van\", \"corners_3d\": [[448.31, 286.07], [329.1, 287.19], [398.72, 253.54], [482.6, 252.98], [448.31, 159.1], [329.1, 158.97], [398.72, 163.05], [482.6, 163.12]]}, {\"category\": \"car\", \"corners_3d\": [[495.63, 240.32], [411.9, 240.01], [462.08, 225.1], [527.33, 225.28], [495.63, 161.2], [411.9, 161.26], [462.08, 163.83], [527.33, 163.8]]}, {\"category\": \"car\", \"corners_3d\": [[663.02, 192.62], [691.81, 192.92], [677.43, 194.32], [646.91, 193.97], [663.02, 162.77], [691.81, 162.62], [677.43, 161.9], [646.91, 162.08]]}]\n```", - "options": null, - "id": 627 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005677", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[959.08, 304.07], [987.57, 304.02], [1049.59, 325.17], [1016.52, 325.24], [959.08, 189.53], [987.57, 189.52], [1049.59, 192.21], [1016.52, 192.22]]}]\n```", - "options": null, - "id": 628 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005677", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[202.85, 571.21], [-176.11, 575.51], [224.12, 366.4], [405.22, 365.4], [202.85, 233.0], [-176.11, 233.65], [224.12, 202.08], [405.22, 201.93]]}, {\"category\": \"car\", \"corners_3d\": [[812.32, 355.83], [997.59, 355.87], [1442.62, 566.07], [1044.44, 565.9], [812.32, 177.17], [997.59, 177.17], [1442.62, 182.13], [1044.44, 182.13]]}, {\"category\": \"car\", \"corners_3d\": [[734.85, 281.62], [832.08, 281.64], [928.32, 328.81], [788.92, 328.79], [734.85, 193.43], [832.08, 193.43], [928.32, 202.36], [788.92, 202.35]]}, {\"category\": \"car\", \"corners_3d\": [[563.81, 224.28], [518.98, 224.32], [530.61, 217.09], [569.14, 217.07], [563.81, 182.97], [518.98, 182.98], [530.61, 181.56], [569.14, 181.55]]}]\n```", - "options": null, - "id": 629 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005677", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[688.19, 247.28], [758.34, 247.07], [814.53, 271.61], [721.29, 271.99], [688.19, 173.26], [758.34, 173.26], [814.53, 173.4], [721.29, 173.4]]}, {\"category\": \"van\", \"corners_3d\": [[662.78, 218.24], [702.75, 218.17], [722.2, 225.88], [675.44, 225.98], [662.78, 172.62], [702.75, 172.62], [722.2, 172.58], [675.44, 172.58]]}]\n```", - "options": null, - "id": 630 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005677", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[959.08, 304.07], [987.57, 304.02], [1049.59, 325.17], [1016.52, 325.24], [959.08, 189.53], [987.57, 189.52], [1049.59, 192.21], [1016.52, 192.22]]}, {\"category\": \"car\", \"corners_3d\": [[202.85, 571.21], [-176.11, 575.51], [224.12, 366.4], [405.22, 365.4], [202.85, 233.0], [-176.11, 233.65], [224.12, 202.08], [405.22, 201.93]]}, {\"category\": \"car\", \"corners_3d\": [[812.32, 355.83], [997.59, 355.87], [1442.62, 566.07], [1044.44, 565.9], [812.32, 177.17], [997.59, 177.17], [1442.62, 182.13], [1044.44, 182.13]]}, {\"category\": \"car\", \"corners_3d\": [[734.85, 281.62], [832.08, 281.64], [928.32, 328.81], [788.92, 328.79], [734.85, 193.43], [832.08, 193.43], [928.32, 202.36], [788.92, 202.35]]}, {\"category\": \"car\", \"corners_3d\": [[563.81, 224.28], [518.98, 224.32], [530.61, 217.09], [569.14, 217.07], [563.81, 182.97], [518.98, 182.98], [530.61, 181.56], [569.14, 181.55]]}]\n```", - "options": null, - "id": 631 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005677", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[959.08, 304.07], [987.57, 304.02], [1049.59, 325.17], [1016.52, 325.24], [959.08, 189.53], [987.57, 189.52], [1049.59, 192.21], [1016.52, 192.22]]}, {\"category\": \"van\", \"corners_3d\": [[688.19, 247.28], [758.34, 247.07], [814.53, 271.61], [721.29, 271.99], [688.19, 173.26], [758.34, 173.26], [814.53, 173.4], [721.29, 173.4]]}, {\"category\": \"van\", \"corners_3d\": [[662.78, 218.24], [702.75, 218.17], [722.2, 225.88], [675.44, 225.98], [662.78, 172.62], [702.75, 172.62], [722.2, 172.58], [675.44, 172.58]]}]\n```", - "options": null, - "id": 632 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005677", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[202.85, 571.21], [-176.11, 575.51], [224.12, 366.4], [405.22, 365.4], [202.85, 233.0], [-176.11, 233.65], [224.12, 202.08], [405.22, 201.93]]}, {\"category\": \"car\", \"corners_3d\": [[812.32, 355.83], [997.59, 355.87], [1442.62, 566.07], [1044.44, 565.9], [812.32, 177.17], [997.59, 177.17], [1442.62, 182.13], [1044.44, 182.13]]}, {\"category\": \"car\", \"corners_3d\": [[734.85, 281.62], [832.08, 281.64], [928.32, 328.81], [788.92, 328.79], [734.85, 193.43], [832.08, 193.43], [928.32, 202.36], [788.92, 202.35]]}, {\"category\": \"van\", \"corners_3d\": [[688.19, 247.28], [758.34, 247.07], [814.53, 271.61], [721.29, 271.99], [688.19, 173.26], [758.34, 173.26], [814.53, 173.4], [721.29, 173.4]]}, {\"category\": \"car\", \"corners_3d\": [[563.81, 224.28], [518.98, 224.32], [530.61, 217.09], [569.14, 217.07], [563.81, 182.97], [518.98, 182.98], [530.61, 181.56], [569.14, 181.55]]}, {\"category\": \"van\", \"corners_3d\": [[662.78, 218.24], [702.75, 218.17], [722.2, 225.88], [675.44, 225.98], [662.78, 172.62], [702.75, 172.62], [722.2, 172.58], [675.44, 172.58]]}]\n```", - "options": null, - "id": 633 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005677", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[959.08, 304.07], [987.57, 304.02], [1049.59, 325.17], [1016.52, 325.24], [959.08, 189.53], [987.57, 189.52], [1049.59, 192.21], [1016.52, 192.22]]}, {\"category\": \"car\", \"corners_3d\": [[202.85, 571.21], [-176.11, 575.51], [224.12, 366.4], [405.22, 365.4], [202.85, 233.0], [-176.11, 233.65], [224.12, 202.08], [405.22, 201.93]]}, {\"category\": \"car\", \"corners_3d\": [[812.32, 355.83], [997.59, 355.87], [1442.62, 566.07], [1044.44, 565.9], [812.32, 177.17], [997.59, 177.17], [1442.62, 182.13], [1044.44, 182.13]]}, {\"category\": \"car\", \"corners_3d\": [[734.85, 281.62], [832.08, 281.64], [928.32, 328.81], [788.92, 328.79], [734.85, 193.43], [832.08, 193.43], [928.32, 202.36], [788.92, 202.35]]}, {\"category\": \"van\", \"corners_3d\": [[688.19, 247.28], [758.34, 247.07], [814.53, 271.61], [721.29, 271.99], [688.19, 173.26], [758.34, 173.26], [814.53, 173.4], [721.29, 173.4]]}, {\"category\": \"car\", \"corners_3d\": [[563.81, 224.28], [518.98, 224.32], [530.61, 217.09], [569.14, 217.07], [563.81, 182.97], [518.98, 182.98], [530.61, 181.56], [569.14, 181.55]]}, {\"category\": \"van\", \"corners_3d\": [[662.78, 218.24], [702.75, 218.17], [722.2, 225.88], [675.44, 225.98], [662.78, 172.62], [702.75, 172.62], [722.2, 172.58], [675.44, 172.58]]}]\n```", - "options": null, - "id": 634 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005683", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[449.75, 324.65], [458.59, 316.33], [550.16, 316.3], [546.93, 324.62], [449.75, 155.32], [458.59, 156.78], [550.16, 156.78], [546.93, 155.33]]}]\n```", - "options": null, - "id": 635 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005741", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[760.35, 290.28], [869.99, 289.02], [1084.02, 369.11], [901.34, 372.72], [760.35, 174.18], [869.99, 174.17], [1084.02, 175.07], [901.34, 175.11]]}, {\"category\": \"car\", \"corners_3d\": [[492.06, 293.25], [392.34, 294.09], [437.89, 263.5], [512.35, 263.03], [492.06, 202.46], [392.34, 202.67], [437.89, 195.15], [512.35, 195.03]]}, {\"category\": \"car\", \"corners_3d\": [[519.55, 218.42], [554.44, 218.22], [555.85, 225.0], [515.72, 225.26], [519.55, 184.74], [554.44, 184.69], [555.85, 186.46], [515.72, 186.53]]}, {\"category\": \"car\", \"corners_3d\": [[636.99, 217.82], [674.12, 217.73], [686.15, 223.54], [644.22, 223.66], [636.99, 176.69], [674.12, 176.68], [686.15, 177.18], [644.22, 177.19]]}]\n```", - "options": null, - "id": 636 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005756", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[634.8, 196.45], [610.89, 196.33], [619.39, 194.47], [641.44, 194.57], [634.8, 175.73], [610.89, 175.71], [619.39, 175.49], [641.44, 175.5]]}]\n```", - "options": null, - "id": 637 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005760", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[698.95, 299.69], [693.12, 292.46], [774.17, 292.44], [785.24, 299.67], [698.95, 141.43], [693.12, 143.8], [774.17, 143.81], [785.24, 141.43]]}]\n```", - "options": null, - "id": 638 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005767", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[793.67, 206.31], [822.72, 206.63], [825.71, 210.19], [793.59, 209.8], [793.67, 180.58], [822.72, 180.66], [825.71, 181.48], [793.59, 181.39]]}]\n```", - "options": null, - "id": 639 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005793", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[603.8, 216.73], [613.39, 216.28], [619.71, 216.77], [610.07, 217.23], [603.8, 175.77], [613.39, 175.74], [619.71, 175.77], [610.07, 175.8]]}]\n```", - "options": null, - "id": 640 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005793", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[246.27, 412.86], [53.14, 413.55], [262.87, 321.54], [381.93, 321.27], [246.27, 254.9], [53.14, 255.13], [262.87, 223.68], [381.93, 223.59]]}, {\"category\": \"car\", \"corners_3d\": [[432.64, 283.46], [331.0, 283.78], [396.71, 256.13], [472.92, 255.94], [432.64, 183.83], [331.0, 183.87], [396.71, 181.12], [472.92, 181.1]]}, {\"category\": \"car\", \"corners_3d\": [[614.55, 210.76], [650.09, 210.72], [656.27, 215.03], [616.69, 215.07], [614.55, 176.8], [650.09, 176.8], [656.27, 177.24], [616.69, 177.25]]}]\n```", - "options": null, - "id": 641 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005793", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[540.77, 213.14], [493.93, 212.93], [524.46, 205.85], [563.13, 205.99], [540.77, 141.19], [493.93, 141.35], [524.46, 146.92], [563.13, 146.81]]}]\n```", - "options": null, - "id": 642 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005793", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[246.27, 412.86], [53.14, 413.55], [262.87, 321.54], [381.93, 321.27], [246.27, 254.9], [53.14, 255.13], [262.87, 223.68], [381.93, 223.59]]}, {\"category\": \"car\", \"corners_3d\": [[432.64, 283.46], [331.0, 283.78], [396.71, 256.13], [472.92, 255.94], [432.64, 183.83], [331.0, 183.87], [396.71, 181.12], [472.92, 181.1]]}, {\"category\": \"car\", \"corners_3d\": [[614.55, 210.76], [650.09, 210.72], [656.27, 215.03], [616.69, 215.07], [614.55, 176.8], [650.09, 176.8], [656.27, 177.24], [616.69, 177.25]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[603.8, 216.73], [613.39, 216.28], [619.71, 216.77], [610.07, 217.23], [603.8, 175.77], [613.39, 175.74], [619.71, 175.77], [610.07, 175.8]]}]\n```", - "options": null, - "id": 643 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005793", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[540.77, 213.14], [493.93, 212.93], [524.46, 205.85], [563.13, 205.99], [540.77, 141.19], [493.93, 141.35], [524.46, 146.92], [563.13, 146.81]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[603.8, 216.73], [613.39, 216.28], [619.71, 216.77], [610.07, 217.23], [603.8, 175.77], [613.39, 175.74], [619.71, 175.77], [610.07, 175.8]]}]\n```", - "options": null, - "id": 644 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005793", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[246.27, 412.86], [53.14, 413.55], [262.87, 321.54], [381.93, 321.27], [246.27, 254.9], [53.14, 255.13], [262.87, 223.68], [381.93, 223.59]]}, {\"category\": \"car\", \"corners_3d\": [[432.64, 283.46], [331.0, 283.78], [396.71, 256.13], [472.92, 255.94], [432.64, 183.83], [331.0, 183.87], [396.71, 181.12], [472.92, 181.1]]}, {\"category\": \"truck\", \"corners_3d\": [[540.77, 213.14], [493.93, 212.93], [524.46, 205.85], [563.13, 205.99], [540.77, 141.19], [493.93, 141.35], [524.46, 146.92], [563.13, 146.81]]}, {\"category\": \"car\", \"corners_3d\": [[614.55, 210.76], [650.09, 210.72], [656.27, 215.03], [616.69, 215.07], [614.55, 176.8], [650.09, 176.8], [656.27, 177.24], [616.69, 177.25]]}]\n```", - "options": null, - "id": 645 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005793", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[246.27, 412.86], [53.14, 413.55], [262.87, 321.54], [381.93, 321.27], [246.27, 254.9], [53.14, 255.13], [262.87, 223.68], [381.93, 223.59]]}, {\"category\": \"car\", \"corners_3d\": [[432.64, 283.46], [331.0, 283.78], [396.71, 256.13], [472.92, 255.94], [432.64, 183.83], [331.0, 183.87], [396.71, 181.12], [472.92, 181.1]]}, {\"category\": \"truck\", \"corners_3d\": [[540.77, 213.14], [493.93, 212.93], [524.46, 205.85], [563.13, 205.99], [540.77, 141.19], [493.93, 141.35], [524.46, 146.92], [563.13, 146.81]]}, {\"category\": \"car\", \"corners_3d\": [[614.55, 210.76], [650.09, 210.72], [656.27, 215.03], [616.69, 215.07], [614.55, 176.8], [650.09, 176.8], [656.27, 177.24], [616.69, 177.25]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[603.8, 216.73], [613.39, 216.28], [619.71, 216.77], [610.07, 217.23], [603.8, 175.77], [613.39, 175.74], [619.71, 175.77], [610.07, 175.8]]}]\n```", - "options": null, - "id": 646 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005794", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[706.25, 190.45], [726.82, 190.69], [708.14, 192.07], [686.26, 191.8], [706.25, 162.65], [726.82, 162.52], [708.14, 161.72], [686.26, 161.87]]}]\n```", - "options": null, - "id": 647 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005794", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[690.87, 244.56], [761.98, 244.49], [822.17, 271.59], [724.18, 271.72], [690.87, 176.49], [761.98, 176.49], [822.17, 177.86], [724.18, 177.87]]}, {\"category\": \"car\", \"corners_3d\": [[350.91, 281.69], [455.15, 281.08], [376.66, 348.45], [206.23, 350.05], [350.91, 185.68], [455.15, 185.61], [376.66, 193.54], [206.23, 193.73]]}, {\"category\": \"car\", \"corners_3d\": [[645.51, 210.18], [682.33, 210.3], [685.86, 214.87], [644.54, 214.73], [645.51, 179.33], [682.33, 179.35], [685.86, 180.14], [644.54, 180.12]]}]\n```", - "options": null, - "id": 648 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005794", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[690.87, 244.56], [761.98, 244.49], [822.17, 271.59], [724.18, 271.72], [690.87, 176.49], [761.98, 176.49], [822.17, 177.86], [724.18, 177.87]]}, {\"category\": \"car\", \"corners_3d\": [[350.91, 281.69], [455.15, 281.08], [376.66, 348.45], [206.23, 350.05], [350.91, 185.68], [455.15, 185.61], [376.66, 193.54], [206.23, 193.73]]}, {\"category\": \"car\", \"corners_3d\": [[645.51, 210.18], [682.33, 210.3], [685.86, 214.87], [644.54, 214.73], [645.51, 179.33], [682.33, 179.35], [685.86, 180.14], [644.54, 180.12]]}, {\"category\": \"van\", \"corners_3d\": [[706.25, 190.45], [726.82, 190.69], [708.14, 192.07], [686.26, 191.8], [706.25, 162.65], [726.82, 162.52], [708.14, 161.72], [686.26, 161.87]]}]\n```", - "options": null, - "id": 649 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005802", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[634.33, 212.67], [673.54, 212.8], [676.35, 218.41], [631.65, 218.24], [634.33, 178.36], [673.54, 178.37], [676.35, 179.15], [631.65, 179.12]]}, {\"category\": \"car\", \"corners_3d\": [[586.94, 203.1], [620.46, 203.25], [612.19, 207.29], [574.28, 207.11], [586.94, 171.06], [620.46, 171.05], [612.19, 170.81], [574.28, 170.82]]}, {\"category\": \"car\", \"corners_3d\": [[648.32, 199.37], [678.26, 199.46], [679.58, 202.88], [645.79, 202.77], [648.32, 171.38], [678.26, 171.38], [679.58, 171.19], [645.79, 171.2]]}, {\"category\": \"car\", \"corners_3d\": [[1055.79, 189.87], [1090.25, 190.85], [1002.35, 191.39], [971.82, 190.36], [1055.79, 152.81], [1090.25, 151.66], [1002.35, 151.02], [971.82, 152.24]]}, {\"category\": \"car\", \"corners_3d\": [[999.92, 189.76], [1028.08, 190.63], [952.2, 191.09], [927.18, 190.17], [999.92, 154.81], [1028.08, 153.87], [952.2, 153.39], [927.18, 154.37]]}]\n```", - "options": null, - "id": 650 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005875", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[513.19, 209.59], [470.94, 209.49], [489.44, 206.48], [527.18, 206.56], [513.19, 175.47], [470.94, 175.49], [489.44, 176.14], [527.18, 176.12]]}]\n```", - "options": null, - "id": 651 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005889", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[826.16, 256.6], [909.86, 259.18], [938.03, 285.55], [828.65, 281.2], [826.16, 173.83], [909.86, 173.86], [938.03, 174.17], [828.65, 174.12]]}, {\"category\": \"car\", \"corners_3d\": [[696.37, 213.66], [657.94, 213.15], [672.9, 208.42], [706.94, 208.83], [696.37, 179.25], [657.94, 179.17], [672.9, 178.43], [706.94, 178.49]]}, {\"category\": \"car\", \"corners_3d\": [[835.89, 206.38], [805.56, 206.05], [804.48, 203.15], [832.12, 203.42], [835.89, 176.37], [805.56, 176.33], [804.48, 176.03], [832.12, 176.06]]}, {\"category\": \"car\", \"corners_3d\": [[732.0, 201.06], [706.3, 200.79], [715.4, 198.48], [739.04, 198.71], [732.0, 178.71], [706.3, 178.65], [715.4, 178.17], [739.04, 178.22]]}]\n```", - "options": null, - "id": 652 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005890", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[400.67, 307.44], [341.22, 336.65], [-38.41, 328.6], [83.56, 301.96], [400.67, 200.06], [341.22, 205.96], [-38.41, 204.34], [83.56, 198.95]]}, {\"category\": \"car\", \"corners_3d\": [[1613.56, 385.8], [2190.94, 513.79], [1170.81, 488.2], [985.77, 375.53], [1613.56, 169.88], [2190.94, 168.1], [1170.81, 168.46], [985.77, 170.03]]}, {\"category\": \"car\", \"corners_3d\": [[1217.57, 305.57], [1372.76, 338.66], [903.03, 340.86], [840.93, 306.98], [1217.57, 154.6], [1372.76, 150.05], [903.03, 149.75], [840.93, 154.41]]}, {\"category\": \"car\", \"corners_3d\": [[320.31, 240.24], [350.18, 234.7], [500.07, 236.06], [483.25, 241.86], [320.31, 182.97], [350.18, 182.14], [500.07, 182.34], [483.25, 183.21]]}, {\"category\": \"car\", \"corners_3d\": [[393.49, 223.63], [403.41, 220.27], [529.76, 219.52], [528.82, 222.78], [393.49, 181.15], [403.41, 180.6], [529.76, 180.48], [528.82, 181.01]]}, {\"category\": \"car\", \"corners_3d\": [[947.06, 243.49], [985.33, 251.67], [815.54, 251.49], [794.94, 243.34], [947.06, 170.47], [985.33, 170.19], [815.54, 170.2], [794.94, 170.47]]}, {\"category\": \"car\", \"corners_3d\": [[1032.31, 254.03], [1095.32, 266.34], [838.16, 266.01], [809.1, 253.78], [1032.31, 170.62], [1095.32, 170.29], [838.16, 170.3], [809.1, 170.63]]}]\n```", - "options": null, - "id": 653 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005890", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[537.47, 209.34], [551.09, 209.0], [561.41, 210.18], [547.44, 210.54], [537.47, 168.01], [551.09, 168.05], [561.41, 167.9], [547.44, 167.85]]}]\n```", - "options": null, - "id": 654 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005890", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[400.67, 307.44], [341.22, 336.65], [-38.41, 328.6], [83.56, 301.96], [400.67, 200.06], [341.22, 205.96], [-38.41, 204.34], [83.56, 198.95]]}, {\"category\": \"car\", \"corners_3d\": [[1613.56, 385.8], [2190.94, 513.79], [1170.81, 488.2], [985.77, 375.53], [1613.56, 169.88], [2190.94, 168.1], [1170.81, 168.46], [985.77, 170.03]]}, {\"category\": \"car\", \"corners_3d\": [[1217.57, 305.57], [1372.76, 338.66], [903.03, 340.86], [840.93, 306.98], [1217.57, 154.6], [1372.76, 150.05], [903.03, 149.75], [840.93, 154.41]]}, {\"category\": \"car\", \"corners_3d\": [[320.31, 240.24], [350.18, 234.7], [500.07, 236.06], [483.25, 241.86], [320.31, 182.97], [350.18, 182.14], [500.07, 182.34], [483.25, 183.21]]}, {\"category\": \"car\", \"corners_3d\": [[393.49, 223.63], [403.41, 220.27], [529.76, 219.52], [528.82, 222.78], [393.49, 181.15], [403.41, 180.6], [529.76, 180.48], [528.82, 181.01]]}, {\"category\": \"car\", \"corners_3d\": [[947.06, 243.49], [985.33, 251.67], [815.54, 251.49], [794.94, 243.34], [947.06, 170.47], [985.33, 170.19], [815.54, 170.2], [794.94, 170.47]]}, {\"category\": \"car\", \"corners_3d\": [[1032.31, 254.03], [1095.32, 266.34], [838.16, 266.01], [809.1, 253.78], [1032.31, 170.62], [1095.32, 170.29], [838.16, 170.3], [809.1, 170.63]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[537.47, 209.34], [551.09, 209.0], [561.41, 210.18], [547.44, 210.54], [537.47, 168.01], [551.09, 168.05], [561.41, 167.9], [547.44, 167.85]]}]\n```", - "options": null, - "id": 655 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005895", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[534.43, 216.39], [490.2, 215.99], [522.08, 209.95], [560.33, 210.25], [534.43, 181.09], [490.2, 181.01], [522.08, 179.87], [560.33, 179.93]]}]\n```", - "options": null, - "id": 656 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005895", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[666.34, 204.22], [666.97, 203.79], [698.86, 203.98], [698.67, 204.42], [666.34, 171.59], [666.97, 171.6], [698.86, 171.6], [698.67, 171.58]]}]\n```", - "options": null, - "id": 657 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005895", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[476.77, 238.25], [399.26, 238.18], [456.82, 220.85], [513.79, 220.88], [476.77, 142.83], [399.26, 142.86], [456.82, 150.82], [513.79, 150.8]]}]\n```", - "options": null, - "id": 658 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005895", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[534.43, 216.39], [490.2, 215.99], [522.08, 209.95], [560.33, 210.25], [534.43, 181.09], [490.2, 181.01], [522.08, 179.87], [560.33, 179.93]]}, {\"category\": \"cyclist\", \"corners_3d\": [[666.34, 204.22], [666.97, 203.79], [698.86, 203.98], [698.67, 204.42], [666.34, 171.59], [666.97, 171.6], [698.86, 171.6], [698.67, 171.58]]}]\n```", - "options": null, - "id": 659 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005895", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[476.77, 238.25], [399.26, 238.18], [456.82, 220.85], [513.79, 220.88], [476.77, 142.83], [399.26, 142.86], [456.82, 150.82], [513.79, 150.8]]}, {\"category\": \"car\", \"corners_3d\": [[534.43, 216.39], [490.2, 215.99], [522.08, 209.95], [560.33, 210.25], [534.43, 181.09], [490.2, 181.01], [522.08, 179.87], [560.33, 179.93]]}]\n```", - "options": null, - "id": 660 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005895", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[476.77, 238.25], [399.26, 238.18], [456.82, 220.85], [513.79, 220.88], [476.77, 142.83], [399.26, 142.86], [456.82, 150.82], [513.79, 150.8]]}, {\"category\": \"cyclist\", \"corners_3d\": [[666.34, 204.22], [666.97, 203.79], [698.86, 203.98], [698.67, 204.42], [666.34, 171.59], [666.97, 171.6], [698.86, 171.6], [698.67, 171.58]]}]\n```", - "options": null, - "id": 661 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005895", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[476.77, 238.25], [399.26, 238.18], [456.82, 220.85], [513.79, 220.88], [476.77, 142.83], [399.26, 142.86], [456.82, 150.82], [513.79, 150.8]]}, {\"category\": \"car\", \"corners_3d\": [[534.43, 216.39], [490.2, 215.99], [522.08, 209.95], [560.33, 210.25], [534.43, 181.09], [490.2, 181.01], [522.08, 179.87], [560.33, 179.93]]}, {\"category\": \"cyclist\", \"corners_3d\": [[666.34, 204.22], [666.97, 203.79], [698.86, 203.98], [698.67, 204.42], [666.34, 171.59], [666.97, 171.6], [698.86, 171.6], [698.67, 171.58]]}]\n```", - "options": null, - "id": 662 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005951", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[548.3, 227.14], [499.84, 227.14], [515.4, 219.4], [556.95, 219.4], [548.3, 185.69], [499.84, 185.69], [515.4, 183.86], [556.95, 183.86]]}, {\"category\": \"car\", \"corners_3d\": [[589.48, 187.07], [570.19, 187.08], [571.67, 186.21], [589.79, 186.21], [589.48, 166.77], [570.19, 166.77], [571.67, 167.14], [589.79, 167.14]]}]\n```", - "options": null, - "id": 663 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005996", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1234.68, 410.64], [1102.1, 356.35], [1570.09, 363.39], [1854.79, 422.61], [1234.68, 222.06], [1102.1, 210.82], [1570.09, 212.28], [1854.79, 224.54]]}, {\"category\": \"car\", \"corners_3d\": [[1283.39, 310.55], [1406.61, 339.11], [1046.62, 330.94], [987.76, 304.9], [1283.39, 177.71], [1406.61, 178.72], [1046.62, 178.43], [987.76, 177.52]]}, {\"category\": \"car\", \"corners_3d\": [[814.03, 252.37], [791.83, 244.35], [977.03, 243.96], [1019.73, 251.88], [814.03, 169.05], [791.83, 169.43], [977.03, 169.45], [1019.73, 169.07]]}, {\"category\": \"car\", \"corners_3d\": [[354.94, 262.06], [318.01, 274.99], [63.85, 274.99], [132.98, 262.06], [354.94, 141.28], [318.01, 136.7], [63.85, 136.7], [132.98, 141.28]]}, {\"category\": \"car\", \"corners_3d\": [[179.48, 252.36], [218.91, 244.93], [368.74, 244.75], [344.83, 252.15], [179.48, 180.99], [218.91, 180.23], [368.74, 180.21], [344.83, 180.97]]}, {\"category\": \"car\", \"corners_3d\": [[312.77, 228.7], [333.65, 224.76], [470.12, 224.74], [459.61, 228.68], [312.77, 178.97], [333.65, 178.54], [470.12, 178.54], [459.61, 178.97]]}, {\"category\": \"car\", \"corners_3d\": [[381.69, 218.8], [390.57, 215.88], [506.73, 215.06], [505.75, 217.87], [381.69, 168.67], [390.57, 168.93], [506.73, 169.01], [505.75, 168.75]]}]\n```", - "options": null, - "id": 664 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005996", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[653.96, 210.01], [702.47, 209.99], [728.88, 219.84], [667.51, 219.88], [653.96, 152.28], [702.47, 152.29], [728.88, 146.84], [667.51, 146.81]]}, {\"category\": \"van\", \"corners_3d\": [[632.71, 198.7], [663.32, 198.69], [673.28, 202.95], [637.63, 202.97], [632.71, 159.15], [663.32, 159.16], [673.28, 156.9], [637.63, 156.89]]}, {\"category\": \"van\", \"corners_3d\": [[565.99, 192.25], [564.5, 192.82], [499.06, 192.8], [502.42, 192.23], [565.99, 160.91], [564.5, 160.55], [499.06, 160.57], [502.42, 160.92]]}]\n```", - "options": null, - "id": 665 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "005996", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1234.68, 410.64], [1102.1, 356.35], [1570.09, 363.39], [1854.79, 422.61], [1234.68, 222.06], [1102.1, 210.82], [1570.09, 212.28], [1854.79, 224.54]]}, {\"category\": \"car\", \"corners_3d\": [[1283.39, 310.55], [1406.61, 339.11], [1046.62, 330.94], [987.76, 304.9], [1283.39, 177.71], [1406.61, 178.72], [1046.62, 178.43], [987.76, 177.52]]}, {\"category\": \"car\", \"corners_3d\": [[814.03, 252.37], [791.83, 244.35], [977.03, 243.96], [1019.73, 251.88], [814.03, 169.05], [791.83, 169.43], [977.03, 169.45], [1019.73, 169.07]]}, {\"category\": \"car\", \"corners_3d\": [[354.94, 262.06], [318.01, 274.99], [63.85, 274.99], [132.98, 262.06], [354.94, 141.28], [318.01, 136.7], [63.85, 136.7], [132.98, 141.28]]}, {\"category\": \"car\", \"corners_3d\": [[179.48, 252.36], [218.91, 244.93], [368.74, 244.75], [344.83, 252.15], [179.48, 180.99], [218.91, 180.23], [368.74, 180.21], [344.83, 180.97]]}, {\"category\": \"van\", \"corners_3d\": [[653.96, 210.01], [702.47, 209.99], [728.88, 219.84], [667.51, 219.88], [653.96, 152.28], [702.47, 152.29], [728.88, 146.84], [667.51, 146.81]]}, {\"category\": \"car\", \"corners_3d\": [[312.77, 228.7], [333.65, 224.76], [470.12, 224.74], [459.61, 228.68], [312.77, 178.97], [333.65, 178.54], [470.12, 178.54], [459.61, 178.97]]}, {\"category\": \"car\", \"corners_3d\": [[381.69, 218.8], [390.57, 215.88], [506.73, 215.06], [505.75, 217.87], [381.69, 168.67], [390.57, 168.93], [506.73, 169.01], [505.75, 168.75]]}, {\"category\": \"van\", \"corners_3d\": [[632.71, 198.7], [663.32, 198.69], [673.28, 202.95], [637.63, 202.97], [632.71, 159.15], [663.32, 159.16], [673.28, 156.9], [637.63, 156.89]]}, {\"category\": \"van\", \"corners_3d\": [[565.99, 192.25], [564.5, 192.82], [499.06, 192.8], [502.42, 192.23], [565.99, 160.91], [564.5, 160.55], [499.06, 160.57], [502.42, 160.92]]}]\n```", - "options": null, - "id": 666 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006016", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1041.31, 215.46], [1084.4, 219.72], [892.15, 219.72], [866.51, 215.46], [1041.31, 131.77], [1084.4, 127.67], [892.15, 127.67], [866.51, 131.77]]}]\n```", - "options": null, - "id": 667 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006016", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1009.48, 442.04], [1092.61, 442.68], [1412.77, 627.24], [1271.73, 625.44], [1009.48, 196.01], [1092.61, 196.06], [1412.77, 211.94], [1271.73, 211.78]]}]\n```", - "options": null, - "id": 668 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006016", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[479.79, 275.34], [388.94, 273.76], [459.17, 250.68], [529.84, 251.62], [479.79, 191.95], [388.94, 191.65], [459.17, 187.35], [529.84, 187.53]]}, {\"category\": \"car\", \"corners_3d\": [[535.34, 252.49], [462.85, 251.52], [510.16, 235.56], [568.27, 236.17], [535.34, 175.03], [462.85, 175.0], [510.16, 174.56], [568.27, 174.58]]}, {\"category\": \"car\", \"corners_3d\": [[726.79, 215.6], [769.79, 215.75], [788.04, 222.49], [738.24, 222.28], [726.79, 179.56], [769.79, 179.58], [788.04, 180.64], [738.24, 180.6]]}, {\"category\": \"car\", \"corners_3d\": [[590.46, 224.37], [545.22, 224.02], [564.03, 217.31], [603.41, 217.58], [590.46, 176.19], [545.22, 176.16], [564.03, 175.73], [603.41, 175.75]]}, {\"category\": \"car\", \"corners_3d\": [[713.67, 208.91], [752.57, 209.03], [766.54, 214.13], [722.12, 213.98], [713.67, 172.64], [752.57, 172.64], [766.54, 172.61], [722.12, 172.61]]}, {\"category\": \"car\", \"corners_3d\": [[604.42, 215.23], [569.81, 215.03], [580.37, 211.04], [611.74, 211.2], [604.42, 176.32], [569.81, 176.3], [580.37, 175.98], [611.74, 175.99]]}, {\"category\": \"car\", \"corners_3d\": [[-382.66, 311.45], [-228.14, 289.02], [79.65, 286.61], [-13.11, 308.03], [-382.66, 183.09], [-228.14, 181.44], [79.65, 181.26], [-13.11, 182.84]]}]\n```", - "options": null, - "id": 669 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006016", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1009.48, 442.04], [1092.61, 442.68], [1412.77, 627.24], [1271.73, 625.44], [1009.48, 196.01], [1092.61, 196.06], [1412.77, 211.94], [1271.73, 211.78]]}, {\"category\": \"van\", \"corners_3d\": [[1041.31, 215.46], [1084.4, 219.72], [892.15, 219.72], [866.51, 215.46], [1041.31, 131.77], [1084.4, 127.67], [892.15, 127.67], [866.51, 131.77]]}]\n```", - "options": null, - "id": 670 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006016", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[479.79, 275.34], [388.94, 273.76], [459.17, 250.68], [529.84, 251.62], [479.79, 191.95], [388.94, 191.65], [459.17, 187.35], [529.84, 187.53]]}, {\"category\": \"car\", \"corners_3d\": [[535.34, 252.49], [462.85, 251.52], [510.16, 235.56], [568.27, 236.17], [535.34, 175.03], [462.85, 175.0], [510.16, 174.56], [568.27, 174.58]]}, {\"category\": \"van\", \"corners_3d\": [[1041.31, 215.46], [1084.4, 219.72], [892.15, 219.72], [866.51, 215.46], [1041.31, 131.77], [1084.4, 127.67], [892.15, 127.67], [866.51, 131.77]]}, {\"category\": \"car\", \"corners_3d\": [[726.79, 215.6], [769.79, 215.75], [788.04, 222.49], [738.24, 222.28], [726.79, 179.56], [769.79, 179.58], [788.04, 180.64], [738.24, 180.6]]}, {\"category\": \"car\", \"corners_3d\": [[590.46, 224.37], [545.22, 224.02], [564.03, 217.31], [603.41, 217.58], [590.46, 176.19], [545.22, 176.16], [564.03, 175.73], [603.41, 175.75]]}, {\"category\": \"car\", \"corners_3d\": [[713.67, 208.91], [752.57, 209.03], [766.54, 214.13], [722.12, 213.98], [713.67, 172.64], [752.57, 172.64], [766.54, 172.61], [722.12, 172.61]]}, {\"category\": \"car\", \"corners_3d\": [[604.42, 215.23], [569.81, 215.03], [580.37, 211.04], [611.74, 211.2], [604.42, 176.32], [569.81, 176.3], [580.37, 175.98], [611.74, 175.99]]}, {\"category\": \"car\", \"corners_3d\": [[-382.66, 311.45], [-228.14, 289.02], [79.65, 286.61], [-13.11, 308.03], [-382.66, 183.09], [-228.14, 181.44], [79.65, 181.26], [-13.11, 182.84]]}]\n```", - "options": null, - "id": 671 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006016", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1009.48, 442.04], [1092.61, 442.68], [1412.77, 627.24], [1271.73, 625.44], [1009.48, 196.01], [1092.61, 196.06], [1412.77, 211.94], [1271.73, 211.78]]}, {\"category\": \"car\", \"corners_3d\": [[479.79, 275.34], [388.94, 273.76], [459.17, 250.68], [529.84, 251.62], [479.79, 191.95], [388.94, 191.65], [459.17, 187.35], [529.84, 187.53]]}, {\"category\": \"car\", \"corners_3d\": [[535.34, 252.49], [462.85, 251.52], [510.16, 235.56], [568.27, 236.17], [535.34, 175.03], [462.85, 175.0], [510.16, 174.56], [568.27, 174.58]]}, {\"category\": \"car\", \"corners_3d\": [[726.79, 215.6], [769.79, 215.75], [788.04, 222.49], [738.24, 222.28], [726.79, 179.56], [769.79, 179.58], [788.04, 180.64], [738.24, 180.6]]}, {\"category\": \"car\", \"corners_3d\": [[590.46, 224.37], [545.22, 224.02], [564.03, 217.31], [603.41, 217.58], [590.46, 176.19], [545.22, 176.16], [564.03, 175.73], [603.41, 175.75]]}, {\"category\": \"car\", \"corners_3d\": [[713.67, 208.91], [752.57, 209.03], [766.54, 214.13], [722.12, 213.98], [713.67, 172.64], [752.57, 172.64], [766.54, 172.61], [722.12, 172.61]]}, {\"category\": \"car\", \"corners_3d\": [[604.42, 215.23], [569.81, 215.03], [580.37, 211.04], [611.74, 211.2], [604.42, 176.32], [569.81, 176.3], [580.37, 175.98], [611.74, 175.99]]}, {\"category\": \"car\", \"corners_3d\": [[-382.66, 311.45], [-228.14, 289.02], [79.65, 286.61], [-13.11, 308.03], [-382.66, 183.09], [-228.14, 181.44], [79.65, 181.26], [-13.11, 182.84]]}]\n```", - "options": null, - "id": 672 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006016", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1009.48, 442.04], [1092.61, 442.68], [1412.77, 627.24], [1271.73, 625.44], [1009.48, 196.01], [1092.61, 196.06], [1412.77, 211.94], [1271.73, 211.78]]}, {\"category\": \"car\", \"corners_3d\": [[479.79, 275.34], [388.94, 273.76], [459.17, 250.68], [529.84, 251.62], [479.79, 191.95], [388.94, 191.65], [459.17, 187.35], [529.84, 187.53]]}, {\"category\": \"car\", \"corners_3d\": [[535.34, 252.49], [462.85, 251.52], [510.16, 235.56], [568.27, 236.17], [535.34, 175.03], [462.85, 175.0], [510.16, 174.56], [568.27, 174.58]]}, {\"category\": \"van\", \"corners_3d\": [[1041.31, 215.46], [1084.4, 219.72], [892.15, 219.72], [866.51, 215.46], [1041.31, 131.77], [1084.4, 127.67], [892.15, 127.67], [866.51, 131.77]]}, {\"category\": \"car\", \"corners_3d\": [[726.79, 215.6], [769.79, 215.75], [788.04, 222.49], [738.24, 222.28], [726.79, 179.56], [769.79, 179.58], [788.04, 180.64], [738.24, 180.6]]}, {\"category\": \"car\", \"corners_3d\": [[590.46, 224.37], [545.22, 224.02], [564.03, 217.31], [603.41, 217.58], [590.46, 176.19], [545.22, 176.16], [564.03, 175.73], [603.41, 175.75]]}, {\"category\": \"car\", \"corners_3d\": [[713.67, 208.91], [752.57, 209.03], [766.54, 214.13], [722.12, 213.98], [713.67, 172.64], [752.57, 172.64], [766.54, 172.61], [722.12, 172.61]]}, {\"category\": \"car\", \"corners_3d\": [[604.42, 215.23], [569.81, 215.03], [580.37, 211.04], [611.74, 211.2], [604.42, 176.32], [569.81, 176.3], [580.37, 175.98], [611.74, 175.99]]}, {\"category\": \"car\", \"corners_3d\": [[-382.66, 311.45], [-228.14, 289.02], [79.65, 286.61], [-13.11, 308.03], [-382.66, 183.09], [-228.14, 181.44], [79.65, 181.26], [-13.11, 182.84]]}]\n```", - "options": null, - "id": 673 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006019", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[429.12, 210.0], [419.87, 209.97], [427.98, 208.67], [436.92, 208.7], [429.12, 182.09], [419.87, 182.08], [427.98, 181.76], [436.92, 181.77]]}]\n```", - "options": null, - "id": 674 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006019", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[555.55, 205.52], [529.49, 205.53], [535.69, 202.73], [559.52, 202.72], [555.55, 180.85], [529.49, 180.85], [535.69, 180.16], [559.52, 180.16]]}, {\"category\": \"car\", \"corners_3d\": [[489.99, 244.39], [430.28, 244.27], [464.04, 231.79], [513.34, 231.87], [489.99, 180.92], [430.28, 180.91], [464.04, 179.5], [513.34, 179.51]]}]\n```", - "options": null, - "id": 675 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006019", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[555.55, 205.52], [529.49, 205.53], [535.69, 202.73], [559.52, 202.72], [555.55, 180.85], [529.49, 180.85], [535.69, 180.16], [559.52, 180.16]]}, {\"category\": \"car\", \"corners_3d\": [[489.99, 244.39], [430.28, 244.27], [464.04, 231.79], [513.34, 231.87], [489.99, 180.92], [430.28, 180.91], [464.04, 179.5], [513.34, 179.51]]}, {\"category\": \"cyclist\", \"corners_3d\": [[429.12, 210.0], [419.87, 209.97], [427.98, 208.67], [436.92, 208.7], [429.12, 182.09], [419.87, 182.08], [427.98, 181.76], [436.92, 181.77]]}]\n```", - "options": null, - "id": 676 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006025", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[709.6, 267.18], [801.71, 267.19], [881.28, 306.37], [750.9, 306.35], [709.6, 188.0], [801.71, 188.0], [881.28, 194.29], [750.9, 194.28]]}, {\"category\": \"car\", \"corners_3d\": [[524.37, 218.1], [487.28, 218.05], [501.46, 213.35], [534.69, 213.38], [524.37, 187.86], [487.28, 187.85], [501.46, 186.28], [534.69, 186.3]]}, {\"category\": \"car\", \"corners_3d\": [[380.36, 283.03], [294.14, 282.77], [369.0, 257.79], [435.71, 257.94], [380.36, 202.54], [294.14, 202.48], [369.0, 195.74], [435.71, 195.78]]}, {\"category\": \"car\", \"corners_3d\": [[443.81, 253.61], [379.58, 253.33], [424.59, 239.31], [477.72, 239.5], [443.81, 195.58], [379.58, 195.5], [424.59, 191.56], [477.72, 191.61]]}, {\"category\": \"car\", \"corners_3d\": [[-36.81, 439.81], [-247.72, 437.47], [143.63, 319.57], [262.02, 320.29], [-36.81, 258.58], [-247.72, 257.83], [143.63, 219.97], [262.02, 220.2]]}]\n```", - "options": null, - "id": 677 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006062", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[620.64, 245.68], [691.1, 245.61], [713.47, 264.07], [625.14, 264.18], [620.64, 178.37], [691.1, 178.36], [713.47, 179.76], [625.14, 179.77]]}]\n```", - "options": null, - "id": 678 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006062", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[301.96, 362.95], [71.38, 364.2], [337.93, 266.71], [450.55, 266.41], [301.96, 19.65], [71.38, 18.64], [337.93, 97.21], [450.55, 97.46]]}]\n```", - "options": null, - "id": 679 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006062", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[600.23, 255.06], [620.13, 253.39], [632.97, 255.28], [612.87, 257.03], [600.23, 170.83], [620.13, 170.87], [632.97, 170.83], [612.87, 170.78]]}]\n```", - "options": null, - "id": 680 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006062", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[301.96, 362.95], [71.38, 364.2], [337.93, 266.71], [450.55, 266.41], [301.96, 19.65], [71.38, 18.64], [337.93, 97.21], [450.55, 97.46]]}, {\"category\": \"car\", \"corners_3d\": [[620.64, 245.68], [691.1, 245.61], [713.47, 264.07], [625.14, 264.18], [620.64, 178.37], [691.1, 178.36], [713.47, 179.76], [625.14, 179.77]]}]\n```", - "options": null, - "id": 681 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006062", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[620.64, 245.68], [691.1, 245.61], [713.47, 264.07], [625.14, 264.18], [620.64, 178.37], [691.1, 178.36], [713.47, 179.76], [625.14, 179.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[600.23, 255.06], [620.13, 253.39], [632.97, 255.28], [612.87, 257.03], [600.23, 170.83], [620.13, 170.87], [632.97, 170.83], [612.87, 170.78]]}]\n```", - "options": null, - "id": 682 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006062", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[301.96, 362.95], [71.38, 364.2], [337.93, 266.71], [450.55, 266.41], [301.96, 19.65], [71.38, 18.64], [337.93, 97.21], [450.55, 97.46]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[600.23, 255.06], [620.13, 253.39], [632.97, 255.28], [612.87, 257.03], [600.23, 170.83], [620.13, 170.87], [632.97, 170.83], [612.87, 170.78]]}]\n```", - "options": null, - "id": 683 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006062", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, truck, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[301.96, 362.95], [71.38, 364.2], [337.93, 266.71], [450.55, 266.41], [301.96, 19.65], [71.38, 18.64], [337.93, 97.21], [450.55, 97.46]]}, {\"category\": \"car\", \"corners_3d\": [[620.64, 245.68], [691.1, 245.61], [713.47, 264.07], [625.14, 264.18], [620.64, 178.37], [691.1, 178.36], [713.47, 179.76], [625.14, 179.77]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[600.23, 255.06], [620.13, 253.39], [632.97, 255.28], [612.87, 257.03], [600.23, 170.83], [620.13, 170.87], [632.97, 170.83], [612.87, 170.78]]}]\n```", - "options": null, - "id": 684 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006072", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-17.26, 242.72], [-71.44, 242.68], [31.87, 232.19], [77.97, 232.22], [-17.26, 185.72], [-71.44, 185.71], [31.87, 183.78], [77.97, 183.79]]}, {\"category\": \"car\", \"corners_3d\": [[111.21, 229.08], [62.2, 228.84], [156.39, 219.9], [197.88, 220.07], [111.21, 171.88], [62.2, 171.88], [156.39, 172.04], [197.88, 172.03]]}, {\"category\": \"car\", \"corners_3d\": [[219.18, 220.15], [183.07, 219.96], [238.25, 214.58], [270.43, 214.74], [219.18, 181.38], [183.07, 181.34], [238.25, 180.37], [270.43, 180.4]]}]\n```", - "options": null, - "id": 685 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006142", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[574.34, 224.73], [525.84, 224.94], [532.15, 216.68], [572.96, 216.53], [574.34, 181.95], [525.84, 181.98], [532.15, 180.54], [572.96, 180.51]]}]\n```", - "options": null, - "id": 686 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006156", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[435.3, 253.95], [407.48, 266.38], [68.23, 265.94], [140.84, 253.62], [435.3, 108.05], [407.48, 98.11], [68.23, 98.46], [140.84, 108.31]]}]\n```", - "options": null, - "id": 687 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006156", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-145.35, 318.63], [-29.42, 294.84], [165.63, 292.91], [89.2, 315.88], [-145.35, 163.16], [-29.42, 164.74], [165.63, 164.87], [89.2, 163.35]]}, {\"category\": \"car\", \"corners_3d\": [[385.16, 270.31], [353.83, 286.52], [22.64, 289.09], [102.73, 272.19], [385.16, 177.18], [353.83, 177.9], [22.64, 178.01], [102.73, 177.26]]}, {\"category\": \"car\", \"corners_3d\": [[733.08, 274.21], [835.11, 274.22], [964.83, 332.67], [803.95, 332.64], [733.08, 185.43], [835.11, 185.43], [964.83, 192.68], [803.95, 192.67]]}, {\"category\": \"car\", \"corners_3d\": [[1002.55, 250.4], [1043.4, 259.65], [865.33, 258.21], [843.79, 249.25], [1002.55, 181.56], [1043.4, 182.6], [865.33, 182.44], [843.79, 181.43]]}, {\"category\": \"car\", \"corners_3d\": [[772.67, 239.66], [759.92, 233.6], [925.85, 234.03], [955.4, 240.19], [772.67, 179.77], [759.92, 179.14], [925.85, 179.18], [955.4, 179.82]]}, {\"category\": \"car\", \"corners_3d\": [[304.16, 228.18], [325.86, 223.94], [465.59, 223.63], [455.58, 227.82], [304.16, 172.48], [325.86, 172.51], [465.59, 172.51], [455.58, 172.48]]}, {\"category\": \"car\", \"corners_3d\": [[633.93, 193.05], [655.43, 193.04], [659.69, 194.68], [636.44, 194.69], [633.93, 175.62], [655.43, 175.62], [659.69, 175.85], [636.44, 175.85]]}, {\"category\": \"car\", \"corners_3d\": [[549.45, 194.95], [547.14, 195.7], [479.81, 195.68], [484.35, 194.93], [549.45, 170.7], [547.14, 170.63], [479.81, 170.63], [484.35, 170.7]]}, {\"category\": \"car\", \"corners_3d\": [[756.55, 203.79], [763.56, 205.13], [688.73, 205.2], [684.83, 203.85], [756.55, 176.0], [763.56, 176.13], [688.73, 176.14], [684.83, 176.0]]}, {\"category\": \"car\", \"corners_3d\": [[518.4, 212.48], [509.84, 214.65], [410.03, 214.12], [423.6, 212.0], [518.4, 174.61], [509.84, 174.71], [410.03, 174.68], [423.6, 174.59]]}, {\"category\": \"car\", \"corners_3d\": [[812.49, 211.59], [826.84, 213.78], [737.24, 214.15], [727.61, 211.92], [812.49, 173.58], [826.84, 173.62], [737.24, 173.63], [727.61, 173.59]]}]\n```", - "options": null, - "id": 688 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006156", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-145.35, 318.63], [-29.42, 294.84], [165.63, 292.91], [89.2, 315.88], [-145.35, 163.16], [-29.42, 164.74], [165.63, 164.87], [89.2, 163.35]]}, {\"category\": \"car\", \"corners_3d\": [[385.16, 270.31], [353.83, 286.52], [22.64, 289.09], [102.73, 272.19], [385.16, 177.18], [353.83, 177.9], [22.64, 178.01], [102.73, 177.26]]}, {\"category\": \"van\", \"corners_3d\": [[435.3, 253.95], [407.48, 266.38], [68.23, 265.94], [140.84, 253.62], [435.3, 108.05], [407.48, 98.11], [68.23, 98.46], [140.84, 108.31]]}, {\"category\": \"car\", \"corners_3d\": [[733.08, 274.21], [835.11, 274.22], [964.83, 332.67], [803.95, 332.64], [733.08, 185.43], [835.11, 185.43], [964.83, 192.68], [803.95, 192.67]]}, {\"category\": \"car\", \"corners_3d\": [[1002.55, 250.4], [1043.4, 259.65], [865.33, 258.21], [843.79, 249.25], [1002.55, 181.56], [1043.4, 182.6], [865.33, 182.44], [843.79, 181.43]]}, {\"category\": \"car\", \"corners_3d\": [[772.67, 239.66], [759.92, 233.6], [925.85, 234.03], [955.4, 240.19], [772.67, 179.77], [759.92, 179.14], [925.85, 179.18], [955.4, 179.82]]}, {\"category\": \"car\", \"corners_3d\": [[304.16, 228.18], [325.86, 223.94], [465.59, 223.63], [455.58, 227.82], [304.16, 172.48], [325.86, 172.51], [465.59, 172.51], [455.58, 172.48]]}, {\"category\": \"car\", \"corners_3d\": [[633.93, 193.05], [655.43, 193.04], [659.69, 194.68], [636.44, 194.69], [633.93, 175.62], [655.43, 175.62], [659.69, 175.85], [636.44, 175.85]]}, {\"category\": \"car\", \"corners_3d\": [[549.45, 194.95], [547.14, 195.7], [479.81, 195.68], [484.35, 194.93], [549.45, 170.7], [547.14, 170.63], [479.81, 170.63], [484.35, 170.7]]}, {\"category\": \"car\", \"corners_3d\": [[756.55, 203.79], [763.56, 205.13], [688.73, 205.2], [684.83, 203.85], [756.55, 176.0], [763.56, 176.13], [688.73, 176.14], [684.83, 176.0]]}, {\"category\": \"car\", \"corners_3d\": [[518.4, 212.48], [509.84, 214.65], [410.03, 214.12], [423.6, 212.0], [518.4, 174.61], [509.84, 174.71], [410.03, 174.68], [423.6, 174.59]]}, {\"category\": \"car\", \"corners_3d\": [[812.49, 211.59], [826.84, 213.78], [737.24, 214.15], [727.61, 211.92], [812.49, 173.58], [826.84, 173.62], [737.24, 173.63], [727.61, 173.59]]}]\n```", - "options": null, - "id": 689 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006206", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[556.12, 203.7], [527.79, 203.74], [532.26, 201.1], [558.17, 201.06], [556.12, 175.8], [527.79, 175.81], [532.26, 175.55], [558.17, 175.55]]}]\n```", - "options": null, - "id": 690 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006209", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[561.38, 209.46], [531.21, 209.46], [538.5, 206.03], [565.84, 206.03], [561.38, 182.96], [531.21, 182.96], [538.5, 182.01], [565.84, 182.01]]}]\n```", - "options": null, - "id": 691 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006234", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[256.78, 249.6], [237.73, 249.52], [265.56, 244.18], [283.32, 244.25], [256.78, 191.94], [237.73, 191.92], [265.56, 190.59], [283.32, 190.61]]}]\n```", - "options": null, - "id": 692 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006234", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[471.16, 275.53], [388.09, 275.31], [442.48, 251.63], [506.41, 251.76], [471.16, 196.32], [388.09, 196.27], [442.48, 190.86], [506.41, 190.89]]}]\n```", - "options": null, - "id": 693 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006234", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[471.16, 275.53], [388.09, 275.31], [442.48, 251.63], [506.41, 251.76], [471.16, 196.32], [388.09, 196.27], [442.48, 190.86], [506.41, 190.89]]}, {\"category\": \"cyclist\", \"corners_3d\": [[256.78, 249.6], [237.73, 249.52], [265.56, 244.18], [283.32, 244.25], [256.78, 191.94], [237.73, 191.92], [265.56, 190.59], [283.32, 190.61]]}]\n```", - "options": null, - "id": 694 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006314", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-367.11, 546.26], [-645.91, 541.87], [-7.87, 357.43], [134.53, 358.52], [-367.11, 277.06], [-645.91, 275.83], [-7.87, 224.36], [134.53, 224.67]]}, {\"category\": \"car\", \"corners_3d\": [[188.31, 335.54], [65.08, 335.83], [225.26, 287.19], [311.55, 287.04], [188.31, 217.43], [65.08, 217.51], [225.26, 204.18], [311.55, 204.14]]}, {\"category\": \"car\", \"corners_3d\": [[353.61, 272.25], [271.15, 272.26], [344.71, 250.61], [409.21, 250.61], [353.61, 202.31], [271.15, 202.31], [344.71, 195.9], [409.21, 195.9]]}, {\"category\": \"car\", \"corners_3d\": [[1056.93, 240.31], [1002.48, 232.08], [1223.71, 232.11], [1308.95, 240.35], [1056.93, 136.88], [1002.48, 141.27], [1223.71, 141.26], [1308.95, 136.86]]}, {\"category\": \"car\", \"corners_3d\": [[1006.8, 227.02], [962.53, 221.21], [1088.73, 221.02], [1147.91, 226.79], [1006.8, 138.26], [962.53, 141.97], [1088.73, 142.09], [1147.91, 138.41]]}, {\"category\": \"car\", \"corners_3d\": [[920.63, 221.12], [890.26, 216.63], [1026.31, 216.45], [1070.41, 220.9], [920.63, 147.15], [890.26, 149.54], [1026.31, 149.63], [1070.41, 147.27]]}, {\"category\": \"car\", \"corners_3d\": [[849.72, 214.66], [830.33, 211.4], [957.79, 211.32], [987.88, 214.57], [849.72, 155.51], [830.33, 156.86], [957.79, 156.9], [987.88, 155.55]]}, {\"category\": \"car\", \"corners_3d\": [[442.35, 238.37], [390.18, 238.14], [430.45, 227.83], [474.47, 227.99], [442.35, 191.28], [390.18, 191.22], [430.45, 188.32], [474.47, 188.36]]}, {\"category\": \"car\", \"corners_3d\": [[693.45, 229.02], [750.22, 228.84], [790.16, 242.08], [720.08, 242.35], [693.45, 176.42], [750.22, 176.41], [790.16, 177.25], [720.08, 177.27]]}, {\"category\": \"car\", \"corners_3d\": [[676.77, 218.09], [720.63, 218.09], [741.11, 226.48], [689.13, 226.48], [676.77, 174.24], [720.63, 174.24], [741.11, 174.5], [689.13, 174.5]]}]\n```", - "options": null, - "id": 695 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006327", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[443.24, 298.42], [321.34, 297.8], [416.51, 259.35], [501.1, 259.65], [443.24, 192.66], [321.34, 192.56], [416.51, 186.5], [501.1, 186.54]]}]\n```", - "options": null, - "id": 696 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006352", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[333.87, 281.82], [226.02, 281.67], [324.43, 254.22], [405.15, 254.31], [333.87, 191.4], [226.02, 191.37], [324.43, 186.7], [405.15, 186.71]]}]\n```", - "options": null, - "id": 697 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006352", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[747.85, 200.36], [758.76, 200.33], [769.51, 201.81], [758.02, 201.85], [747.85, 162.2], [758.76, 162.21], [769.51, 161.63], [758.02, 161.62]]}]\n```", - "options": null, - "id": 698 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006352", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[333.87, 281.82], [226.02, 281.67], [324.43, 254.22], [405.15, 254.31], [333.87, 191.4], [226.02, 191.37], [324.43, 186.7], [405.15, 186.71]]}, {\"category\": \"cyclist\", \"corners_3d\": [[747.85, 200.36], [758.76, 200.33], [769.51, 201.81], [758.02, 201.85], [747.85, 162.2], [758.76, 162.21], [769.51, 161.63], [758.02, 161.62]]}]\n```", - "options": null, - "id": 699 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006359", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[608.87, 198.68], [586.42, 198.69], [587.07, 196.61], [607.72, 196.6], [608.87, 178.78], [586.42, 178.78], [587.07, 178.3], [607.72, 178.3]]}]\n```", - "options": null, - "id": 700 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006391", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[64.73, 333.49], [188.97, 310.97], [382.5, 335.14], [270.41, 367.16], [64.73, 212.37], [188.97, 206.83], [382.5, 212.78], [270.41, 220.66]]}, {\"category\": \"car\", \"corners_3d\": [[799.11, 236.55], [875.73, 239.06], [867.31, 256.98], [771.36, 252.96], [799.11, 170.78], [875.73, 170.7], [867.31, 170.12], [771.36, 170.25]]}, {\"category\": \"car\", \"corners_3d\": [[1060.32, 240.57], [1157.43, 242.94], [1281.59, 269.8], [1143.26, 265.32], [1060.32, 160.79], [1157.43, 160.37], [1281.59, 155.59], [1143.26, 156.39]]}, {\"category\": \"car\", \"corners_3d\": [[-124.58, 321.57], [6.61, 303.43], [167.77, 326.22], [36.82, 351.87], [-124.58, 232.2], [6.61, 224.96], [167.77, 234.05], [36.82, 244.29]]}, {\"category\": \"car\", \"corners_3d\": [[1018.59, 213.73], [1093.1, 215.21], [1137.29, 226.27], [1042.25, 223.94], [1018.59, 161.31], [1093.1, 160.89], [1137.29, 157.77], [1042.25, 158.42]]}, {\"category\": \"car\", \"corners_3d\": [[986.69, 200.18], [1034.89, 200.92], [1044.05, 205.34], [988.2, 204.36], [986.69, 156.68], [1034.89, 156.25], [1044.05, 153.63], [988.2, 154.21]]}]\n```", - "options": null, - "id": 701 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006429", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[375.82, 399.16], [167.77, 397.9], [351.62, 307.1], [476.05, 307.55], [375.82, 184.65], [167.77, 184.59], [351.62, 179.85], [476.05, 179.88]]}, {\"category\": \"car\", \"corners_3d\": [[796.26, 350.34], [945.38, 348.88], [1265.76, 503.76], [988.38, 508.94], [796.26, 188.12], [945.38, 188.0], [1265.76, 201.32], [988.38, 201.76]]}, {\"category\": \"car\", \"corners_3d\": [[692.94, 276.24], [783.92, 275.72], [862.2, 315.46], [736.38, 316.45], [692.94, 187.77], [783.92, 187.7], [862.2, 193.43], [736.38, 193.57]]}, {\"category\": \"car\", \"corners_3d\": [[430.79, 258.4], [506.79, 257.88], [487.76, 284.85], [387.29, 285.76], [430.79, 180.89], [506.79, 180.84], [487.76, 183.37], [387.29, 183.46]]}, {\"category\": \"car\", \"corners_3d\": [[513.56, 250.75], [440.56, 250.76], [469.33, 237.45], [529.86, 237.44], [513.56, 186.27], [440.56, 186.27], [469.33, 183.98], [529.86, 183.98]]}, {\"category\": \"car\", \"corners_3d\": [[667.65, 238.57], [725.28, 238.15], [760.6, 251.48], [691.39, 252.09], [667.65, 180.02], [725.28, 179.97], [760.6, 181.43], [691.39, 181.49]]}, {\"category\": \"car\", \"corners_3d\": [[643.77, 222.85], [690.87, 222.72], [710.96, 231.86], [655.26, 232.04], [643.77, 179.06], [690.87, 179.04], [710.96, 180.18], [655.26, 180.2]]}, {\"category\": \"car\", \"corners_3d\": [[551.88, 215.03], [516.11, 214.97], [528.07, 210.57], [560.12, 210.62], [551.88, 178.87], [516.11, 178.86], [528.07, 178.24], [560.12, 178.24]]}, {\"category\": \"car\", \"corners_3d\": [[571.61, 207.84], [538.82, 207.83], [546.84, 204.2], [576.23, 204.22], [571.61, 170.08], [538.82, 170.08], [546.84, 170.37], [576.23, 170.37]]}, {\"category\": \"car\", \"corners_3d\": [[585.34, 204.19], [556.69, 204.13], [564.78, 201.27], [590.82, 201.32], [585.34, 173.02], [556.69, 173.02], [564.78, 173.0], [590.82, 173.0]]}]\n```", - "options": null, - "id": 702 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006554", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[707.56, 185.11], [725.51, 185.14], [728.97, 185.98], [709.78, 185.95], [707.56, 169.07], [725.51, 169.06], [728.97, 168.8], [709.78, 168.81]]}]\n```", - "options": null, - "id": 703 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006554", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[216.95, 227.45], [186.81, 234.57], [-269.24, 242.28], [-177.5, 233.4], [216.95, 118.84], [186.81, 111.8], [-269.24, 104.17], [-177.5, 112.95]]}]\n```", - "options": null, - "id": 704 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006554", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[707.56, 185.11], [725.51, 185.14], [728.97, 185.98], [709.78, 185.95], [707.56, 169.07], [725.51, 169.06], [728.97, 168.8], [709.78, 168.81]]}, {\"category\": \"truck\", \"corners_3d\": [[216.95, 227.45], [186.81, 234.57], [-269.24, 242.28], [-177.5, 233.4], [216.95, 118.84], [186.81, 111.8], [-269.24, 104.17], [-177.5, 112.95]]}]\n```", - "options": null, - "id": 705 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006584", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[445.76, 261.36], [364.39, 261.67], [411.45, 242.92], [475.56, 242.73], [445.76, 181.74], [364.39, 181.77], [411.45, 179.89], [475.56, 179.87]]}]\n```", - "options": null, - "id": 706 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006607", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[519.44, 241.42], [460.38, 242.14], [470.4, 229.35], [518.57, 228.87], [519.44, 189.99], [460.38, 190.17], [470.4, 186.97], [518.57, 186.85]]}]\n```", - "options": null, - "id": 707 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006617", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[567.7, 200.32], [527.01, 200.4], [530.11, 198.11], [566.12, 198.05], [567.7, 162.13], [527.01, 162.06], [530.11, 164.18], [566.12, 164.24]]}, {\"category\": \"car\", \"corners_3d\": [[578.07, 193.45], [547.17, 193.46], [552.17, 192.14], [579.93, 192.14], [578.07, 166.71], [547.17, 166.7], [552.17, 168.1], [579.93, 168.11]]}, {\"category\": \"car\", \"corners_3d\": [[581.64, 182.76], [597.91, 182.75], [609.43, 182.89], [592.19, 182.91], [581.64, 167.94], [597.91, 168.02], [609.43, 167.22], [592.19, 167.14]]}]\n```", - "options": null, - "id": 708 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006617", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1262.26, 223.06], [1212.76, 219.51], [1375.75, 220.28], [1441.47, 223.98], [1262.26, 156.29], [1212.76, 158.31], [1375.75, 157.88], [1441.47, 155.77]]}]\n```", - "options": null, - "id": 709 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006617", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[996.89, 220.24], [1005.87, 219.6], [1050.21, 221.47], [1041.55, 222.17], [996.89, 154.74], [1005.87, 155.15], [1050.21, 153.94], [1041.55, 153.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1000.89, 220.32], [1013.03, 219.84], [1055.72, 222.03], [1043.45, 222.57], [1000.89, 157.07], [1013.03, 157.35], [1055.72, 156.07], [1043.45, 155.75]]}]\n```", - "options": null, - "id": 710 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006617", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[567.7, 200.32], [527.01, 200.4], [530.11, 198.11], [566.12, 198.05], [567.7, 162.13], [527.01, 162.06], [530.11, 164.18], [566.12, 164.24]]}, {\"category\": \"van\", \"corners_3d\": [[1262.26, 223.06], [1212.76, 219.51], [1375.75, 220.28], [1441.47, 223.98], [1262.26, 156.29], [1212.76, 158.31], [1375.75, 157.88], [1441.47, 155.77]]}, {\"category\": \"car\", \"corners_3d\": [[578.07, 193.45], [547.17, 193.46], [552.17, 192.14], [579.93, 192.14], [578.07, 166.71], [547.17, 166.7], [552.17, 168.1], [579.93, 168.11]]}, {\"category\": \"car\", \"corners_3d\": [[581.64, 182.76], [597.91, 182.75], [609.43, 182.89], [592.19, 182.91], [581.64, 167.94], [597.91, 168.02], [609.43, 167.22], [592.19, 167.14]]}]\n```", - "options": null, - "id": 711 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006617", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[567.7, 200.32], [527.01, 200.4], [530.11, 198.11], [566.12, 198.05], [567.7, 162.13], [527.01, 162.06], [530.11, 164.18], [566.12, 164.24]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[996.89, 220.24], [1005.87, 219.6], [1050.21, 221.47], [1041.55, 222.17], [996.89, 154.74], [1005.87, 155.15], [1050.21, 153.94], [1041.55, 153.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1000.89, 220.32], [1013.03, 219.84], [1055.72, 222.03], [1043.45, 222.57], [1000.89, 157.07], [1013.03, 157.35], [1055.72, 156.07], [1043.45, 155.75]]}, {\"category\": \"car\", \"corners_3d\": [[578.07, 193.45], [547.17, 193.46], [552.17, 192.14], [579.93, 192.14], [578.07, 166.71], [547.17, 166.7], [552.17, 168.1], [579.93, 168.11]]}, {\"category\": \"car\", \"corners_3d\": [[581.64, 182.76], [597.91, 182.75], [609.43, 182.89], [592.19, 182.91], [581.64, 167.94], [597.91, 168.02], [609.43, 167.22], [592.19, 167.14]]}]\n```", - "options": null, - "id": 712 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006617", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[996.89, 220.24], [1005.87, 219.6], [1050.21, 221.47], [1041.55, 222.17], [996.89, 154.74], [1005.87, 155.15], [1050.21, 153.94], [1041.55, 153.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1000.89, 220.32], [1013.03, 219.84], [1055.72, 222.03], [1043.45, 222.57], [1000.89, 157.07], [1013.03, 157.35], [1055.72, 156.07], [1043.45, 155.75]]}, {\"category\": \"van\", \"corners_3d\": [[1262.26, 223.06], [1212.76, 219.51], [1375.75, 220.28], [1441.47, 223.98], [1262.26, 156.29], [1212.76, 158.31], [1375.75, 157.88], [1441.47, 155.77]]}]\n```", - "options": null, - "id": 713 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006617", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[567.7, 200.32], [527.01, 200.4], [530.11, 198.11], [566.12, 198.05], [567.7, 162.13], [527.01, 162.06], [530.11, 164.18], [566.12, 164.24]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[996.89, 220.24], [1005.87, 219.6], [1050.21, 221.47], [1041.55, 222.17], [996.89, 154.74], [1005.87, 155.15], [1050.21, 153.94], [1041.55, 153.48]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[1000.89, 220.32], [1013.03, 219.84], [1055.72, 222.03], [1043.45, 222.57], [1000.89, 157.07], [1013.03, 157.35], [1055.72, 156.07], [1043.45, 155.75]]}, {\"category\": \"van\", \"corners_3d\": [[1262.26, 223.06], [1212.76, 219.51], [1375.75, 220.28], [1441.47, 223.98], [1262.26, 156.29], [1212.76, 158.31], [1375.75, 157.88], [1441.47, 155.77]]}, {\"category\": \"car\", \"corners_3d\": [[578.07, 193.45], [547.17, 193.46], [552.17, 192.14], [579.93, 192.14], [578.07, 166.71], [547.17, 166.7], [552.17, 168.1], [579.93, 168.11]]}, {\"category\": \"car\", \"corners_3d\": [[581.64, 182.76], [597.91, 182.75], [609.43, 182.89], [592.19, 182.91], [581.64, 167.94], [597.91, 168.02], [609.43, 167.22], [592.19, 167.14]]}]\n```", - "options": null, - "id": 714 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006659", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1060.99, 342.21], [1097.5, 342.13], [1204.67, 378.81], [1160.31, 378.92], [1060.99, 195.27], [1097.5, 195.26], [1204.67, 200.11], [1160.31, 200.13]]}]\n```", - "options": null, - "id": 715 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006659", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[799.98, 334.7], [945.53, 334.72], [1223.03, 468.65], [957.01, 468.56], [799.98, 202.69], [945.53, 202.69], [1223.03, 227.38], [957.01, 227.37]]}, {\"category\": \"car\", \"corners_3d\": [[560.4, 228.32], [508.26, 228.36], [523.18, 219.5], [567.0, 219.47], [560.4, 180.28], [508.26, 180.29], [523.18, 179.1], [567.0, 179.1]]}, {\"category\": \"car\", \"corners_3d\": [[657.92, 211.25], [689.26, 211.21], [701.88, 216.39], [666.31, 216.43], [657.92, 178.39], [689.26, 178.39], [701.88, 179.13], [666.31, 179.14]]}]\n```", - "options": null, - "id": 716 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006659", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[716.19, 266.91], [806.49, 266.69], [905.61, 310.81], [773.04, 311.3], [716.19, 171.74], [806.49, 171.74], [905.61, 171.22], [773.04, 171.22]]}, {\"category\": \"van\", \"corners_3d\": [[670.06, 216.23], [715.3, 216.15], [740.29, 224.68], [686.17, 224.79], [670.06, 164.58], [715.3, 164.6], [740.29, 162.97], [686.17, 162.95]]}]\n```", - "options": null, - "id": 717 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006659", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1060.99, 342.21], [1097.5, 342.13], [1204.67, 378.81], [1160.31, 378.92], [1060.99, 195.27], [1097.5, 195.26], [1204.67, 200.11], [1160.31, 200.13]]}, {\"category\": \"car\", \"corners_3d\": [[799.98, 334.7], [945.53, 334.72], [1223.03, 468.65], [957.01, 468.56], [799.98, 202.69], [945.53, 202.69], [1223.03, 227.38], [957.01, 227.37]]}, {\"category\": \"car\", \"corners_3d\": [[560.4, 228.32], [508.26, 228.36], [523.18, 219.5], [567.0, 219.47], [560.4, 180.28], [508.26, 180.29], [523.18, 179.1], [567.0, 179.1]]}, {\"category\": \"car\", \"corners_3d\": [[657.92, 211.25], [689.26, 211.21], [701.88, 216.39], [666.31, 216.43], [657.92, 178.39], [689.26, 178.39], [701.88, 179.13], [666.31, 179.14]]}]\n```", - "options": null, - "id": 718 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006659", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1060.99, 342.21], [1097.5, 342.13], [1204.67, 378.81], [1160.31, 378.92], [1060.99, 195.27], [1097.5, 195.26], [1204.67, 200.11], [1160.31, 200.13]]}, {\"category\": \"van\", \"corners_3d\": [[716.19, 266.91], [806.49, 266.69], [905.61, 310.81], [773.04, 311.3], [716.19, 171.74], [806.49, 171.74], [905.61, 171.22], [773.04, 171.22]]}, {\"category\": \"van\", \"corners_3d\": [[670.06, 216.23], [715.3, 216.15], [740.29, 224.68], [686.17, 224.79], [670.06, 164.58], [715.3, 164.6], [740.29, 162.97], [686.17, 162.95]]}]\n```", - "options": null, - "id": 719 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006659", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[799.98, 334.7], [945.53, 334.72], [1223.03, 468.65], [957.01, 468.56], [799.98, 202.69], [945.53, 202.69], [1223.03, 227.38], [957.01, 227.37]]}, {\"category\": \"van\", \"corners_3d\": [[716.19, 266.91], [806.49, 266.69], [905.61, 310.81], [773.04, 311.3], [716.19, 171.74], [806.49, 171.74], [905.61, 171.22], [773.04, 171.22]]}, {\"category\": \"car\", \"corners_3d\": [[560.4, 228.32], [508.26, 228.36], [523.18, 219.5], [567.0, 219.47], [560.4, 180.28], [508.26, 180.29], [523.18, 179.1], [567.0, 179.1]]}, {\"category\": \"van\", \"corners_3d\": [[670.06, 216.23], [715.3, 216.15], [740.29, 224.68], [686.17, 224.79], [670.06, 164.58], [715.3, 164.6], [740.29, 162.97], [686.17, 162.95]]}, {\"category\": \"car\", \"corners_3d\": [[657.92, 211.25], [689.26, 211.21], [701.88, 216.39], [666.31, 216.43], [657.92, 178.39], [689.26, 178.39], [701.88, 179.13], [666.31, 179.14]]}]\n```", - "options": null, - "id": 720 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006659", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (cyclist, car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"cyclist\", \"corners_3d\": [[1060.99, 342.21], [1097.5, 342.13], [1204.67, 378.81], [1160.31, 378.92], [1060.99, 195.27], [1097.5, 195.26], [1204.67, 200.11], [1160.31, 200.13]]}, {\"category\": \"car\", \"corners_3d\": [[799.98, 334.7], [945.53, 334.72], [1223.03, 468.65], [957.01, 468.56], [799.98, 202.69], [945.53, 202.69], [1223.03, 227.38], [957.01, 227.37]]}, {\"category\": \"van\", \"corners_3d\": [[716.19, 266.91], [806.49, 266.69], [905.61, 310.81], [773.04, 311.3], [716.19, 171.74], [806.49, 171.74], [905.61, 171.22], [773.04, 171.22]]}, {\"category\": \"car\", \"corners_3d\": [[560.4, 228.32], [508.26, 228.36], [523.18, 219.5], [567.0, 219.47], [560.4, 180.28], [508.26, 180.29], [523.18, 179.1], [567.0, 179.1]]}, {\"category\": \"van\", \"corners_3d\": [[670.06, 216.23], [715.3, 216.15], [740.29, 224.68], [686.17, 224.79], [670.06, 164.58], [715.3, 164.6], [740.29, 162.97], [686.17, 162.95]]}, {\"category\": \"car\", \"corners_3d\": [[657.92, 211.25], [689.26, 211.21], [701.88, 216.39], [666.31, 216.43], [657.92, 178.39], [689.26, 178.39], [701.88, 179.13], [666.31, 179.14]]}]\n```", - "options": null, - "id": 721 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006691", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[521.0, 243.52], [449.32, 243.45], [481.08, 230.02], [539.13, 230.06], [521.0, 186.67], [449.32, 186.66], [481.08, 184.03], [539.13, 184.04]]}]\n```", - "options": null, - "id": 722 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006733", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[1189.95, 187.42], [1233.47, 188.47], [1111.6, 188.56], [1076.12, 187.49], [1189.95, 138.99], [1233.47, 136.54], [1111.6, 136.36], [1076.12, 138.83]]}]\n```", - "options": null, - "id": 723 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006733", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-11.73, 596.67], [-442.22, 604.26], [167.63, 348.72], [340.55, 347.44], [-11.73, 266.45], [-442.22, 268.13], [167.63, 211.69], [340.55, 211.41]]}, {\"category\": \"car\", \"corners_3d\": [[-422.74, 521.69], [-734.61, 523.29], [-88.36, 353.82], [71.52, 353.4], [-422.74, 264.51], [-734.61, 264.93], [-88.36, 220.4], [71.52, 220.29]]}, {\"category\": \"car\", \"corners_3d\": [[654.37, 216.9], [694.95, 216.88], [710.02, 224.09], [662.8, 224.12], [654.37, 178.96], [694.95, 178.96], [710.02, 179.96], [662.8, 179.97]]}, {\"category\": \"car\", \"corners_3d\": [[405.58, 295.81], [293.37, 295.82], [368.75, 266.45], [454.15, 266.45], [405.58, 200.53], [293.37, 200.54], [368.75, 193.93], [454.15, 193.92]]}, {\"category\": \"car\", \"corners_3d\": [[823.41, 221.82], [806.31, 217.88], [965.09, 217.9], [996.09, 221.84], [823.41, 167.28], [806.31, 167.72], [965.09, 167.72], [996.09, 167.27]]}, {\"category\": \"car\", \"corners_3d\": [[801.17, 216.36], [786.92, 213.11], [916.15, 213.12], [940.86, 216.38], [801.17, 154.79], [786.92, 156.14], [916.15, 156.14], [940.86, 154.78]]}, {\"category\": \"car\", \"corners_3d\": [[451.95, 218.57], [486.83, 218.57], [470.9, 224.48], [431.51, 224.47], [451.95, 183.25], [486.83, 183.25], [470.9, 184.59], [431.51, 184.59]]}, {\"category\": \"car\", \"corners_3d\": [[397.38, 250.08], [337.26, 250.02], [380.85, 237.97], [431.6, 238.01], [397.38, 195.13], [337.26, 195.11], [380.85, 191.63], [431.6, 191.65]]}, {\"category\": \"car\", \"corners_3d\": [[753.21, 206.36], [745.75, 204.51], [849.0, 204.57], [862.5, 206.42], [753.21, 165.29], [745.75, 165.7], [849.0, 165.69], [862.5, 165.27]]}, {\"category\": \"car\", \"corners_3d\": [[507.98, 212.29], [478.35, 212.3], [488.98, 208.91], [516.06, 208.89], [507.98, 185.35], [478.35, 185.35], [488.98, 184.28], [516.06, 184.27]]}]\n```", - "options": null, - "id": 724 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006733", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-11.73, 596.67], [-442.22, 604.26], [167.63, 348.72], [340.55, 347.44], [-11.73, 266.45], [-442.22, 268.13], [167.63, 211.69], [340.55, 211.41]]}, {\"category\": \"car\", \"corners_3d\": [[-422.74, 521.69], [-734.61, 523.29], [-88.36, 353.82], [71.52, 353.4], [-422.74, 264.51], [-734.61, 264.93], [-88.36, 220.4], [71.52, 220.29]]}, {\"category\": \"car\", \"corners_3d\": [[654.37, 216.9], [694.95, 216.88], [710.02, 224.09], [662.8, 224.12], [654.37, 178.96], [694.95, 178.96], [710.02, 179.96], [662.8, 179.97]]}, {\"category\": \"car\", \"corners_3d\": [[405.58, 295.81], [293.37, 295.82], [368.75, 266.45], [454.15, 266.45], [405.58, 200.53], [293.37, 200.54], [368.75, 193.93], [454.15, 193.92]]}, {\"category\": \"car\", \"corners_3d\": [[823.41, 221.82], [806.31, 217.88], [965.09, 217.9], [996.09, 221.84], [823.41, 167.28], [806.31, 167.72], [965.09, 167.72], [996.09, 167.27]]}, {\"category\": \"car\", \"corners_3d\": [[801.17, 216.36], [786.92, 213.11], [916.15, 213.12], [940.86, 216.38], [801.17, 154.79], [786.92, 156.14], [916.15, 156.14], [940.86, 154.78]]}, {\"category\": \"car\", \"corners_3d\": [[451.95, 218.57], [486.83, 218.57], [470.9, 224.48], [431.51, 224.47], [451.95, 183.25], [486.83, 183.25], [470.9, 184.59], [431.51, 184.59]]}, {\"category\": \"car\", \"corners_3d\": [[397.38, 250.08], [337.26, 250.02], [380.85, 237.97], [431.6, 238.01], [397.38, 195.13], [337.26, 195.11], [380.85, 191.63], [431.6, 191.65]]}, {\"category\": \"car\", \"corners_3d\": [[753.21, 206.36], [745.75, 204.51], [849.0, 204.57], [862.5, 206.42], [753.21, 165.29], [745.75, 165.7], [849.0, 165.69], [862.5, 165.27]]}, {\"category\": \"car\", \"corners_3d\": [[507.98, 212.29], [478.35, 212.3], [488.98, 208.91], [516.06, 208.89], [507.98, 185.35], [478.35, 185.35], [488.98, 184.28], [516.06, 184.27]]}, {\"category\": \"van\", \"corners_3d\": [[1189.95, 187.42], [1233.47, 188.47], [1111.6, 188.56], [1076.12, 187.49], [1189.95, 138.99], [1233.47, 136.54], [1111.6, 136.36], [1076.12, 138.83]]}]\n```", - "options": null, - "id": 725 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006735", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[388.43, 289.27], [490.05, 289.95], [398.52, 361.9], [235.9, 360.14], [388.43, 185.71], [490.05, 185.78], [398.52, 193.72], [235.9, 193.53]]}, {\"category\": \"car\", \"corners_3d\": [[540.73, 253.99], [476.79, 253.78], [503.33, 239.83], [556.27, 239.97], [540.73, 192.75], [476.79, 192.69], [503.33, 189.27], [556.27, 189.31]]}, {\"category\": \"car\", \"corners_3d\": [[594.88, 212.04], [569.8, 212.01], [574.05, 208.91], [597.15, 208.93], [594.88, 190.55], [569.8, 190.54], [574.05, 189.14], [597.15, 189.15]]}, {\"category\": \"car\", \"corners_3d\": [[1569.03, 278.96], [1668.49, 296.76], [1170.75, 271.88], [1147.74, 260.18], [1569.03, 111.15], [1668.49, 100.8], [1170.75, 115.27], [1147.74, 122.08]]}]\n```", - "options": null, - "id": 726 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006743", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[868.5, 187.37], [896.14, 187.38], [926.74, 189.02], [895.96, 189.0], [868.5, 161.26], [896.14, 161.25], [926.74, 159.95], [895.96, 159.96]]}, {\"category\": \"van\", \"corners_3d\": [[944.95, 191.87], [986.46, 191.89], [1045.42, 194.99], [997.07, 194.96], [944.95, 154.4], [986.46, 154.38], [1045.42, 151.37], [997.07, 151.4]]}]\n```", - "options": null, - "id": 727 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006743", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[1072.53, 193.09], [1133.02, 193.09], [1275.27, 198.59], [1198.31, 198.59], [1072.53, 132.1], [1133.02, 132.1], [1275.27, 121.01], [1198.31, 121.01]]}]\n```", - "options": null, - "id": 728 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006743", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[567.18, 220.2], [611.85, 220.18], [613.06, 226.58], [562.34, 226.62], [567.18, 176.88], [611.85, 176.88], [613.06, 177.42], [562.34, 177.43]]}, {\"category\": \"car\", \"corners_3d\": [[549.12, 195.65], [533.63, 195.65], [536.95, 194.64], [551.76, 194.64], [549.12, 179.68], [533.63, 179.68], [536.95, 179.38], [551.76, 179.38]]}, {\"category\": \"car\", \"corners_3d\": [[841.7, 188.32], [861.32, 188.31], [884.1, 189.6], [862.86, 189.61], [841.7, 170.34], [861.32, 170.34], [884.1, 170.13], [862.86, 170.13]]}]\n```", - "options": null, - "id": 729 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006743", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[868.5, 187.37], [896.14, 187.38], [926.74, 189.02], [895.96, 189.0], [868.5, 161.26], [896.14, 161.25], [926.74, 159.95], [895.96, 159.96]]}, {\"category\": \"van\", \"corners_3d\": [[944.95, 191.87], [986.46, 191.89], [1045.42, 194.99], [997.07, 194.96], [944.95, 154.4], [986.46, 154.38], [1045.42, 151.37], [997.07, 151.4]]}, {\"category\": \"truck\", \"corners_3d\": [[1072.53, 193.09], [1133.02, 193.09], [1275.27, 198.59], [1198.31, 198.59], [1072.53, 132.1], [1133.02, 132.1], [1275.27, 121.01], [1198.31, 121.01]]}]\n```", - "options": null, - "id": 730 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006743", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[567.18, 220.2], [611.85, 220.18], [613.06, 226.58], [562.34, 226.62], [567.18, 176.88], [611.85, 176.88], [613.06, 177.42], [562.34, 177.43]]}, {\"category\": \"car\", \"corners_3d\": [[549.12, 195.65], [533.63, 195.65], [536.95, 194.64], [551.76, 194.64], [549.12, 179.68], [533.63, 179.68], [536.95, 179.38], [551.76, 179.38]]}, {\"category\": \"car\", \"corners_3d\": [[841.7, 188.32], [861.32, 188.31], [884.1, 189.6], [862.86, 189.61], [841.7, 170.34], [861.32, 170.34], [884.1, 170.13], [862.86, 170.13]]}, {\"category\": \"van\", \"corners_3d\": [[868.5, 187.37], [896.14, 187.38], [926.74, 189.02], [895.96, 189.0], [868.5, 161.26], [896.14, 161.25], [926.74, 159.95], [895.96, 159.96]]}, {\"category\": \"van\", \"corners_3d\": [[944.95, 191.87], [986.46, 191.89], [1045.42, 194.99], [997.07, 194.96], [944.95, 154.4], [986.46, 154.38], [1045.42, 151.37], [997.07, 151.4]]}]\n```", - "options": null, - "id": 731 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006743", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[567.18, 220.2], [611.85, 220.18], [613.06, 226.58], [562.34, 226.62], [567.18, 176.88], [611.85, 176.88], [613.06, 177.42], [562.34, 177.43]]}, {\"category\": \"car\", \"corners_3d\": [[549.12, 195.65], [533.63, 195.65], [536.95, 194.64], [551.76, 194.64], [549.12, 179.68], [533.63, 179.68], [536.95, 179.38], [551.76, 179.38]]}, {\"category\": \"car\", \"corners_3d\": [[841.7, 188.32], [861.32, 188.31], [884.1, 189.6], [862.86, 189.61], [841.7, 170.34], [861.32, 170.34], [884.1, 170.13], [862.86, 170.13]]}, {\"category\": \"truck\", \"corners_3d\": [[1072.53, 193.09], [1133.02, 193.09], [1275.27, 198.59], [1198.31, 198.59], [1072.53, 132.1], [1133.02, 132.1], [1275.27, 121.01], [1198.31, 121.01]]}]\n```", - "options": null, - "id": 732 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006743", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, truck, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[567.18, 220.2], [611.85, 220.18], [613.06, 226.58], [562.34, 226.62], [567.18, 176.88], [611.85, 176.88], [613.06, 177.42], [562.34, 177.43]]}, {\"category\": \"car\", \"corners_3d\": [[549.12, 195.65], [533.63, 195.65], [536.95, 194.64], [551.76, 194.64], [549.12, 179.68], [533.63, 179.68], [536.95, 179.38], [551.76, 179.38]]}, {\"category\": \"car\", \"corners_3d\": [[841.7, 188.32], [861.32, 188.31], [884.1, 189.6], [862.86, 189.61], [841.7, 170.34], [861.32, 170.34], [884.1, 170.13], [862.86, 170.13]]}, {\"category\": \"van\", \"corners_3d\": [[868.5, 187.37], [896.14, 187.38], [926.74, 189.02], [895.96, 189.0], [868.5, 161.26], [896.14, 161.25], [926.74, 159.95], [895.96, 159.96]]}, {\"category\": \"van\", \"corners_3d\": [[944.95, 191.87], [986.46, 191.89], [1045.42, 194.99], [997.07, 194.96], [944.95, 154.4], [986.46, 154.38], [1045.42, 151.37], [997.07, 151.4]]}, {\"category\": \"truck\", \"corners_3d\": [[1072.53, 193.09], [1133.02, 193.09], [1275.27, 198.59], [1198.31, 198.59], [1072.53, 132.1], [1133.02, 132.1], [1275.27, 121.01], [1198.31, 121.01]]}]\n```", - "options": null, - "id": 733 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006774", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[535.68, 233.25], [479.28, 233.35], [497.6, 223.74], [545.02, 223.67], [535.68, 174.05], [479.28, 174.05], [497.6, 173.86], [545.02, 173.86]]}, {\"category\": \"car\", \"corners_3d\": [[566.22, 206.04], [543.49, 206.02], [549.32, 203.6], [570.4, 203.62], [566.22, 180.74], [543.49, 180.73], [549.32, 180.16], [570.4, 180.16]]}]\n```", - "options": null, - "id": 734 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006784", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[572.7, 219.14], [543.33, 219.35], [542.09, 215.24], [568.88, 215.07], [572.7, 189.0], [543.33, 189.07], [542.09, 187.64], [568.88, 187.58]]}]\n```", - "options": null, - "id": 735 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006799", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[380.48, 311.05], [360.39, 322.41], [264.25, 322.22], [292.0, 310.88], [380.48, 161.18], [360.39, 159.5], [264.25, 159.53], [292.0, 161.21]]}]\n```", - "options": null, - "id": 736 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006857", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[122.82, 324.84], [-6.08, 324.86], [169.76, 281.4], [261.79, 281.39], [122.82, 206.43], [-6.08, 206.43], [169.76, 196.83], [261.79, 196.83]]}, {\"category\": \"car\", \"corners_3d\": [[618.59, 195.93], [640.1, 195.94], [641.74, 197.6], [618.68, 197.59], [618.59, 175.62], [640.1, 175.62], [641.74, 175.82], [618.68, 175.82]]}, {\"category\": \"car\", \"corners_3d\": [[508.35, 207.34], [482.52, 207.12], [502.16, 204.43], [526.06, 204.62], [508.35, 180.83], [482.52, 180.78], [502.16, 180.16], [526.06, 180.2]]}, {\"category\": \"car\", \"corners_3d\": [[529.06, 202.48], [505.52, 202.32], [522.52, 200.04], [544.32, 200.18], [529.06, 172.85], [505.52, 172.85], [522.52, 172.85], [544.32, 172.85]]}, {\"category\": \"car\", \"corners_3d\": [[658.21, 195.38], [680.3, 195.42], [682.32, 197.12], [658.57, 197.07], [658.21, 174.56], [680.3, 174.57], [682.32, 174.69], [658.57, 174.69]]}]\n```", - "options": null, - "id": 737 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006872", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[758.81, 292.66], [920.73, 293.67], [1430.54, 490.15], [979.42, 482.59], [758.81, 94.86], [920.73, 94.07], [1430.54, -59.11], [979.42, -53.22]]}]\n```", - "options": null, - "id": 738 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006872", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[288.68, 270.24], [327.04, 266.39], [342.23, 268.43], [303.61, 272.48], [288.68, 156.65], [327.04, 157.73], [342.23, 157.16], [303.61, 156.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[613.98, 239.16], [582.04, 236.91], [608.65, 234.68], [640.28, 236.75], [613.98, 173.64], [582.04, 173.94], [608.65, 174.25], [640.28, 173.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[519.51, 241.91], [496.38, 243.64], [477.48, 241.71], [500.43, 240.08], [519.51, 169.18], [496.38, 168.83], [477.48, 169.22], [500.43, 169.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[572.76, 227.91], [541.48, 228.15], [540.28, 226.17], [570.24, 225.95], [572.76, 166.92], [541.48, 166.84], [540.28, 167.47], [570.24, 167.54]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[617.1, 215.81], [644.05, 215.88], [644.33, 217.2], [616.34, 217.13], [617.1, 165.16], [644.05, 165.13], [644.33, 164.5], [616.34, 164.53]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[622.93, 206.39], [607.78, 206.36], [608.86, 205.73], [623.62, 205.76], [622.93, 167.21], [607.78, 167.23], [608.86, 167.59], [623.62, 167.57]]}]\n```", - "options": null, - "id": 739 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006872", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[288.68, 270.24], [327.04, 266.39], [342.23, 268.43], [303.61, 272.48], [288.68, 156.65], [327.04, 157.73], [342.23, 157.16], [303.61, 156.02]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[613.98, 239.16], [582.04, 236.91], [608.65, 234.68], [640.28, 236.75], [613.98, 173.64], [582.04, 173.94], [608.65, 174.25], [640.28, 173.97]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[519.51, 241.91], [496.38, 243.64], [477.48, 241.71], [500.43, 240.08], [519.51, 169.18], [496.38, 168.83], [477.48, 169.22], [500.43, 169.55]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[572.76, 227.91], [541.48, 228.15], [540.28, 226.17], [570.24, 225.95], [572.76, 166.92], [541.48, 166.84], [540.28, 167.47], [570.24, 167.54]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[617.1, 215.81], [644.05, 215.88], [644.33, 217.2], [616.34, 217.13], [617.1, 165.16], [644.05, 165.13], [644.33, 164.5], [616.34, 164.53]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[622.93, 206.39], [607.78, 206.36], [608.86, 205.73], [623.62, 205.76], [622.93, 167.21], [607.78, 167.23], [608.86, 167.59], [623.62, 167.57]]}, {\"category\": \"van\", \"corners_3d\": [[758.81, 292.66], [920.73, 293.67], [1430.54, 490.15], [979.42, 482.59], [758.81, 94.86], [920.73, 94.07], [1430.54, -59.11], [979.42, -53.22]]}]\n```", - "options": null, - "id": 740 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006890", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[543.21, 222.73], [500.27, 222.73], [514.22, 216.33], [551.65, 216.33], [543.21, 185.01], [500.27, 185.01], [514.22, 183.45], [551.65, 183.45]]}, {\"category\": \"car\", \"corners_3d\": [[572.22, 200.88], [549.22, 200.87], [553.51, 199.08], [575.04, 199.08], [572.22, 181.38], [549.22, 181.38], [553.51, 180.83], [575.04, 180.83]]}]\n```", - "options": null, - "id": 741 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006905", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[507.56, 274.76], [414.6, 276.45], [434.29, 256.06], [508.93, 254.97], [507.56, 189.39], [414.6, 189.67], [434.29, 186.36], [508.93, 186.18]]}, {\"category\": \"car\", \"corners_3d\": [[519.11, 244.07], [466.11, 244.08], [489.05, 232.64], [533.54, 232.64], [519.11, 190.05], [466.11, 190.05], [489.05, 187.29], [533.54, 187.29]]}]\n```", - "options": null, - "id": 742 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "006941", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[326.24, 311.92], [306.2, 318.64], [226.06, 315.93], [249.56, 309.46], [326.24, 169.9], [306.2, 169.36], [226.06, 169.57], [249.56, 170.1]]}]\n```", - "options": null, - "id": 743 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007024", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[528.58, 237.3], [462.67, 237.42], [486.95, 225.54], [540.71, 225.45], [528.58, 181.14], [462.67, 181.16], [486.95, 179.63], [540.71, 179.62]]}, {\"category\": \"car\", \"corners_3d\": [[588.03, 181.24], [566.8, 181.25], [568.64, 180.69], [588.46, 180.69], [588.03, 158.9], [566.8, 158.89], [568.64, 159.82], [588.46, 159.83]]}]\n```", - "options": null, - "id": 744 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007057", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[253.98, 207.3], [285.91, 207.22], [253.06, 211.23], [217.29, 211.34], [253.98, 181.28], [285.91, 181.26], [253.06, 182.24], [217.29, 182.27]]}, {\"category\": \"car\", \"corners_3d\": [[201.56, 206.75], [233.98, 206.66], [195.15, 210.6], [158.82, 210.71], [201.56, 178.74], [233.98, 178.73], [195.15, 179.41], [158.82, 179.43]]}, {\"category\": \"car\", \"corners_3d\": [[85.88, 228.88], [35.55, 229.03], [106.91, 221.67], [150.49, 221.55], [85.88, 184.18], [35.55, 184.21], [106.91, 182.72], [150.49, 182.7]]}]\n```", - "options": null, - "id": 745 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007069", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[663.63, 212.07], [699.83, 212.09], [712.16, 217.95], [670.55, 217.92], [663.63, 181.17], [699.83, 181.18], [712.16, 182.42], [670.55, 182.41]]}, {\"category\": \"car\", \"corners_3d\": [[517.08, 217.83], [510.2, 221.18], [377.72, 221.18], [393.78, 217.83], [517.08, 177.84], [510.2, 178.21], [377.72, 178.21], [393.78, 177.84]]}, {\"category\": \"car\", \"corners_3d\": [[480.45, 231.37], [465.34, 236.73], [303.82, 236.01], [332.17, 230.77], [480.45, 180.92], [465.34, 181.65], [303.82, 181.55], [332.17, 180.83]]}, {\"category\": \"car\", \"corners_3d\": [[472.24, 242.19], [451.04, 249.72], [274.13, 248.19], [311.96, 240.94], [472.24, 181.9], [451.04, 182.89], [274.13, 182.69], [311.96, 181.74]]}, {\"category\": \"car\", \"corners_3d\": [[448.78, 252.03], [425.11, 260.07], [237.19, 257.73], [277.18, 250.1], [448.78, 183.37], [425.11, 184.43], [237.19, 184.12], [277.18, 183.11]]}, {\"category\": \"car\", \"corners_3d\": [[397.91, 262.48], [368.89, 274.77], [151.76, 274.77], [206.95, 262.48], [397.91, 181.32], [368.89, 182.48], [151.76, 182.48], [206.95, 181.32]]}, {\"category\": \"car\", \"corners_3d\": [[-256.37, 351.87], [-99.77, 319.54], [191.67, 319.63], [99.23, 352.01], [-256.37, 190.72], [-99.77, 187.49], [191.67, 187.5], [99.23, 190.74]]}]\n```", - "options": null, - "id": 746 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007069", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[527.69, 213.93], [524.12, 216.7], [388.3, 217.03], [400.55, 214.22], [527.69, 163.46], [524.12, 162.83], [388.3, 162.75], [400.55, 163.4]]}]\n```", - "options": null, - "id": 747 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007069", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"truck\", \"corners_3d\": [[640.84, 199.51], [676.69, 199.51], [686.62, 203.49], [645.43, 203.49], [640.84, 165.11], [676.69, 165.11], [686.62, 163.95], [645.43, 163.95]]}]\n```", - "options": null, - "id": 748 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007069", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[663.63, 212.07], [699.83, 212.09], [712.16, 217.95], [670.55, 217.92], [663.63, 181.17], [699.83, 181.18], [712.16, 182.42], [670.55, 182.41]]}, {\"category\": \"van\", \"corners_3d\": [[527.69, 213.93], [524.12, 216.7], [388.3, 217.03], [400.55, 214.22], [527.69, 163.46], [524.12, 162.83], [388.3, 162.75], [400.55, 163.4]]}, {\"category\": \"car\", \"corners_3d\": [[517.08, 217.83], [510.2, 221.18], [377.72, 221.18], [393.78, 217.83], [517.08, 177.84], [510.2, 178.21], [377.72, 178.21], [393.78, 177.84]]}, {\"category\": \"car\", \"corners_3d\": [[480.45, 231.37], [465.34, 236.73], [303.82, 236.01], [332.17, 230.77], [480.45, 180.92], [465.34, 181.65], [303.82, 181.55], [332.17, 180.83]]}, {\"category\": \"car\", \"corners_3d\": [[472.24, 242.19], [451.04, 249.72], [274.13, 248.19], [311.96, 240.94], [472.24, 181.9], [451.04, 182.89], [274.13, 182.69], [311.96, 181.74]]}, {\"category\": \"car\", \"corners_3d\": [[448.78, 252.03], [425.11, 260.07], [237.19, 257.73], [277.18, 250.1], [448.78, 183.37], [425.11, 184.43], [237.19, 184.12], [277.18, 183.11]]}, {\"category\": \"car\", \"corners_3d\": [[397.91, 262.48], [368.89, 274.77], [151.76, 274.77], [206.95, 262.48], [397.91, 181.32], [368.89, 182.48], [151.76, 182.48], [206.95, 181.32]]}, {\"category\": \"car\", \"corners_3d\": [[-256.37, 351.87], [-99.77, 319.54], [191.67, 319.63], [99.23, 352.01], [-256.37, 190.72], [-99.77, 187.49], [191.67, 187.5], [99.23, 190.74]]}]\n```", - "options": null, - "id": 749 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007069", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[663.63, 212.07], [699.83, 212.09], [712.16, 217.95], [670.55, 217.92], [663.63, 181.17], [699.83, 181.18], [712.16, 182.42], [670.55, 182.41]]}, {\"category\": \"car\", \"corners_3d\": [[517.08, 217.83], [510.2, 221.18], [377.72, 221.18], [393.78, 217.83], [517.08, 177.84], [510.2, 178.21], [377.72, 178.21], [393.78, 177.84]]}, {\"category\": \"car\", \"corners_3d\": [[480.45, 231.37], [465.34, 236.73], [303.82, 236.01], [332.17, 230.77], [480.45, 180.92], [465.34, 181.65], [303.82, 181.55], [332.17, 180.83]]}, {\"category\": \"car\", \"corners_3d\": [[472.24, 242.19], [451.04, 249.72], [274.13, 248.19], [311.96, 240.94], [472.24, 181.9], [451.04, 182.89], [274.13, 182.69], [311.96, 181.74]]}, {\"category\": \"car\", \"corners_3d\": [[448.78, 252.03], [425.11, 260.07], [237.19, 257.73], [277.18, 250.1], [448.78, 183.37], [425.11, 184.43], [237.19, 184.12], [277.18, 183.11]]}, {\"category\": \"car\", \"corners_3d\": [[397.91, 262.48], [368.89, 274.77], [151.76, 274.77], [206.95, 262.48], [397.91, 181.32], [368.89, 182.48], [151.76, 182.48], [206.95, 181.32]]}, {\"category\": \"car\", \"corners_3d\": [[-256.37, 351.87], [-99.77, 319.54], [191.67, 319.63], [99.23, 352.01], [-256.37, 190.72], [-99.77, 187.49], [191.67, 187.5], [99.23, 190.74]]}, {\"category\": \"truck\", \"corners_3d\": [[640.84, 199.51], [676.69, 199.51], [686.62, 203.49], [645.43, 203.49], [640.84, 165.11], [676.69, 165.11], [686.62, 163.95], [645.43, 163.95]]}]\n```", - "options": null, - "id": 750 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007069", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[527.69, 213.93], [524.12, 216.7], [388.3, 217.03], [400.55, 214.22], [527.69, 163.46], [524.12, 162.83], [388.3, 162.75], [400.55, 163.4]]}, {\"category\": \"truck\", \"corners_3d\": [[640.84, 199.51], [676.69, 199.51], [686.62, 203.49], [645.43, 203.49], [640.84, 165.11], [676.69, 165.11], [686.62, 163.95], [645.43, 163.95]]}]\n```", - "options": null, - "id": 751 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007069", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van, truck). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[663.63, 212.07], [699.83, 212.09], [712.16, 217.95], [670.55, 217.92], [663.63, 181.17], [699.83, 181.18], [712.16, 182.42], [670.55, 182.41]]}, {\"category\": \"van\", \"corners_3d\": [[527.69, 213.93], [524.12, 216.7], [388.3, 217.03], [400.55, 214.22], [527.69, 163.46], [524.12, 162.83], [388.3, 162.75], [400.55, 163.4]]}, {\"category\": \"car\", \"corners_3d\": [[517.08, 217.83], [510.2, 221.18], [377.72, 221.18], [393.78, 217.83], [517.08, 177.84], [510.2, 178.21], [377.72, 178.21], [393.78, 177.84]]}, {\"category\": \"car\", \"corners_3d\": [[480.45, 231.37], [465.34, 236.73], [303.82, 236.01], [332.17, 230.77], [480.45, 180.92], [465.34, 181.65], [303.82, 181.55], [332.17, 180.83]]}, {\"category\": \"car\", \"corners_3d\": [[472.24, 242.19], [451.04, 249.72], [274.13, 248.19], [311.96, 240.94], [472.24, 181.9], [451.04, 182.89], [274.13, 182.69], [311.96, 181.74]]}, {\"category\": \"car\", \"corners_3d\": [[448.78, 252.03], [425.11, 260.07], [237.19, 257.73], [277.18, 250.1], [448.78, 183.37], [425.11, 184.43], [237.19, 184.12], [277.18, 183.11]]}, {\"category\": \"car\", \"corners_3d\": [[397.91, 262.48], [368.89, 274.77], [151.76, 274.77], [206.95, 262.48], [397.91, 181.32], [368.89, 182.48], [151.76, 182.48], [206.95, 181.32]]}, {\"category\": \"car\", \"corners_3d\": [[-256.37, 351.87], [-99.77, 319.54], [191.67, 319.63], [99.23, 352.01], [-256.37, 190.72], [-99.77, 187.49], [191.67, 187.5], [99.23, 190.74]]}, {\"category\": \"truck\", \"corners_3d\": [[640.84, 199.51], [676.69, 199.51], [686.62, 203.49], [645.43, 203.49], [640.84, 165.11], [676.69, 165.11], [686.62, 163.95], [645.43, 163.95]]}]\n```", - "options": null, - "id": 752 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007075", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[1072.78, 295.27], [1099.42, 301.9], [996.87, 301.72], [975.85, 295.12], [1072.78, 143.4], [1099.42, 141.26], [996.87, 141.32], [975.85, 143.45]]}]\n```", - "options": null, - "id": 753 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007086", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[579.63, 196.77], [559.78, 196.76], [563.74, 195.27], [582.35, 195.28], [579.63, 180.13], [559.78, 180.13], [563.74, 179.67], [582.35, 179.68]]}]\n```", - "options": null, - "id": 754 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007131", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[646.77, 194.22], [667.26, 194.24], [669.73, 196.11], [647.45, 196.08], [646.77, 169.65], [667.26, 169.65], [669.73, 169.37], [647.45, 169.37]]}]\n```", - "options": null, - "id": 755 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007131", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1037.51, 367.73], [855.19, 366.8], [766.93, 292.94], [879.34, 293.29], [1037.51, 181.99], [855.19, 181.94], [766.93, 178.48], [879.34, 178.5]]}, {\"category\": \"car\", \"corners_3d\": [[493.21, 272.71], [399.29, 272.33], [458.88, 246.71], [528.71, 246.92], [493.21, 190.43], [399.29, 190.36], [458.88, 185.85], [528.71, 185.89]]}, {\"category\": \"car\", \"corners_3d\": [[701.76, 249.44], [769.17, 249.16], [820.25, 269.91], [734.68, 270.36], [701.76, 186.53], [769.17, 186.48], [820.25, 190.19], [734.68, 190.27]]}, {\"category\": \"car\", \"corners_3d\": [[524.64, 240.27], [461.76, 240.1], [488.97, 229.26], [541.75, 229.38], [524.64, 181.5], [461.76, 181.48], [488.97, 180.09], [541.75, 180.11]]}, {\"category\": \"car\", \"corners_3d\": [[754.66, 242.25], [689.91, 242.19], [675.37, 228.52], [727.33, 228.55], [754.66, 185.46], [689.91, 185.45], [675.37, 182.96], [727.33, 182.97]]}, {\"category\": \"car\", \"corners_3d\": [[560.05, 218.47], [521.25, 218.52], [528.82, 213.8], [563.61, 213.76], [560.05, 181.05], [521.25, 181.06], [528.82, 180.21], [563.61, 180.2]]}, {\"category\": \"car\", \"corners_3d\": [[704.46, 212.06], [669.16, 212.01], [664.78, 207.55], [696.05, 207.6], [704.46, 181.49], [669.16, 181.47], [664.78, 180.49], [696.05, 180.5]]}, {\"category\": \"car\", \"corners_3d\": [[653.67, 197.33], [673.61, 197.35], [676.63, 199.12], [655.25, 199.1], [653.67, 178.08], [673.61, 178.08], [676.63, 178.46], [655.25, 178.45]]}]\n```", - "options": null, - "id": 756 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007131", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[1037.51, 367.73], [855.19, 366.8], [766.93, 292.94], [879.34, 293.29], [1037.51, 181.99], [855.19, 181.94], [766.93, 178.48], [879.34, 178.5]]}, {\"category\": \"car\", \"corners_3d\": [[493.21, 272.71], [399.29, 272.33], [458.88, 246.71], [528.71, 246.92], [493.21, 190.43], [399.29, 190.36], [458.88, 185.85], [528.71, 185.89]]}, {\"category\": \"car\", \"corners_3d\": [[701.76, 249.44], [769.17, 249.16], [820.25, 269.91], [734.68, 270.36], [701.76, 186.53], [769.17, 186.48], [820.25, 190.19], [734.68, 190.27]]}, {\"category\": \"car\", \"corners_3d\": [[524.64, 240.27], [461.76, 240.1], [488.97, 229.26], [541.75, 229.38], [524.64, 181.5], [461.76, 181.48], [488.97, 180.09], [541.75, 180.11]]}, {\"category\": \"car\", \"corners_3d\": [[754.66, 242.25], [689.91, 242.19], [675.37, 228.52], [727.33, 228.55], [754.66, 185.46], [689.91, 185.45], [675.37, 182.96], [727.33, 182.97]]}, {\"category\": \"car\", \"corners_3d\": [[560.05, 218.47], [521.25, 218.52], [528.82, 213.8], [563.61, 213.76], [560.05, 181.05], [521.25, 181.06], [528.82, 180.21], [563.61, 180.2]]}, {\"category\": \"car\", \"corners_3d\": [[704.46, 212.06], [669.16, 212.01], [664.78, 207.55], [696.05, 207.6], [704.46, 181.49], [669.16, 181.47], [664.78, 180.49], [696.05, 180.5]]}, {\"category\": \"car\", \"corners_3d\": [[653.67, 197.33], [673.61, 197.35], [676.63, 199.12], [655.25, 199.1], [653.67, 178.08], [673.61, 178.08], [676.63, 178.46], [655.25, 178.45]]}, {\"category\": \"van\", \"corners_3d\": [[646.77, 194.22], [667.26, 194.24], [669.73, 196.11], [647.45, 196.08], [646.77, 169.65], [667.26, 169.65], [669.73, 169.37], [647.45, 169.37]]}]\n```", - "options": null, - "id": 757 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007149", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[942.19, 303.36], [916.63, 293.97], [954.29, 293.77], [982.72, 303.13], [942.19, 129.63], [916.63, 132.74], [954.29, 132.81], [982.72, 129.71]]}]\n```", - "options": null, - "id": 758 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007149", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[481.25, 243.29], [551.16, 243.44], [528.91, 264.83], [437.92, 264.59], [481.25, 188.81], [551.16, 188.85], [528.91, 193.69], [437.92, 193.64]]}, {\"category\": \"car\", \"corners_3d\": [[672.57, 206.2], [708.1, 206.16], [725.98, 211.46], [684.81, 211.51], [672.57, 172.16], [708.1, 172.16], [725.98, 172.05], [684.81, 172.05]]}, {\"category\": \"car\", \"corners_3d\": [[532.74, 219.24], [574.17, 219.32], [563.59, 227.85], [514.6, 227.73], [532.74, 180.71], [574.17, 180.73], [563.59, 182.17], [514.6, 182.15]]}, {\"category\": \"car\", \"corners_3d\": [[707.45, 235.21], [767.45, 235.32], [803.82, 251.22], [728.5, 251.05], [707.45, 175.47], [767.45, 175.48], [803.82, 176.14], [728.5, 176.14]]}, {\"category\": \"car\", \"corners_3d\": [[657.87, 194.89], [682.59, 194.96], [683.16, 196.69], [656.51, 196.61], [657.87, 174.2], [682.59, 174.2], [683.16, 174.31], [656.51, 174.3]]}]\n```", - "options": null, - "id": 759 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007149", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[723.27, 245.17], [815.67, 245.08], [983.95, 302.21], [818.67, 302.49], [723.27, 117.35], [815.67, 117.41], [983.95, 73.56], [818.67, 73.35]]}]\n```", - "options": null, - "id": 760 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007149", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[481.25, 243.29], [551.16, 243.44], [528.91, 264.83], [437.92, 264.59], [481.25, 188.81], [551.16, 188.85], [528.91, 193.69], [437.92, 193.64]]}, {\"category\": \"car\", \"corners_3d\": [[672.57, 206.2], [708.1, 206.16], [725.98, 211.46], [684.81, 211.51], [672.57, 172.16], [708.1, 172.16], [725.98, 172.05], [684.81, 172.05]]}, {\"category\": \"car\", \"corners_3d\": [[532.74, 219.24], [574.17, 219.32], [563.59, 227.85], [514.6, 227.73], [532.74, 180.71], [574.17, 180.73], [563.59, 182.17], [514.6, 182.15]]}, {\"category\": \"car\", \"corners_3d\": [[707.45, 235.21], [767.45, 235.32], [803.82, 251.22], [728.5, 251.05], [707.45, 175.47], [767.45, 175.48], [803.82, 176.14], [728.5, 176.14]]}, {\"category\": \"car\", \"corners_3d\": [[657.87, 194.89], [682.59, 194.96], [683.16, 196.69], [656.51, 196.61], [657.87, 174.2], [682.59, 174.2], [683.16, 174.31], [656.51, 174.3]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[942.19, 303.36], [916.63, 293.97], [954.29, 293.77], [982.72, 303.13], [942.19, 129.63], [916.63, 132.74], [954.29, 132.81], [982.72, 129.71]]}]\n```", - "options": null, - "id": 761 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007149", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[723.27, 245.17], [815.67, 245.08], [983.95, 302.21], [818.67, 302.49], [723.27, 117.35], [815.67, 117.41], [983.95, 73.56], [818.67, 73.35]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[942.19, 303.36], [916.63, 293.97], [954.29, 293.77], [982.72, 303.13], [942.19, 129.63], [916.63, 132.74], [954.29, 132.81], [982.72, 129.71]]}]\n```", - "options": null, - "id": 762 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007149", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[481.25, 243.29], [551.16, 243.44], [528.91, 264.83], [437.92, 264.59], [481.25, 188.81], [551.16, 188.85], [528.91, 193.69], [437.92, 193.64]]}, {\"category\": \"van\", \"corners_3d\": [[723.27, 245.17], [815.67, 245.08], [983.95, 302.21], [818.67, 302.49], [723.27, 117.35], [815.67, 117.41], [983.95, 73.56], [818.67, 73.35]]}, {\"category\": \"car\", \"corners_3d\": [[672.57, 206.2], [708.1, 206.16], [725.98, 211.46], [684.81, 211.51], [672.57, 172.16], [708.1, 172.16], [725.98, 172.05], [684.81, 172.05]]}, {\"category\": \"car\", \"corners_3d\": [[532.74, 219.24], [574.17, 219.32], [563.59, 227.85], [514.6, 227.73], [532.74, 180.71], [574.17, 180.73], [563.59, 182.17], [514.6, 182.15]]}, {\"category\": \"car\", \"corners_3d\": [[707.45, 235.21], [767.45, 235.32], [803.82, 251.22], [728.5, 251.05], [707.45, 175.47], [767.45, 175.48], [803.82, 176.14], [728.5, 176.14]]}, {\"category\": \"car\", \"corners_3d\": [[657.87, 194.89], [682.59, 194.96], [683.16, 196.69], [656.51, 196.61], [657.87, 174.2], [682.59, 174.2], [683.16, 174.31], [656.51, 174.3]]}]\n```", - "options": null, - "id": 763 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007149", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian, car, van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[481.25, 243.29], [551.16, 243.44], [528.91, 264.83], [437.92, 264.59], [481.25, 188.81], [551.16, 188.85], [528.91, 193.69], [437.92, 193.64]]}, {\"category\": \"van\", \"corners_3d\": [[723.27, 245.17], [815.67, 245.08], [983.95, 302.21], [818.67, 302.49], [723.27, 117.35], [815.67, 117.41], [983.95, 73.56], [818.67, 73.35]]}, {\"category\": \"car\", \"corners_3d\": [[672.57, 206.2], [708.1, 206.16], [725.98, 211.46], [684.81, 211.51], [672.57, 172.16], [708.1, 172.16], [725.98, 172.05], [684.81, 172.05]]}, {\"category\": \"car\", \"corners_3d\": [[532.74, 219.24], [574.17, 219.32], [563.59, 227.85], [514.6, 227.73], [532.74, 180.71], [574.17, 180.73], [563.59, 182.17], [514.6, 182.15]]}, {\"category\": \"car\", \"corners_3d\": [[707.45, 235.21], [767.45, 235.32], [803.82, 251.22], [728.5, 251.05], [707.45, 175.47], [767.45, 175.48], [803.82, 176.14], [728.5, 176.14]]}, {\"category\": \"car\", \"corners_3d\": [[657.87, 194.89], [682.59, 194.96], [683.16, 196.69], [656.51, 196.61], [657.87, 174.2], [682.59, 174.2], [683.16, 174.31], [656.51, 174.3]]}, {\"category\": \"pedestrian\", \"corners_3d\": [[942.19, 303.36], [916.63, 293.97], [954.29, 293.77], [982.72, 303.13], [942.19, 129.63], [916.63, 132.74], [954.29, 132.81], [982.72, 129.71]]}]\n```", - "options": null, - "id": 764 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007150", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (pedestrian). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (707.05, 707.05), (cx, cy) = (604.08, 180.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"pedestrian\", \"corners_3d\": [[957.65, 287.79], [941.52, 283.01], [1006.33, 282.9], [1025.46, 287.67], [957.65, 153.52], [941.52, 154.72], [1006.33, 154.75], [1025.46, 153.55]]}]\n```", - "options": null, - "id": 765 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007282", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[975.4, 352.49], [786.39, 344.68], [769.54, 281.91], [887.07, 285.01], [975.4, 185.57], [786.39, 185.02], [769.54, 180.58], [887.07, 180.8]]}]\n```", - "options": null, - "id": 766 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007298", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[478.44, 228.81], [432.16, 228.45], [464.0, 221.34], [504.52, 221.62], [478.44, 184.98], [432.16, 184.91], [464.0, 183.37], [504.52, 183.43]]}, {\"category\": \"car\", \"corners_3d\": [[589.75, 199.34], [569.9, 199.17], [582.46, 197.58], [601.18, 197.73], [589.75, 179.08], [569.9, 179.04], [582.46, 178.67], [601.18, 178.71]]}]\n```", - "options": null, - "id": 767 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007311", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"van\", \"corners_3d\": [[567.55, 203.95], [538.68, 203.88], [547.34, 201.36], [573.87, 201.41], [567.55, 174.74], [538.68, 174.74], [547.34, 174.59], [573.87, 174.59]]}]\n```", - "options": null, - "id": 768 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007311", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-41.8, 682.59], [-540.07, 682.87], [187.67, 359.86], [370.28, 359.82], [-41.8, 163.59], [-540.07, 163.58], [187.67, 169.45], [370.28, 169.45]]}, {\"category\": \"car\", \"corners_3d\": [[417.3, 261.19], [501.82, 261.41], [458.15, 292.9], [343.82, 292.51], [417.3, 190.1], [501.82, 190.14], [458.15, 196.29], [343.82, 196.21]]}, {\"category\": \"car\", \"corners_3d\": [[683.21, 224.05], [640.83, 224.17], [633.51, 217.99], [670.81, 217.89], [683.21, 184.72], [640.83, 184.75], [633.51, 183.32], [670.81, 183.29]]}]\n```", - "options": null, - "id": 769 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007311", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (van, car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[-41.8, 682.59], [-540.07, 682.87], [187.67, 359.86], [370.28, 359.82], [-41.8, 163.59], [-540.07, 163.58], [187.67, 169.45], [370.28, 169.45]]}, {\"category\": \"car\", \"corners_3d\": [[417.3, 261.19], [501.82, 261.41], [458.15, 292.9], [343.82, 292.51], [417.3, 190.1], [501.82, 190.14], [458.15, 196.29], [343.82, 196.21]]}, {\"category\": \"car\", \"corners_3d\": [[683.21, 224.05], [640.83, 224.17], [633.51, 217.99], [670.81, 217.89], [683.21, 184.72], [640.83, 184.75], [633.51, 183.32], [670.81, 183.29]]}, {\"category\": \"van\", \"corners_3d\": [[567.55, 203.95], [538.68, 203.88], [547.34, 201.36], [573.87, 201.41], [567.55, 174.74], [538.68, 174.74], [547.34, 174.59], [573.87, 174.59]]}]\n```", - "options": null, - "id": 770 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007384", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (721.54, 721.54), (cx, cy) = (609.56, 172.85). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[863.1, 346.93], [1027.5, 352.28], [1314.07, 515.16], [992.81, 496.18], [863.1, 183.64], [1027.5, 183.97], [1314.07, 194.06], [992.81, 192.89]]}, {\"category\": \"car\", \"corners_3d\": [[599.36, 225.77], [548.72, 225.51], [565.66, 217.47], [608.6, 217.65], [599.36, 180.63], [548.72, 180.59], [565.66, 179.41], [608.6, 179.43]]}, {\"category\": \"car\", \"corners_3d\": [[762.1, 214.36], [727.14, 214.08], [725.43, 209.8], [756.73, 210.02], [762.1, 178.96], [727.14, 178.92], [725.43, 178.29], [756.73, 178.33]]}, {\"category\": \"car\", \"corners_3d\": [[639.94, 204.5], [610.2, 204.32], [620.05, 201.23], [646.91, 201.38], [639.94, 178.42], [610.2, 178.39], [620.05, 177.85], [646.91, 177.87]]}]\n```", - "options": null, - "id": 771 - }, - { - "dataset": "Mono3DRefer", - "scene_name": "007399", - "question_type": "mobj_3dBbox_json", - "question": "I hope you can understand all the target objects in this scene within the following categories: (car). Please ignore the objects that are too small or too far away to be recognized. Additionally, known the camera intrinsic parameters: (fx, fy) = (718.34, 718.34), (cx, cy) = (600.39, 181.51). You'll uderstand the meaning from the phrases I have provided, and answer the 3D attributes of all target objects with JSON format including: \n- corners_3d: the projected points of 8 corners of the 3D bounding box in the image coordinate (list of [x, y] with unit in pixels).\n\nNoting the answered JSON format should be like:\n```JSON\n[{\"category\": category, \"corners_3d\": [[x1, y1], [x2, y2], ..., [x8, y8]]}, ...]\n```", - "ground_truth": "```JSON\n[{\"category\": \"car\", \"corners_3d\": [[417.93, 245.5], [340.79, 245.16], [394.21, 233.63], [457.55, 233.86], [417.93, 182.77], [340.79, 182.76], [394.21, 182.54], [457.55, 182.54]]}]\n```", - "options": null, - "id": 772 - } -] \ No newline at end of file