| { | |
| "data": [ | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nhermit crab\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 0, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 283, | |
| 196, | |
| 272, | |
| 236 | |
| ] | |
| } | |
| ], | |
| "id": 0, | |
| "images": [ | |
| "localization_Orca_Simple_hermit_crab_Hermit crab8.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nhermit crab\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\natlantic blue crab\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 0, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 107, | |
| 209, | |
| 468, | |
| 188 | |
| ] | |
| } | |
| ], | |
| "id": 1, | |
| "images": [ | |
| "localization_Orca_Simple_atlantic_blue_crab_005.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\natlantic blue crab\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\npagophilus groenlandicus\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 0, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 6, | |
| 192, | |
| 459, | |
| 131 | |
| ] | |
| } | |
| ], | |
| "id": 2, | |
| "images": [ | |
| "localization_Orca_Simple_pagophilus_groenlandicus_Pagophilus groenlandicus18.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\npagophilus groenlandicus\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\npygoscelis adeliae\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 0, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 168, | |
| 113, | |
| 263, | |
| 285 | |
| ] | |
| } | |
| ], | |
| "id": 3, | |
| "images": [ | |
| "localization_Orca_Simple_pygoscelis_adeliae_Pygoscelis adeliae7.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\npygoscelis adeliae\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nspotted sharpnose puffer\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 0, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 157, | |
| 139, | |
| 353, | |
| 203 | |
| ] | |
| } | |
| ], | |
| "id": 4, | |
| "images": [ | |
| "localization_Orca_Simple_spotted_sharpnose_puffer_Spotted sharpnose puffer7.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nspotted sharpnose puffer\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\norca\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 0, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 99, | |
| 77, | |
| 423, | |
| 301 | |
| ] | |
| } | |
| ], | |
| "id": 5, | |
| "images": [ | |
| "localization_Orca_Simple_orca_(killer_whale)_002.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\norca\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nbox jellyfish\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 0, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 133, | |
| 8, | |
| 430, | |
| 476 | |
| ] | |
| } | |
| ], | |
| "id": 6, | |
| "images": [ | |
| "localization_Orca_Simple_box_jellyfish_Google_0069.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nbox jellyfish\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nemperor angelfish\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 0, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 127, | |
| 34, | |
| 427, | |
| 369 | |
| ] | |
| } | |
| ], | |
| "id": 7, | |
| "images": [ | |
| "localization_Orca_Simple_emperor_angelfish_022.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nemperor angelfish\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nleatherback sea turtle\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 0, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 8, | |
| 61, | |
| 635, | |
| 421 | |
| ] | |
| } | |
| ], | |
| "id": 8, | |
| "images": [ | |
| "localization_Orca_Simple_leatherback_sea_turtle_007.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nleatherback sea turtle\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\npike\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 0, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 22, | |
| 285, | |
| 1035, | |
| 416 | |
| ] | |
| } | |
| ], | |
| "id": 9, | |
| "images": [ | |
| "localization_Orca_Simple_pike_Google_0014.png" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\npike\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 1090, | |
| "image_height": 719 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\narbacia punctulata\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 0, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 41, | |
| 37, | |
| 557, | |
| 422 | |
| ] | |
| } | |
| ], | |
| "id": 10, | |
| "images": [ | |
| "localization_Orca_Simple_arbacia_punctulata_Arbacia punctulata34.jpeg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\narbacia punctulata\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\npagrus major\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 0, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 45, | |
| 112, | |
| 564, | |
| 245 | |
| ] | |
| } | |
| ], | |
| "id": 11, | |
| "images": [ | |
| "localization_Orca_Simple_pagrus_major_Image_25.JPG" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\npagrus major\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nzebra shark\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 0, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 93, | |
| 44, | |
| 550, | |
| 329 | |
| ] | |
| } | |
| ], | |
| "id": 12, | |
| "images": [ | |
| "localization_Orca_Simple_zebra_shark_juveniles_(rarely_seen_with_stripes)_Zebra shark juveniles (rarely seen with stripes)10.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nzebra shark\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ngreen turtle\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 0, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 89, | |
| 79, | |
| 464, | |
| 366 | |
| ] | |
| } | |
| ], | |
| "id": 13, | |
| "images": [ | |
| "localization_Orca_Simple_green_turtle_Green turtle9.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ngreen turtle\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\njellyfish\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 0, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 155, | |
| 109, | |
| 289, | |
| 302 | |
| ] | |
| } | |
| ], | |
| "id": 14, | |
| "images": [ | |
| "localization_Orca_Simple_jellyfish_Jellyfish9.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\njellyfish\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nbohadschia marmorata\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 0, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 49, | |
| 47, | |
| 552, | |
| 385 | |
| ] | |
| } | |
| ], | |
| "id": 15, | |
| "images": [ | |
| "localization_Orca_Simple_bohadschia_marmorata_Bohadschia marmorata21.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nbohadschia marmorata\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\npaddlefish\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 0, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 98, | |
| 159, | |
| 511, | |
| 238 | |
| ] | |
| } | |
| ], | |
| "id": 16, | |
| "images": [ | |
| "localization_Orca_Simple_paddlefish_Google_0027.jpeg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\npaddlefish\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nstomolophus meleagris\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 0, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 86, | |
| 64, | |
| 470, | |
| 387 | |
| ] | |
| } | |
| ], | |
| "id": 17, | |
| "images": [ | |
| "localization_Orca_Simple_stomolophus_meleagris_Image_10.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nstomolophus meleagris\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\npaddlefish\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 0, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 38, | |
| 136, | |
| 606, | |
| 171 | |
| ] | |
| } | |
| ], | |
| "id": 18, | |
| "images": [ | |
| "localization_Orca_Simple_paddlefish_Google_0072.jpeg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\npaddlefish\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ndevil ray\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 0, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 158, | |
| 114, | |
| 94, | |
| 208 | |
| ] | |
| } | |
| ], | |
| "id": 19, | |
| "images": [ | |
| "localization_Orca_Simple_devil_ray_Google_0003.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ndevil ray\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis image appears to show two sharks swimming in the ocean. The sharks are swimming near some coral and there are green plants in the background. The sharks are large and have black and white markings on their bodies. The dorsal fin has a white tip. The one we are focusin is located at the left.\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 2, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 6, | |
| 198, | |
| 318, | |
| 160 | |
| ] | |
| } | |
| ], | |
| "id": 20, | |
| "images": [ | |
| "localization_Orca_Medium_black_rock_sharks_(unspecified_species)_Black Rock sharks (unspecified species)3.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis image appears to show two sharks swimming in the ocean. The sharks are swimming near some coral and there are green plants in the background. The sharks are large and have black and white markings on their bodies. The dorsal fin has a white tip. The one we are focusin is located at the left.\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThe image pictures a white oval shell. There is a yellow line almost forming a circle on top of the shell. The shell is smooth and shiny. It has a small circular whole on the bottom. It is sitting on a sandy bottom and there are five other similar shells around. The one we are focusing is located at the bottom.\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 2, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 317, | |
| 266, | |
| 156, | |
| 166 | |
| ] | |
| } | |
| ], | |
| "id": 21, | |
| "images": [ | |
| "localization_Orca_Medium_cypraea_annulus_Cypraea annulus3.JPG" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThe image pictures a white oval shell. There is a yellow line almost forming a circle on top of the shell. The shell is smooth and shiny. It has a small circular whole on the bottom. It is sitting on a sandy bottom and there are five other similar shells around. The one we are focusing is located at the bottom.\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis image shows an oblong, concave, white and brown spotted object that appears to be a cowrie shell with a small projection from a light on the surface. It appears to have a smooth and shiny surface. The edges of the shell are flat, the center is a rounded bump. It is located at the left.\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 2, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 37, | |
| 36, | |
| 263, | |
| 407 | |
| ] | |
| } | |
| ], | |
| "id": 22, | |
| "images": [ | |
| "localization_Orca_Medium_cypraea_erosa_Cypraea erosa18.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis image shows an oblong, concave, white and brown spotted object that appears to be a cowrie shell with a small projection from a light on the surface. It appears to have a smooth and shiny surface. The edges of the shell are flat, the center is a rounded bump. It is located at the left.\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThe image is of a crab. It has a light-blue body with red claws. One claw is much larger. It is crawling across mud. It is located at the right.\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 2, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 384, | |
| 223, | |
| 176, | |
| 208 | |
| ] | |
| } | |
| ], | |
| "id": 23, | |
| "images": [ | |
| "localization_Orca_Medium_stone_crab_005.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThe image is of a crab. It has a light-blue body with red claws. One claw is much larger. It is crawling across mud. It is located at the right.\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThe object in the figure is two jellyfish swimming in the ocean. The jellyfish has a translucent body and it is surrounded by blue background. It is located at the left side of the image.\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 2, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 13, | |
| 100, | |
| 370, | |
| 318 | |
| ] | |
| } | |
| ], | |
| "id": 24, | |
| "images": [ | |
| "localization_Orca_Medium_crystal_jellyfish_015.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThe object in the figure is two jellyfish swimming in the ocean. The jellyfish has a translucent body and it is surrounded by blue background. It is located at the left side of the image.\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis is a halfmoon swimming to the left under the sea. It is located at the center-right position of the image.\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 2, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 294, | |
| 113, | |
| 195, | |
| 110 | |
| ] | |
| } | |
| ], | |
| "id": 25, | |
| "images": [ | |
| "localization_Orca_Medium_halfmoon_000.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis is a halfmoon swimming to the left under the sea. It is located at the center-right position of the image.\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThere are two mud crab lying on the ice and the one we are focusing is located at the left. It has a robust and flattened body covered in a hard shell with deep brown color. It has a pair of strong claws and multiple elongated legs covered in segmented shell.\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 2, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 66, | |
| 77, | |
| 328, | |
| 279 | |
| ] | |
| } | |
| ], | |
| "id": 26, | |
| "images": [ | |
| "localization_Orca_Medium_mud_crab_Google_0060.jpeg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThere are two mud crab lying on the ice and the one we are focusing is located at the left. It has a robust and flattened body covered in a hard shell with deep brown color. It has a pair of strong claws and multiple elongated legs covered in segmented shell.\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis image depicts two sharks swimming in the ocean with coral reefs in the background. The sharks are of different species, one is a blacktip shark and the other is a whitetip shark. The one we are focusing is heading to the right.\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 2, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 212, | |
| 155, | |
| 353, | |
| 212 | |
| ] | |
| } | |
| ], | |
| "id": 27, | |
| "images": [ | |
| "localization_Orca_Medium_whitetip_reef_sharks_Whitetip reef sharks7.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis image depicts two sharks swimming in the ocean with coral reefs in the background. The sharks are of different species, one is a blacktip shark and the other is a whitetip shark. The one we are focusing is heading to the right.\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis figure is an image of a jellyfish with long tentacles in an aquarium. It appears to be an orange or yellow colored creature with long, wavy tendrils that are moving in the water. It has a bell-shaped body. It is closer to the camear compared with another one.\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 2, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 275, | |
| 134, | |
| 369, | |
| 350 | |
| ] | |
| } | |
| ], | |
| "id": 28, | |
| "images": [ | |
| "localization_Orca_Medium_chrysaora_fuscescens_Chrysaora fuscescens28.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis figure is an image of a jellyfish with long tentacles in an aquarium. It appears to be an orange or yellow colored creature with long, wavy tendrils that are moving in the water. It has a bell-shaped body. It is closer to the camear compared with another one.\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThere is a small greenish-yellow fish swimming under the sea. It is located at the top of the image.\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 2, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 336, | |
| 28, | |
| 69, | |
| 18 | |
| ] | |
| } | |
| ], | |
| "id": 29, | |
| "images": [ | |
| "localization_Orca_Medium_green_sea_turtle_021.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThere is a small greenish-yellow fish swimming under the sea. It is located at the top of the image.\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nIt seems to be a moorish idol swimming on top of a coral reef. The moorish idol has a striking appearance characterized by a combination of black, white, and yellow hues. Its predominantly black body features three vertical white stripes\u2014one extending through the head, another across the midsection, and the third near the base of the tail fin. The tail fin itself is black with a contrasting white edge. Elongated dorsal fins display alternating black and yellow stripes. The fish\u2019s pointed snout houses a small mouth at its tip. It is located at the center of the image.\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 2, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 71, | |
| 128, | |
| 525, | |
| 240 | |
| ] | |
| } | |
| ], | |
| "id": 30, | |
| "images": [ | |
| "localization_Orca_Medium_moorish_idol_022.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nIt seems to be a moorish idol swimming on top of a coral reef. The moorish idol has a striking appearance characterized by a combination of black, white, and yellow hues. Its predominantly black body features three vertical white stripes\u2014one extending through the head, another across the midsection, and the third near the base of the tail fin. The tail fin itself is black with a contrasting white edge. Elongated dorsal fins display alternating black and yellow stripes. The fish\u2019s pointed snout houses a small mouth at its tip. It is located at the center of the image.\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis is an image of a blue fish on a black background. It appears to be swimming in the water and is made up of several different blue colors. The fish is located at the right-most position of the image.\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 2, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 876, | |
| 223, | |
| 309, | |
| 155 | |
| ] | |
| } | |
| ], | |
| "id": 31, | |
| "images": [ | |
| "localization_Orca_Medium_damsel_fish_Google_0057.png" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis is an image of a blue fish on a black background. It appears to be swimming in the water and is made up of several different blue colors. The fish is located at the right-most position of the image.\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 1266, | |
| "image_height": 847 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThe image shows a group of sea lions sitting on rocks in a large group. The sea lions are brown and have thick fur. The sea lion is looking to the left and has its mouth open. There are many other sea lions on the rock behind it. It seems to be the largest one from the view of the camera.\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 2, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 341, | |
| 184, | |
| 139, | |
| 229 | |
| ] | |
| } | |
| ], | |
| "id": 32, | |
| "images": [ | |
| "localization_Orca_Medium_eumetopias_jubatus_Eumetopias jubatus15.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThe image shows a group of sea lions sitting on rocks in a large group. The sea lions are brown and have thick fur. The sea lion is looking to the left and has its mouth open. There are many other sea lions on the rock behind it. It seems to be the largest one from the view of the camera.\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThe image shows a shell with a brown and white striped pattern. It appears to be a cone-shaped object with a small opening at the top. The colorful, striped pattern gives the shell a unique and interesting appearance. It is located at the bottom left.\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 2, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 15, | |
| 235, | |
| 389, | |
| 228 | |
| ] | |
| } | |
| ], | |
| "id": 33, | |
| "images": [ | |
| "localization_Orca_Medium_voluta_musica_Image_7.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThe image shows a shell with a brown and white striped pattern. It appears to be a cone-shaped object with a small opening at the top. The colorful, striped pattern gives the shell a unique and interesting appearance. It is located at the bottom left.\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis appears to be an image of two turtles laying on top of each other. They appear to be red-eared sliders, which are a common species of turtle found in the united states. The one we are focusing is on top of another.\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 2, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 224, | |
| 52, | |
| 344, | |
| 256 | |
| ] | |
| } | |
| ], | |
| "id": 34, | |
| "images": [ | |
| "localization_Orca_Medium_trachemys_scripta_elegans_Trachemys scripta elegans34.jpeg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis appears to be an image of two turtles laying on top of each other. They appear to be red-eared sliders, which are a common species of turtle found in the united states. The one we are focusing is on top of another.\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis is a bigeye tuna at the bottom of the image, being placed on a white surface in parallel with other tuna above. The tuna has a torpedo-shaped body with a dark metallic blue color on the back and silvery down to the belly. It has yellow finlets with black edges. The fish appears to be dead.\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 2, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 61, | |
| 257, | |
| 561, | |
| 179 | |
| ] | |
| } | |
| ], | |
| "id": 35, | |
| "images": [ | |
| "localization_Orca_Medium_bigeye_tuna_Google_0065.jpeg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis is a bigeye tuna at the bottom of the image, being placed on a white surface in parallel with other tuna above. The tuna has a torpedo-shaped body with a dark metallic blue color on the back and silvery down to the belly. It has yellow finlets with black edges. The fish appears to be dead.\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis image shows a shrimp or prawn with blue legs and tail sitting on a surface. the shrimp has relatively large head in orange and a pair of long and slender pincers in blue. This shrimp is on the left of the image.\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 2, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 16, | |
| 148, | |
| 436, | |
| 308 | |
| ] | |
| } | |
| ], | |
| "id": 36, | |
| "images": [ | |
| "localization_Orca_Medium_macrobrachium_rosenbergii_Image_1.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis image shows a shrimp or prawn with blue legs and tail sitting on a surface. the shrimp has relatively large head in orange and a pair of long and slender pincers in blue. This shrimp is on the left of the image.\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nIt seems to be a shark resting on a seabed. It is located at the center of the image.\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 2, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 77, | |
| 203, | |
| 409, | |
| 144 | |
| ] | |
| } | |
| ], | |
| "id": 37, | |
| "images": [ | |
| "localization_Orca_Medium_port_jackson_shark_014.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nIt seems to be a shark resting on a seabed. It is located at the center of the image.\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis image shows a group of blue crabs sitting on top of each other in a container. It is put in the right-most position in the container.\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 2, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 322, | |
| 134, | |
| 140, | |
| 348 | |
| ] | |
| } | |
| ], | |
| "id": 38, | |
| "images": [ | |
| "localization_Orca_Medium_blue_crab_Google_0009.jpeg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis image shows a group of blue crabs sitting on top of each other in a container. It is put in the right-most position in the container.\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis is an image of a red and black fish with spots. It appears to be swimming in the water. It has a white belly and red spots on its back. It is located at the left-most position of the image.\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 2, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 91, | |
| 277, | |
| 131, | |
| 173 | |
| ] | |
| } | |
| ], | |
| "id": 39, | |
| "images": [ | |
| "localization_Orca_Medium_clown_wrasse_Google_0041.jpeg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nThis is an image of a red and black fish with spots. It appears to be swimming in the water. It has a white belly and red spots on its back. It is located at the left-most position of the image.\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nstrongylocentrotus franciscanus\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 128, | |
| 33, | |
| 403, | |
| 435 | |
| ] | |
| } | |
| ], | |
| "id": 40, | |
| "images": [ | |
| "localization_Orca_Medium_strongylocentrotus_franciscanus_Strongylocentrotus franciscanus10.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nstrongylocentrotus franciscanus\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nfish\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 267, | |
| 332, | |
| 64, | |
| 88 | |
| ] | |
| } | |
| ], | |
| "id": 41, | |
| "images": [ | |
| "localization_Orca_Medium_chelonia_mydas_Chelonia mydas26.JPG" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nfish\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nbowfin\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 46, | |
| 250, | |
| 598, | |
| 178 | |
| ] | |
| } | |
| ], | |
| "id": 42, | |
| "images": [ | |
| "localization_Orca_Medium_bowfin_Google_0042.jpeg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nbowfin\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncoral\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 6, | |
| 287, | |
| 638, | |
| 197 | |
| ] | |
| } | |
| ], | |
| "id": 43, | |
| "images": [ | |
| "localization_Orca_Medium_blue-ringed_octopus_Google_0034.jpeg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncoral\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsea fan\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 41, | |
| 26, | |
| 438, | |
| 346 | |
| ] | |
| } | |
| ], | |
| "id": 44, | |
| "images": [ | |
| "localization_Orca_Medium_sea_fan_005.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsea fan\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nregal tang fish\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 364, | |
| 194, | |
| 175, | |
| 195 | |
| ] | |
| } | |
| ], | |
| "id": 45, | |
| "images": [ | |
| "localization_Orca_Medium_regal_tang_fish_016.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nregal tang fish\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nbasket star\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 39, | |
| 10, | |
| 581, | |
| 432 | |
| ] | |
| } | |
| ], | |
| "id": 46, | |
| "images": [ | |
| "localization_Orca_Medium_basket_star_Google_0006.jpeg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nbasket star\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ndottyback\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 152, | |
| 156, | |
| 321, | |
| 190 | |
| ] | |
| } | |
| ], | |
| "id": 47, | |
| "images": [ | |
| "localization_Orca_Medium_dottyback_021.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ndottyback\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nmoorish idol\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 188, | |
| 117, | |
| 263, | |
| 225 | |
| ] | |
| } | |
| ], | |
| "id": 48, | |
| "images": [ | |
| "localization_Orca_Medium_moorish_idol_021.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nmoorish idol\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nclark's anemonefish\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 9, | |
| 123, | |
| 334, | |
| 208 | |
| ] | |
| } | |
| ], | |
| "id": 49, | |
| "images": [ | |
| "localization_Orca_Medium_clark_s_anemone_fish_clark_s anemone fish6.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nclark's anemonefish\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsea anemone\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 320, | |
| 23, | |
| 325, | |
| 463 | |
| ] | |
| } | |
| ], | |
| "id": 50, | |
| "images": [ | |
| "localization_Orca_Medium_clark_s_anemone_fish_clark_s anemone fish6.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsea anemone\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncypraea erosa\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 256, | |
| 215, | |
| 263, | |
| 152 | |
| ] | |
| } | |
| ], | |
| "id": 51, | |
| "images": [ | |
| "localization_Orca_Medium_cypraea_erosa_Cypraea erosa20.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncypraea erosa\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nnurse shark\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 6, | |
| 7, | |
| 529, | |
| 464 | |
| ] | |
| } | |
| ], | |
| "id": 52, | |
| "images": [ | |
| "localization_Orca_Medium_nurse_sharks_Nurse sharks2.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nnurse shark\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncarpet anemone\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 40, | |
| 61, | |
| 603, | |
| 419 | |
| ] | |
| } | |
| ], | |
| "id": 53, | |
| "images": [ | |
| "localization_Orca_Medium_carpet_anemone_Google_0018.jpeg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncarpet anemone\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nfish\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 142, | |
| 231, | |
| 44, | |
| 41 | |
| ] | |
| } | |
| ], | |
| "id": 54, | |
| "images": [ | |
| "localization_Orca_Medium_chelonia_mydas_Chelonia mydas5.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nfish\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nblue whale\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 50, | |
| 136, | |
| 547, | |
| 153 | |
| ] | |
| } | |
| ], | |
| "id": 55, | |
| "images": [ | |
| "localization_Orca_Medium_blue_whale_Google_0016.jpeg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nblue whale\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nalpheus bellulus\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 37, | |
| 51, | |
| 490, | |
| 344 | |
| ] | |
| } | |
| ], | |
| "id": 56, | |
| "images": [ | |
| "localization_Orca_Medium_alpheus_bellulus_Image_15.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nalpheus bellulus\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ngymnothorax javanicus\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 112, | |
| 115, | |
| 400, | |
| 284 | |
| ] | |
| } | |
| ], | |
| "id": 57, | |
| "images": [ | |
| "localization_Orca_Medium_gymnothorax_javanicus_Gymnothorax javanicus3.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ngymnothorax javanicus\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nblack marlin\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 56, | |
| 246, | |
| 160, | |
| 142 | |
| ] | |
| } | |
| ], | |
| "id": 58, | |
| "images": [ | |
| "localization_Orca_Medium_black_marlin_Google_0010.jpeg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nblack marlin\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncoral reef shrimp\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 7, | |
| 6, | |
| 428, | |
| 319 | |
| ] | |
| } | |
| ], | |
| "id": 59, | |
| "images": [ | |
| "localization_Orca_Medium_coral_reef_shrimp_Google_0068.jpeg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncoral reef shrimp\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncookiecutter shark\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 56, | |
| 295, | |
| 588, | |
| 189 | |
| ] | |
| } | |
| ], | |
| "id": 60, | |
| "images": [ | |
| "localization_Orca_Medium_cookiecutter_shark_Google_0005.jpeg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncookiecutter shark\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncoral\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 6, | |
| 195, | |
| 638, | |
| 289 | |
| ] | |
| } | |
| ], | |
| "id": 61, | |
| "images": [ | |
| "localization_Orca_Medium_banded_sea_kraits_Banded sea kraits1.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncoral\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nbanded sea krait\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 5, | |
| 119, | |
| 357, | |
| 283 | |
| ] | |
| } | |
| ], | |
| "id": 62, | |
| "images": [ | |
| "localization_Orca_Medium_banded_sea_kraits_Banded sea kraits1.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nbanded sea krait\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nstarfish\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 482, | |
| 264, | |
| 162, | |
| 158 | |
| ] | |
| } | |
| ], | |
| "id": 63, | |
| "images": [ | |
| "localization_Orca_Medium_cusk_Google_0039.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nstarfish\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncusk\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 72, | |
| 91, | |
| 464, | |
| 356 | |
| ] | |
| } | |
| ], | |
| "id": 64, | |
| "images": [ | |
| "localization_Orca_Medium_cusk_Google_0039.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncusk\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsea cucumber\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 176, | |
| 133, | |
| 59, | |
| 44 | |
| ] | |
| } | |
| ], | |
| "id": 65, | |
| "images": [ | |
| "localization_Orca_Medium_snake_eel_Google_0045.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsea cucumber\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsnake eel\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 68, | |
| 121, | |
| 401, | |
| 132 | |
| ] | |
| } | |
| ], | |
| "id": 66, | |
| "images": [ | |
| "localization_Orca_Medium_snake_eel_Google_0045.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsnake eel\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsea urchin\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 6, | |
| 8, | |
| 288, | |
| 340 | |
| ] | |
| } | |
| ], | |
| "id": 67, | |
| "images": [ | |
| "localization_Orca_Medium_parastichopus_parvimensis_Image_16.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsea urchin\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nlontra canadensis\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 180, | |
| 173, | |
| 256, | |
| 194 | |
| ] | |
| } | |
| ], | |
| "id": 68, | |
| "images": [ | |
| "localization_Orca_Medium_starry_flounder_016.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nlontra canadensis\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsnake pipefish\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 67, | |
| 171, | |
| 511, | |
| 85 | |
| ] | |
| } | |
| ], | |
| "id": 69, | |
| "images": [ | |
| "localization_Orca_Medium_snake_pipefish_Google_0098.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsnake pipefish\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsea turtle\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 40, | |
| 91, | |
| 581, | |
| 292 | |
| ] | |
| } | |
| ], | |
| "id": 70, | |
| "images": [ | |
| "localization_Orca_Medium_sea_turtle_014.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsea turtle\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nturtle\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 172, | |
| 172, | |
| 443, | |
| 262 | |
| ] | |
| } | |
| ], | |
| "id": 71, | |
| "images": [ | |
| "localization_Orca_Medium_snapper_002.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nturtle\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ngiant moray\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 110, | |
| 63, | |
| 412, | |
| 318 | |
| ] | |
| } | |
| ], | |
| "id": 72, | |
| "images": [ | |
| "localization_Orca_Medium_giant_moray_Giant moray10.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ngiant moray\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncoral\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 267, | |
| 7, | |
| 377, | |
| 475 | |
| ] | |
| } | |
| ], | |
| "id": 73, | |
| "images": [ | |
| "localization_Orca_Medium_sleeper_goby_021.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncoral\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nblackspotted puffer\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 148, | |
| 7, | |
| 469, | |
| 307 | |
| ] | |
| } | |
| ], | |
| "id": 74, | |
| "images": [ | |
| "localization_Orca_Medium_blackspotted_puffer_Blackspotted puffer1.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nblackspotted puffer\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nchelonia mydas\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 21, | |
| 77, | |
| 555, | |
| 301 | |
| ] | |
| } | |
| ], | |
| "id": 75, | |
| "images": [ | |
| "localization_Orca_Medium_chelonia_mydas_Chelonia mydas20.JPG" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nchelonia mydas\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsea anemone\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 50, | |
| 7, | |
| 561, | |
| 469 | |
| ] | |
| } | |
| ], | |
| "id": 76, | |
| "images": [ | |
| "localization_Orca_Medium_ocellaris_clownfish_Ocellaris clownfish3.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsea anemone\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nocellaris clownfish\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 68, | |
| 190, | |
| 362, | |
| 150 | |
| ] | |
| } | |
| ], | |
| "id": 77, | |
| "images": [ | |
| "localization_Orca_Medium_ocellaris_clownfish_Ocellaris clownfish3.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nocellaris clownfish\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nhoneycomb morays\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 5, | |
| 44, | |
| 479, | |
| 440 | |
| ] | |
| } | |
| ], | |
| "id": 78, | |
| "images": [ | |
| "localization_Orca_Medium_honeycomb_morays_Honeycomb morays1.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nhoneycomb morays\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nshrimp\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 333, | |
| 44, | |
| 197, | |
| 440 | |
| ] | |
| } | |
| ], | |
| "id": 79, | |
| "images": [ | |
| "localization_Orca_Medium_honeycomb_morays_Honeycomb morays1.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nshrimp\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncushion star\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 137, | |
| 27, | |
| 389, | |
| 364 | |
| ] | |
| } | |
| ], | |
| "id": 80, | |
| "images": [ | |
| "localization_Orca_Medium_cushion_star_010.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncushion star\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsea anemone\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 6, | |
| 114, | |
| 509, | |
| 370 | |
| ] | |
| } | |
| ], | |
| "id": 81, | |
| "images": [ | |
| "localization_Orca_Medium_anemonefish_012.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsea anemone\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nshark\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 78, | |
| 67, | |
| 470, | |
| 254 | |
| ] | |
| } | |
| ], | |
| "id": 82, | |
| "images": [ | |
| "localization_Orca_Medium_shark_000.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nshark\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nfeather star\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 94, | |
| 78, | |
| 469, | |
| 381 | |
| ] | |
| } | |
| ], | |
| "id": 83, | |
| "images": [ | |
| "localization_Orca_Medium_feather_star_Google_0033.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nfeather star\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nclam\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 515, | |
| 7, | |
| 44, | |
| 76 | |
| ] | |
| } | |
| ], | |
| "id": 84, | |
| "images": [ | |
| "localization_Orca_Medium_squilla_empusa_Squilla empusa5.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nclam\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nshrimpgoby\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 180, | |
| 73, | |
| 424, | |
| 327 | |
| ] | |
| } | |
| ], | |
| "id": 85, | |
| "images": [ | |
| "localization_Orca_Medium_shrimp_goby_008.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nshrimpgoby\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncleaner shrimp\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 5, | |
| 5, | |
| 388, | |
| 455 | |
| ] | |
| } | |
| ], | |
| "id": 86, | |
| "images": [ | |
| "localization_Orca_Medium_cleaner_shrimp_Cleaner shrimp4.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncleaner shrimp\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\neel\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 148, | |
| 37, | |
| 495, | |
| 445 | |
| ] | |
| } | |
| ], | |
| "id": 87, | |
| "images": [ | |
| "localization_Orca_Medium_cleaner_shrimp_Cleaner shrimp4.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\neel\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nocean sunfish\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 269, | |
| 68, | |
| 361, | |
| 277 | |
| ] | |
| } | |
| ], | |
| "id": 88, | |
| "images": [ | |
| "localization_Orca_Medium_ocean_sunfish_Google_0083.jpeg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nocean sunfish\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nseal\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 130, | |
| 38, | |
| 201, | |
| 222 | |
| ] | |
| } | |
| ], | |
| "id": 89, | |
| "images": [ | |
| "localization_Orca_Medium_ocean_sunfish_Google_0083.jpeg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nseal\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsea anemone\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 439, | |
| 180, | |
| 205, | |
| 304 | |
| ] | |
| } | |
| ], | |
| "id": 90, | |
| "images": [ | |
| "localization_Orca_Medium_copperband_butterflyfish_Copperband butterflyfish8.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsea anemone\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncopperband butterflyfish\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 34, | |
| 56, | |
| 408, | |
| 373 | |
| ] | |
| } | |
| ], | |
| "id": 91, | |
| "images": [ | |
| "localization_Orca_Medium_copperband_butterflyfish_Copperband butterflyfish8.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ncopperband butterflyfish\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nporcupinefish\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 179, | |
| 166, | |
| 224, | |
| 197 | |
| ] | |
| } | |
| ], | |
| "id": 92, | |
| "images": [ | |
| "localization_Orca_Medium_porcupine_fish_012.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nporcupinefish\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ntitan triggerfish\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 6, | |
| 158, | |
| 544, | |
| 293 | |
| ] | |
| } | |
| ], | |
| "id": 93, | |
| "images": [ | |
| "localization_Orca_Medium_titan_triggerfish_Titan triggerfish7.JPG" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ntitan triggerfish\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsea whip\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 104, | |
| 238, | |
| 377, | |
| 236 | |
| ] | |
| } | |
| ], | |
| "id": 94, | |
| "images": [ | |
| "localization_Orca_Medium_barracuda_Barracuda7.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsea whip\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsea anemone\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 204, | |
| 283, | |
| 441, | |
| 201 | |
| ] | |
| } | |
| ], | |
| "id": 95, | |
| "images": [ | |
| "localization_Orca_Medium_pink_anemone_fish_pink anemone fish10.jpeg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nsea anemone\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\npink anemone fish\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 223, | |
| 78, | |
| 365, | |
| 255 | |
| ] | |
| } | |
| ], | |
| "id": 96, | |
| "images": [ | |
| "localization_Orca_Medium_pink_anemone_fish_pink anemone fish10.jpeg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\npink anemone fish\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nzebra shark\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 32, | |
| 175, | |
| 326, | |
| 309 | |
| ] | |
| } | |
| ], | |
| "id": 97, | |
| "images": [ | |
| "localization_Orca_Medium_zebra_shark_juveniles_(rarely_seen_with_stripes)_Zebra shark juveniles (rarely seen with stripes)9.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\nzebra shark\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\njellyfish\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 508, | |
| 360, | |
| 103, | |
| 106 | |
| ] | |
| } | |
| ], | |
| "id": 98, | |
| "images": [ | |
| "localization_Orca_Medium_dermochelys_coriacea_Dermochelys coriacea12.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\njellyfish\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| }, | |
| { | |
| "question": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ndermochelys coriacea\n```\n\nPlease provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n Please output the answer in JSON format without any additional text.\n", | |
| "difficulty": 1, | |
| "dimension": "VisualGrounding", | |
| "sources": "Orca: https://orca.hkustvgd.com/", | |
| "question_format": 3, | |
| "answers": [ | |
| { | |
| "answer": [ | |
| 23, | |
| 47, | |
| 467, | |
| 280 | |
| ] | |
| } | |
| ], | |
| "id": 99, | |
| "images": [ | |
| "localization_Orca_Medium_dermochelys_coriacea_Dermochelys coriacea12.jpg" | |
| ], | |
| "question_raw": "\n\nBased on the descrption below, locate the target object in the image by providing a bounding box. The description is as follows:\n\n```\ndermochelys coriacea\n```\n\n", | |
| "question_choice": null, | |
| "question_rule": "Please provide the bounding boxes of all targets in the format:\n ```json\n {\n \"bboxes\": [\n [x_topleft, y_topleft, width, height],\n [x_topleft, y_topleft, width, height]\n ...\n ],\n }\n ```\n where `bboxes` is a list of bounding boxes in the format [x_topleft, y_topleft, width, height], which is coco format. You may normalize the value.\n ", | |
| "image_width": 650, | |
| "image_height": 490 | |
| } | |
| ] | |
| } |