Data Download

• COCO Format

  
  

  The COCO format is a widely adopted dataset standard used for benchmarking computer vision tasks. It supports object detection, instance segmentation, and keypoint annotations, making it suitable for complex workflows.

• YOLO Format

  
  

  YOLO format is a lightweight, text-based annotation structure designed for real-time and high-speed model training.It supports object detection, segmentation, and keypoints across modern YOLO variants.

• PASCAL VOC Format

  
  

  PASCAL VOC uses an XML-based annotation structure commonly applied in classical vision pipelines. It supports object detection and is compatible with many traditional model training frameworks.

Examples Provided

• COCO Keypoints Format

  {
    "info": {
      "year": " ",
      "version": " ",
      "description": "",
      "contributor": "",
      "url": "",
      "date_created": "2025-10-27T05:10:30+00:00"
    },
    "licenses": [
      {
        "id": 1,
        "url": "",
        "name": ""
      }
    ],
    "categories": [
      {
        "id": 0,
        "name": "pig-keypoint",
        "supercategory": "none"
      },
      {
        "id": 1,
        "name": "pig-keypoint",
        "supercategory": "pig-keypoint",
        "keypoints": [
          "1",
          "2",
          "3",
          "4",
          "5-Left",
          "6-Left",
          "7",
          "8",
          "9",
          "10",
          "11-Left",
          "12-Left"
        ],
        "skeleton": [
          [
            1,
            2
          ],
          [
            2,
            3
          ],
          [
            3,
            4
          ],
          [
            5,
            2
          ],
          [
            6,
            5
          ],
          [
            8,
            7
          ],
          [
            7,
            2
          ],
          [
            9,
            10
          ],
          [
            9,
            4
          ],
          [
            4,
            11
          ],
          [
            11,
            12
          ]
        ]
      }
    ],
    "images": [
      {
        "id": 0,
        "file_name": "pig-3159-_jpg.rf.3d36c12410f85a60398cc0e4187da65b.jpg",
        "height": 640,
        "width": 640
      },
  "annotations": [
      {
        "id": 0,
        "image_id": 0,
        "category_id": 1,
        "bbox": [
          324,
          24,
          114,
          111
        ],
        "area": 12709,
        "segmentation": [],
        "iscrowd": 0,
        "keypoints": [
          435.992,
          67.978,
          2,
          407.221,
          46.641,
          2,
          380.942,
          32.148,
          2,
          351.039,
          46.238,
          2,
          403.143,
          94.146,
          2,
          400.878,
          107.028,
          2,
          417.869,
          94.951,
          2,
          421.946,
          113.872,
          2,
          343.11,
          100.99,
          2,
          348.547,
          125.145,
          2,
          331.33,
          103.002,
          2,
          327.025,
          129.17,
          2
        ]
      },
      {
        "id": 1,
        "image_id": 0,
        "category_id": 1,
        "bbox": [
          488,
          54,
          52,
          189
        ],}
  ]}

• COCO to Pascal VOC (Object Detection) – Example

• COCO to YOLO (Segmentation) – Example

Supported format