{
  "dataset": "m2caiSeg",
  "description": "m2caiSeg dataset with 19 semantic classes for surgical scene segmentation. Masks are single-channel integer labels (0-18).",
  "num_classes": 19,
  "class_names": [
    "background",
    "grasper",
    "bipolar",
    "hook",
    "scissors",
    "clipper",
    "irrigator",
    "specimen-bag",
    "trocars",
    "clip",
    "liver",
    "gall-bladder",
    "fat",
    "upperwall",
    "artery",
    "intestine",
    "bile",
    "blood",
    "unknown"
  ],
  "mask_format": "Single-channel grayscale PNG. Pixel values are class IDs (0-18).",
  "class_mapping": {
    "0": {
      "class_id": 0,
      "class_name": "background"
    },
    "1": {
      "class_id": 1,
      "class_name": "grasper"
    },
    "2": {
      "class_id": 2,
      "class_name": "bipolar"
    },
    "3": {
      "class_id": 3,
      "class_name": "hook"
    },
    "4": {
      "class_id": 4,
      "class_name": "scissors"
    },
    "5": {
      "class_id": 5,
      "class_name": "clipper"
    },
    "6": {
      "class_id": 6,
      "class_name": "irrigator"
    },
    "7": {
      "class_id": 7,
      "class_name": "specimen-bag"
    },
    "8": {
      "class_id": 8,
      "class_name": "trocars"
    },
    "9": {
      "class_id": 9,
      "class_name": "clip"
    },
    "10": {
      "class_id": 10,
      "class_name": "liver"
    },
    "11": {
      "class_id": 11,
      "class_name": "gall-bladder"
    },
    "12": {
      "class_id": 12,
      "class_name": "fat"
    },
    "13": {
      "class_id": 13,
      "class_name": "upperwall"
    },
    "14": {
      "class_id": 14,
      "class_name": "artery"
    },
    "15": {
      "class_id": 15,
      "class_name": "intestine"
    },
    "16": {
      "class_id": 16,
      "class_name": "bile"
    },
    "17": {
      "class_id": 17,
      "class_name": "blood"
    },
    "18": {
      "class_id": 18,
      "class_name": "unknown"
    }
  }
}