Two commonly used data sets for object detection, COCO and VOC

{
    "info": info, # basic information about the dataset
    "licenses": [license], # license
    "images": [image], # image information, name, and height
    "annotations": [annotation], # annotation
    "categories": [category] # tag information
}
info{ # Dataset information description
    "year": int, # dataset year
    "version": str, # dataset version
    "description": str, # dataset description
    "contributor": str, # dataset provider
    "url": str, # dataset download link
    "date_created": datetime, # dataset creation date
}
license{
    "id": int,
    "name": str,
    "url": str,
}
image{ # images is a list that stores all the image (dict) information. image is a dict that stores information about a single image
    "id": int, # ID number of the image (unique for each image ID)
    "width": int, # image width
    "height": int, # image height
    "file_name": str, # image name
    "license": int, # agreement
    "flickr_url": str, # flickr link address
    "coco_url": str, # network connection address
    "date_captured": datetime, # dataset fetch date
}
annotation{ # annotations is a list that stores all dict information. An annotation is a dict that stores a single target annotation information.
    "id": int, # Target object ID (unique for each object ID), each image may have multiple targets
    "image_id": int, # corresponds to the image ID
    "category_id": int, # corresponds to the category ID, corresponding to the ID in the categories
    "segmentation": RLE or [polygon], # instance segmentation, the boundary point coordinates of the object [x1,y1,x2,y2,....,xn,yn]
    "area": float, # The area of the object area
    "bbox": [xmin,ymin,width,height], # object detection, object positioning border[x,y,w,h]
    "iscrowd": 0 or 1, # indicates whether it is a crowd or not
}
categories{ # Category description
    "id": int, # The ID corresponding to the category (0 defaults to the background)
    "name": str, # subcategory name
    "supercategory": str, # main category name
}

.NET/C# calculates the area of a polygon
https://www.itsvse.com/thread-10870-1-1.html