Process all your datasets in one place

Comparison of Annotation Formats

Technical information and the difference between YOLO Darknet .TXT and Microsoft COCO (Common Objects in Context) .JSON Annotation Formats





General

46

Score

62

Score

Winner
Publication 2015 2014
Example
0 0.643 0.5522388059701493 0.386 0.1791044776119403
                                                                                                        
{
    "info": {
        "year": "2021",
        "version": "1.0",
        "description": "Exported from NanoML.org",
        "contributor": "Peter",
        "url": "https://NanoML.org/datasets/test",
        "date_created": "2021-04-10T14:54:31+00:00"
    },
    "licenses": [
        {
	"url": "http://creativecommons.org/licenses/by-nc-sa/2.0/",
	"id": 1,
	"name": "Attribution-NonCommercial-ShareAlike License"
	}
    ],
    "categories": [
        {
	"id": 1,
	"name": "Car",
	"supercategory": "vehicle"
        }
    ],
    "images": [
        {
	"id": 0,
	"license": 1,
	"file_name": "000001.png",
	"height": 292,
	"width": 400,
	"date_captured": "2021-04-10T14:54:31+00:00"
    ],
    "annotations": [
        {
	"id": 0,
	"image_id": 0,
	"category_id": 1,
	"bbox": [
	128,
	259,
	103,
	30
	],
	"area": 3090,
	"segmentation": [],
	"iscrowd": 0
        }
    ]
}
                                                                                                        
Properties
min. Size 0.06 kB 1.05 kB
Size per additional object 0.05 kB 0.3 kB
File Format .TXT .JSON
Human readability Bad Okay
Popularity
Google Trends Popularity Very Good Good
Model availability Very Good Good
Structure
Uses normalized values
One Annotation File per Image
LabelMap required
Supported Types
Bounding Box Support
Polygonal Segmentation Support
Dot annotation Support
Other
Recommended Tool Labellmg Coco Annotation Tool