udt-labelme-converter
v0.1.7
Published
```json { "version": "4.5.0", "flags": "", "shapes": "", "imagePath": "", "imageData": "", "imageHeight": "", "imageWidth": "" } ```
Downloads
6
Maintainers
Readme
Run script
npx udt-labelme-converter --input=my-precious-samples.json --output=where-i-want-to-see-my-output-files
Example Labelme Format
{
"version": "4.5.0",
"flags": "",
"shapes": {
"label": "my_lovely_box",
"points": [
[160, 120],
[210, 190],
[100,30],
[32, 70]
],
"shape_type": "polygon",
"group_id": null,
"flags": {}
},
"imagePath": "./image.jpeg",
"imageData": "base64String",
"imageHeight": 330,
"imageWidth": 450
}
- version:
- It gives an information about Labelme version.
- flags:
- Flags for label.
- shapes:
- label:
- It contains string label
- points:
- It's an array of points which contains x and y
- shape_type:
- It defines type of shape you made on image.
- group_id:
- Every label group has their unique color. This is for specifying label group.
- flags:
- It's an flags object.
- label:
- imagePath:
- It's file path of image used for labelling.
- imageData:
- It's base64 version of image used for labelling.
- imageHeight:
- It's exact height of image used for labelling.
- imageWidth:
- It's exact width of image used for labelling.