PyQt-based Image/Video multi-object key-point annotator
- Support keypoints labelling for multiple objects
- Custom naming: Prompt for custom object name to save in JSON (name should be unique and non-empty)
- Mode Switch: Mode 1 (multiple objects on single image) and Mode 2 (multiple objects on separate images)
- Clear annoation: clear current object annotation in case of wrong labelling
{
"bowl": {
"keypoints": [
[
287,
190
],
[
291,
386
],
[
360,
301
],
[
217,
284
]
],
"no_keypoints": 4
},
"arm": {
"keypoints": [
[
95,
204
],
[
182,
346
]
],
"no_keypoints": 2
}
}

