pose skeleton (PoseTrack)

[sanjeevini178]

How do I extract pose skeletons alone as graphs, preferably? From videos, though I'm able to save them with annotated boxes and skeletons, how do I extract just those skeletons?

this is the command I currently use:
!python scripts/demo_inference.py --cfg configs/halpe_136/resnet/256x192_res50_lr1e-3_2x-regression.yaml --checkpoint pretrained_models/halpe136_fast50_regression_256x192.pth --showbox --video /content/drive/MyDrive/fingertap21_.MOV --outdir /content/drive/MyDrive/outputvideos/ --save_video --sp --pose_track

[zhou-jered]

Currenyly there is no way to reach your goal by a single commandline, you need to write code, after get the keypoints, you draw it yourself, that's it.

[sanjeevini178]

Currenyly there is no way to reach your goal by a single commandline, you need to write code, after get the keypoints, you draw it yourself, that's it.

yep, did that but just wanted to know if this repo had a way. thanks!