3 ms·
It looks like they’re using OpenPose [0] images fed to a special “pose guider” model. You can make them from just regular video. [0] https://github.com/CMU-Per
by netruk44 3y ago
It looks like they’re using OpenPose [0] images fed to a special “pose guider” model. You can make them from just regular video.
[0] https://github.com/CMU-Perceptual-Computing-Lab/openpose https://github.com/CMU-Perceptual-Computing-Lab/openpose