[Paper]

The implementation of our paper Unsupervised 3D Human Pose Representation with Viewpoint and Pose Disentanglement (ECCV2020).

To train the pose encoder using corrupted human skeletons, please run

python train_poseAE.py

Some extracted features of action datasets using the trained encoder can be download from here.

The use of this software is RESTRICTED to non-commercial research and educational purposes.

@inproceedings{Unsup3DPose,
  title={Unsupervised 3D Human Pose Representation with Viewpoint and Pose Disentanglement},
  author={Nie, Qiang and Liu, Ziwei and Liu, Yunhui},
  booktitle={European Conference on Computer Vision (ECCV)},
  year={2020}
}