PMnet: Learning of Disentangled Pose and Movement for Unsupervised Motion Retargeting

Tensorflow implementation of paper:

PMnet: Learning of Disentangled Pose and Movement for Unsupervised Motion Retargeting (BMVC 2019)

Environments

• Python 2.7

• Tensorflow 1.3.0

• Blender 2.79 (for visualizing character animations)

Our implementation is based on the code for the CVPR 2018 paper (Neural Kinematics Networks for Unsupervised Motion Retargeting) by Ruben Villegas, Jimei Yang, Duygu Ceylan and Honglak Lee.

Please follow the instructions for further details.

Contact ljin0429@gmail.com if you have any questions.

Citation

Please cite our work as follows:

@InProceedings{Lim_pmnet_BMVC19,
  author = {Lim, Jongin and Chang, Hyung Jin and Choi, Jin Young},
  title = {PMnet: Learning of Disentangled Pose and Movement for Unsupervised Motion Retargeting},
  booktitle = {British Machine Vision Conference (BMVC)},
  year = {2019}
}