This is the pytorch version of LieNet, which is from the paper Zhiwu Huang, Chengde Wan, Thomas Probst, Luc Van Gool. Deep Learning on Lie Groups for Skeleton-based Action Recognition, In Proc. CVPR 2017.
The original code written by Matlab is provided here. https://github.com/zzhiwu/LieNet
step 1 download g3d data from http://www.vision.ee.ethz.ch/~zzhiwu/ and save them in g3d folder.
step 2 run prediction.py in any IDE.
The code is very simplified with some functions not implemented like load trained model, etc. I will implement them in the future! :)