Raymond A. Yeh*,
Yuan-Ting Hu*, Alexander G. Schwing
University of Illinois at Urbana-Champaign
(* indicates equal contribution)
The repository contains Pytorch implementation of Chirality Nets for Human Pose Regression.
If you used this code for your experiments or found it helpful, please consider citing the following paper:
@inproceedings{YehNeurIPS2019,
author = {R.~A. Yeh^\ast$ and Y.-T. Hu^\ast$ and A.~G. Schwing},
title = {Chirality Nets for Human Pose Regression},
booktitle = {Proc. NeurIPS},
year = {2019},
note = {$^\ast$ equal contribution},
}
- Python 3+
- Pytorch 1.1.0
We recommend reading through our short tutorial on chirality equivariance. The tutorial illustrates the chirality definition and API for the chiral layers.
We support chirality equivariant versions of the following layers:
To verify that these layers satisfy chirality equivariance, we have provided some test cases in the test directory
Coming soon.
- 3D human pose estimation in video with temporal convolutions and semi-supervised training in CVPR 2019
- Equivariance Through Parameter-Sharing in ICML 2017
This work is licensed under the MIT License