Code accompanying the paper
SAR-NAS: Skeleton-based Action Recognition via Neural Architecture Searching
Accepted by the Journal of Visual Communication and Image Representation.
Python >= 3.5.5, PyTorch == 0.4.1, torchvision == 0.2.1.
Skeleton data in NTU RGB+D can be obtained via:
https://drive.google.com/open?id=1CUZnBtYwifVXS21yVg62T-vrPVayso5H
Or via (Kinetics+NTU)：
https://pan.baidu.com/s/1O1azJwxkzh04cOuSWyXi1Q
extracted code: data
python train_search_ntu.py
python train_ntu.py
python draw_confusion_matrix.py
Package graphviz is required to visualize the learned cells
python visualize.py DARTS
where DARTS can be replaced by any customized architectures in genotypes.py
If you use any part of this code in your research, please cite our paper:
@article{Zhang2020sar-darts,
title={SAR-NAS: Skeleton-based Action Recognition via Neural Architecture Searching},
author={Zhang, Haoyuan and Hou, Yonghong and Wang, Pichao and Guo, Zihui and Li, Wanqing},
journal={Journal of Visual Communication and Image Representation, preprint},
year={2020}
}