Dynamic Semantic-based Spatial-Temporal Graph Convolution Network for Skeleton-based Human Action Recognition

This project is generated based on the the pyskl (https://github.com/kennymckormick/pyskl)

Below are quick steps for installation:

git clone https://github.com/davelailai/DS-STGCN

cd DS-STGCN

conda env create -f pyskl.yaml

conda activate pyskl

pip install -e .

Run our method :

bash tools/dist_train.sh  configs/dsstgcn/ntu60_xsub_3dkp/j.py 1

Please cite the paper :

Xie, J., Meng, Y., Zhao, Y., Nguyen, A., Yang, X., & Zheng, Y. (2024). Dynamic Semantic-Based Spatial Graph Convolution Network for Skeleton-Based Human Action Recognition. Proceedings of the AAAI Conference on Artificial Intelligence, 38(6), 6225-6233. https://doi.org/10.1609/aaai.v38i6.28440

Xie, J., Meng, Y., Zhao, Y., Nguyen, A., Yang, X., & Zheng, Y. (2024). Dynamic Semantic-based Spatial-Temporal Graph Convolution Network for Skeleton-based Human Action Recognition. Trans on Image processing (TIP).