ICE-GCN

code for "ICE-GCN: Interactional channel excitation enhanced graph convolutional network for skeleton-based action recognition"

Fig. 1: Schematic diagram of the ICE on the skeleton sequence of action “kicking something”

Fig. 2: Schematic comparison between the proposed CSE (c), CTE (d) and the classical attention module of SENet (a), CBAM (b).

Prerequisites

conda env create -f ice_gcn_environment.yml

Data Preparation

follow CTR-GCN.

Acknowledgements

This repo is based on CTR-GCN and 2s-AGCN. The data processing is borrowed from SGN and HCN.

Thanks to the original authors for their work!

Citation

Please cite this work if you find it useful:.

@article{wang2023ice,
  title={ICE-GCN: An interactional channel excitation-enhanced graph convolutional network for skeleton-based action recognition},
  author={Wang, Shuxi and Pan, Jiahui and Huang, Binyuan and Liu, Pingzhi and Li, Zina and Zhou, Chengju},
  journal={Machine Vision and Applications},
  volume={34},
  number={3},
  pages={40},
  year={2023},
  publisher={Springer}
}