CHASE: Learning Convex Hull Adaptive Shift for Skeleton-based Multi-Entity Action Recognition

Here's the official implementation of

1. De-biasing Skeleton-based Action Recognition with Convex Hull Adaptive Shift (Under Review)
2. CHASE: Learning Convex Hull Adaptive Shift for Skeleton-based Multi-Entity Action Recognition accepted in NeurIPS 2024.

1. De-biasing Skeleton-based Action Recognition with Convex Hull Adaptive Shift

Detailed Implementation: See folder JournalSubmission README 1 & README 2

2. [NeurIPS'24] CHASE: Learning Convex Hull Adaptive Shift for Skeleton-based Multi-Entity Action Recognition

Detailed Implementation: See folder NeurIPS24 README

To clone the main branch only (for code) and exclude the gh-pages branch (for project page), use the following git command:

git clone -b main https://github.com/Necolizer/CHASE.git
cd ./NeurIPS24
pip install -r requirements.txt 

For datasets:

• Please refer to ISTA-Net and follow the instructions in section Prepare the Datasets to prepare NTU Mutual 11 & 26, H2O, and Assembly101.
• Please refer to COMPOSER repo's section Dataset Preparation to get Collective Activity and Volleyball. You could directly download the data using their provided google drive links.

To run the code:

python main.py --config config/[yourBackboneName]/[dataset]/[yourSetting]_chase.yaml
python main_group.py --config config/[yourBackboneName]/[cadORvol]/[yourSetting]_chase.yaml

Checkpoints of the best backbone for each benchmark are provided in this Hugging Face repo.

3. Citation

If you find this work or code helpful in your research, please consider citing:

@inproceedings{NEURIPS2024_wen2024chase,
    author = {Wen, Yuhang and Liu, Mengyuan and Wu, Songtao and Ding, Beichen},
    booktitle = {Advances in Neural Information Processing Systems},
    editor = {A. Globerson and L. Mackey and D. Belgrave and A. Fan and U. Paquet and J. Tomczak and C. Zhang},
    pages = {9388--9420},
    publisher = {Curran Associates, Inc.},
    title = {CHASE: Learning Convex Hull Adaptive Shift for Skeleton-based Multi-Entity Action Recognition},
    url = {https://proceedings.neurips.cc/paper_files/paper/2024/file/11f5520daf9132775e8604e89f53925a-Paper-Conference.pdf},
    volume = {37},
    year = {2024}
}

@INPROCEEDINGS{wen2023interactive,
    author={Wen, Yuhang and Tang, Zixuan and Pang, Yunsheng and Ding, Beichen and Liu, Mengyuan},
    booktitle={2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)}, 
    title={Interactive Spatiotemporal Token Attention Network for Skeleton-Based General Interactive Action Recognition}, 
    year={2023},
    pages={7886-7892},
    doi={10.1109/IROS55552.2023.10342472}
}