f18298335152h/dg-sta

Construct Dynamic Graphs for Hand Gesture Recognition via Spatial-Temporal Attention

This repository holds the Pytorch implementation of Construct Dynamic Graphs for Hand Gesture Recognition via Spatial-Temporal Attention by Yuxiao Chen, Long Zhao, Xi Peng, Jianbo Yuan, and Dimitris N. Metaxas.

Introduction

We propose a Dynamic Graph-Based Spatial-Temporal Attention (DG-STA) method for hand gesture recognition. The key idea is to first construct a fully-connected graph from a hand skeleton, where the node features and edges are then automatically learned via a self-attention mechanism that performs in both spatial and temporal domains. The code of training our approach for skeleton-based hand gesture recognition on the DHG-14/28 Dataset and the SHREC’17 Track Dataset is provided in this repository.

Prerequisites

This package has the following requirements:

• Python 3.6
• Pytorch v1.0.1

Training

1. Download the DHG-14/28 Dataset or the SHREC’17 Track Dataset.

2. Set the path to your downloaded dataset folder in the /util/DHG_parse_data.py (line 2) or the /util/SHREC_parse_data.py (line 5).

3. Set the path for saving your trained models in the train_on_DHG.py (line 117) or the train_on_SHREC.py (line 109) .

4. Run one of following commands.

python train_on_SHREC.py       # on SHREC’17 Track Dataset
python train_on_DHC.py         # on DHG-14/28 Dataset

Citation

If you find this code useful in your research, please consider citing:

@inproceedings{chenBMVC19dynamic,
  author    = {Chen, Yuxiao and Zhao, Long and Peng, Xi and Yuan, Jianbo and Metaxas, Dimitris N.},
  title     = {Construct Dynamic Graphs for Hand Gesture Recognition via Spatial-Temporal Attention},
  booktitle = {BMVC},
  year      = {2019}
}

Acknowledgement

Part of our code is borrowed from the pytorch implementation of Transformer. We thank to the authors for releasing their codes.