This repository holds the codebase for the paper:
Spatial Temporal Graph Convolutional Networks for Skeleton-Based Action Recognition Sijie Yan, Yuanjun Xiong and Dahua Lin, AAAI 2018. [Arxiv Preprint]
We experimented on two skeleton-based action recognition datasts: Kinetics-skeleton and NTU RGB+D. Before training and testing, for the convenience of fast data loading, the datasets should be converted to the proper format. Please download the pre-processed data from GoogleDrive and extract files with
cd st-gcn
unzip <path to st-gcn-processed-data.zip>
If you want to process data by yourself, please refer to SKELETON_DATA.md for more details.
The evaluation of pre-trained models on three datasets can be achieved by:
mmskl configs/recognition/st_gcn_aaai18/$DATASET/test.yamlwhere the $DATASET must be ntu-rgbd-xsub, ntu-rgbd-xview or kinetics-skeleton.
Models will be downloaded automatically before testing.
The expected accuracies are shown here:
| Dataset | Top-1 Accuracy (%) | Top-5 Accuracy (%) | Download |
|---|---|---|---|
| Kinetics-skeleton | 31.60 | 53.68 | model |
| NTU RGB+D Cross View | 88.76 | 98.83 | model |
| NTU RGB+D Cross Subject | 81.57 | 96.85 | model |
To train a ST-GCN model, run
mmskl configs/recognition/st_gcn_aaai18/$DATASET/train.yaml [optional arguments]The usage of optional arguments can be checked via adding --help argument.
All outputs (log files and ) will be saved to the default working directory.
That can be changed by modifying the configuration file
or adding a optional argument --work_dir $WORKING_DIRECTORY in the command line.
After that, evaluate your models by:
mmskl configs/recognition/st_gcn_aaai18/$DATASET/test.yaml --checkpoint $CHECKPOINT_FILE