Yi-Fan Song, Zhang Zhang, Caifeng Shan and Liang Wang. Stronger, Faster and More Explainable: A Graph Convolutional Baseline for Skeleton-based Action Recognition ACM MultiMedia, 2020. [ACMMM 2020] [Arxiv Preprint]
The following pictures are the pipeline of ResGCNv1 and the illustration of PartAtt block, respectively.
This code is based on Python3 (anaconda, >=3.5) and PyTorch (>=1.2.0).
Other Python libraries are presented in the 'scripts/requirements.txt', which can be installed by
pip install -r scripts/requirements.txt
Our models are experimented on the NTU RGB+D 60 & 120 datasets, which can be downloaded from here.
There are 302 samples of NTU RGB+D 60 and 532 samples of NTU RGB+D 120 need to be ignored, which are shown in the 'src/preprocess/ignore.txt'.
Several pretrained models are provided, which include ResGCN-N51, PA-ResGCN-N51, ResGCN-B19, and PA-ResGCN-B19 for the cross-subject (X-sub) and cross-view (X-view) benchmarks of the NTU RGB+D 60 dataset and the cross-subject (X-sub120) and cross-setup (X-set120) benchmarks of the NTU RGB+D 120 dataset.
These models can be downloaded from BaiduYun (Extraction code: d3ea) or GoogleDrive.
Before training and evaluating, there are some parameters should be noticed.
- (1) '--config' or '-c': The config of RA-GCN. You must use this parameter in the command line or the program will output an error. There are 8 configs given in the configs folder, which can be illustrated in the following tabel.
| config | 1001 | 1002 | 1003 | 1004 | 1005 | 1006 | 1007 | 1008 |
|---|---|---|---|---|---|---|---|---|
| model | N51 | N51 | N51 | N51 | B19 | B19 | B19 | B19 |
| benchmark | Xsub | Xview | Xsub120 | Xset120 | Xsub | Xview | Xsub120 | Xset120 |
| config | 1009 | 1010 | 1011 | 1012 | 1013 | 1014 | 1015 | 1016 |
|---|---|---|---|---|---|---|---|---|
| model | PA-N51 | PA-N51 | PA-N51 | PA-N51 | PA-B19 | PA-B19 | PA-B19 | PA-B19 |
| benchmark | Xsub | Xview | Xsub120 | Xset120 | Xsub | Xview | Xsub120 | Xset120 |
-
(2) '--work_dir' or '-w': The path to workdir, for saving checkpoints and other running files. Default is './workdir' in all config files.
-
(3) '--pretrained_path' or '-pp': The path to the downloaded pretrained models. Default is './pretrained' in all config files.
-
(4) '--resume' or '-r': Resume from the recent checkpoint ('<--work_dir>/checkpoint.pth.tar').
-
(5) '--evaluate' or '-e': Only evaluate models. You can choose the evaluating model according to the instructions.
-
(6) '--extract' or '-ex': Extract features from a trained model for visualization. Using this parameter will make a data file named 'extraction_<--config>.npz' at the './visualization' folder.
-
(7) '--visualization' or '-v': Show the information and details of a trained model. You should extract features by using <--extract> parameter before visualizing.
-
(8) '--dataset' or '-d': Choose the dataset. (Choice: [ntu-xsub, ntu-xview, ntu-xsub120, ntu-xset120])
-
(9) '--model_type' or '-mt': Choose the model. (Format: {attention}-resgcn-{structure}-{reduction}, attention: [pa, ca, fa, sa, pca, psa, None], structure: [b15, b19, b23, b29, n39, n51, n57, n75], reduction: [r1, r2, r4, r8, None], e.g., resgcn-b19, resgcn-n51-r4, pa-resgcn-b19, pa-resgcn-n51-r4)
Other parameters can be updated by modifying the corresponding config file in the 'configs' folder or using command line to send parameters to the model, and the parameter priority is command line > yaml config > default value.
Firstly, you should modify the 'path' parameters in all config files of the 'configs' folder.
A python file 'scripts/modify_configs.py' will help you to do this. You need only to change three parameters in this file to your path to NTU datasets.
python scripts/modify_configs.py --path <path/to/save/preprocessed/data> --ntu60_path <path/to/ntu60/dataset> --ntu120_path <path/to/ntu120/dataset>
After modifing the path to datasets, please generate preprocessed datasets by using 'scripts/auto_gen_data.sh'.
bash scripts/auto_gen_data.sh
Or you can preprocess data (same as 2s-AGCN) by the following command (only the first time to use this benchmark). It may take you several hours.
python main.py -c <config> -gd
where <config> is the config file name in the 'configs' folder, e.g., 1001.
Note: only training the NTU X-view benchmark with the PA-ResGCN-N51 (config: 1010) and PA-ResGCN-B19 (config: 1014) models require this preprocessed data.
You can simply train the model by
python main.py -c <config>
If you want to restart training from the saved checkpoint last time, you can run
python main.py -c <config> -r
Before evaluating, you should ensure that the trained model corresponding the config is already existed in the <--pretrained_path> or '<--work_dir>' folder. Then run
python main.py -c <config> -e
To visualize the details of the trained model, you can run
python main.py -c <config> -ex -v
where '-ex' can be removed if the data file 'extraction_<config>.npz' already exists in the './visualization' folder.
Top-1 Accuracy for the provided models on NTU RGB+D 60 & 120 datasets.
| models | parameters | NTU Xsub | NTU Xview | NTU Xsub120 | NTU Xset120 |
|---|---|---|---|---|---|
| ResGCN-N51 | 0.77M | 89.1% | 93.5% | 84.0% | 84.2% |
| ResGCN-B19 | 1.14M | 90.3% | 95.6% | 86.6% | 87.1% |
| PA-ResGCN-N51 | 3.26M | 90.0% | 94.8% | 85.2% | 85.7% |
| PA-ResGCN-B19 | 3.64M | 90.9% | 96.0% | 87.3% | 88.3% |
If you have any question, please send e-mail to [email protected].
Please cite our paper when you use this code in your reseach.
@inproceedings{song2020stronger,
author = {Song, Yi-Fan and Zhang, Zhang and Shan, Caifeng and Wang, Liang},
title = {Stronger, Faster and More Explainable: A Graph Convolutional Baseline for Skeleton-Based Action Recognition},
booktitle = {Proceedings of the 28th ACM International Conference on Multimedia (ACMMM)},
pages = {1625--1633},
year = {2020},
isbn = {9781450379885},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3394171.3413802},
doi = {10.1145/3394171.3413802},
}