This repository contains codes for data generation about our paper in the Pattern Recognition journal:
Radar gait recognition using Dual-branch Swin Transformer with Asymmetric Attention Fusion
Wentao He, Jianfeng Ren, Ruibin Bai, Xudong Jiang
Pattern Recognition, 2025.
Step 1: Create file directories ./dataset and ./DataMatrix.
Step 2: Download the raw signal files (~14GB) from Google Drive or 123 Cloud Drive and save data to ./dataset. File path should be ./dataset/dataset/001/...
Step 3: Run generate_spectrogram_for_all.m to convert raw gait signals into spectrograms in .mat format.
Step 4: Generate a data partition sequence to split train and test subsets with different settings.
- Run
generate_half_partition.m: For each subject with$N$ sequences ($N=10$ or$20$ ), sequences$[1:\frac{N}{2}]$ will be used for training and sequences$[\frac{N}{2}:N]$ will be used for testing. - Run
generate_random_partition.m: For each subject with$N$ sequences ($N=10$ or$20$ ), randomly selected$\frac{N}{2}$ sequences will be used for training and the remaining will be used for testing. - Run
generate_second_partition.m: The train and test subsets will be partitioned based on Collecting Phase I/II for Protocol II evaluation. Please refer to the original paper for more details.
Step 5: Run mat2pngSpec.m or mat2pngCVD.m to generate spectrogram/CVD patches with size of
We have also uploaded the image files of generated spectrograms/CVDs to Google Drive. Alternatively, you may use the image dataset with customized train/test partitions.
We thank you for showing interest in our work. If you find the paper and/or the code helpful, please consider citing us using:
@article{he2025radar,
title={Radar gait recognition using Dual-branch Swin Transformer with Asymmetric Attention Fusion},
author={He, Wentao and Ren, Jianfeng and Bai, Ruibin and Jiang, Xudong},
journal={Pattern Recognition},
volume={159},
pages={111101},
year={2025}
}
@inproceedings{chen2022attention,
title={Attention-based dual-stream vision transformer for radar gait recognition},
author={Chen, Shiliang and He, Wentao and Ren, Jianfeng and Jiang, Xudong},
booktitle={Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
pages={3668--3672},
year={2022}
}
We'd like to express our sincere gratitude towards all the advisors and anonymous reviewers for helping us improve the paper. We'd like to thank authors for all the pioneer works in this research field.