Awesome

SAAI

The implementation of ICCV 2023 Visible-Infrared Person Re-Identification via Semantic Alignment and Affinity Inference.

:sparkles: News

• 2023-8-26: Release codes and our pretrained models at Baidudisk (7p3w).

Getting Started

Testing

1. Download the dataset and pretrained models (checkpoints) from Baidudisk (7p3w), unzip them.
2. Download the training data SYSU, unzip and put it in correct position.
3. Download the rand_perm_cam.mat, and put it in correct position.
4. Change the dataset path in the file configs/default/dataset.py
5. Run the following command to retrain the model. You need about 22G GPU for the training.

chmod 755 test.sh
./test

Training

1. Download the training data SYSU, unzip and put it in correct position.
2. Change the dataset path in the file configs/default/dataset.py
3. Run the following command to retrain the model. You need about 22G GPU for the training.

chmod 755 train.sh
./train

Requirement

numpy==1.24.2
torch==2.0.0
torchvision==0.15.1
PyYAML==6.0
pytorch-ignite==0.1.2

Citation

If you find our work useful for your research, please consider citing the following papers :)

@InProceedings{Fang_2023_ICCV,
    author    = {Fang, Xingye and Yang, Yang and Fu, Ying},
    title     = {Visible-Infrared Person Re-Identification via Semantic Alignment and Affinity Inference},
    booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
    month     = {October},
    year      = {2023},
    pages     = {11270-11279}
}

Contact

If you find any problem, please feel free to contact me (fangxingye@bit.edu.cn). A brief self-introduction is required, if you would like to get an in-depth help from me.