Pytorch Code of our approach for "Homogeneous and Heterogeneous Relational Graph for Visible-infrared Person Re-identification" in PDF

Results on the SYSU-MM01 Dataset an the RegDB Dataset

Method	Datasets	Rank@1	mAP
Ours	#SYSU-MM01 (All-Search)	~ 78.10 %	~ 72.88 %
Ours	#SYSU-MM01 (Indoor-Search)	~ 81.70 %	~ 81.44 %
Ours	#RegDB	~ 94.92 %	~ 94.58 %

*The code has been tested in Python 3.7, PyTorch=1.0. Both of these two datasets may have some fluctuation due to random splitting

1. Prepare the datasets.

• (1) RegDB Dataset [1]: The RegDB dataset can be downloaded from this website by submitting a copyright form.

• (2) SYSU-MM01 Dataset [2]: The SYSU-MM01 dataset can be downloaded from this website.

  • run python pre_process_sysu.py in to pepare the dataset, the training data will be stored in ".npy" format.

2. Training.

• --dataset: which dataset "sysu" or "regdb".

• --lr: initial learning rate.

• --gpu: which gpu to run.

You may need manually define the data path first.

3. Citation

Please kindly cite the references in your publications if it helps your research:

@article{feng2021homogeneous,
  title={Homogeneous and Heterogeneous Relational Graph for Visible-infrared Person Re-identification},
  author={Feng, Yujian and Chen, Feng and Yu, Jian and Ji, Yimu and Wu, Fei and Liu, Shangdong},
  journal={arXiv preprint arXiv:2109.08811},
  year={2021}
}

4. References

[1] D. T. Nguyen, H. G. Hong, K. W. Kim, and K. R. Park. Person recognition system based on a combination of body images from visible light and thermal cameras. Sensors, 17(3):605, 2017.

[2] A. Wu, W.-s. Zheng, H.-X. Yu, S. Gong, and J. Lai. Rgb-infrared crossmodality person re-identification. In IEEE International Conference on Computer Vision (ICCV), pages 5380–5389, 2017.

[3] M. Ye, Z. Wang, X. Lan, and P. C. Yuen. Visible thermal person reidentification via dual-constrained top-ranking. In International Joint Conference on Artificial Intelligence (IJCAI), pages 1092–1099, 2018.

[4] Liu H, Tan X, Zhou X. Parameter sharing exploration and hetero-center triplet loss for visible-thermal person re-identification[J]. IEEE Transactions on Multimedia, 2020.

[5] Ye, Mang, et al. "Deep learning for person re-identification: A survey and outlook." IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021.

[6] Zhang X, Luo H, Fan X, et al. Alignedreid: Surpassing human-level performance in person re-identification[J]. arXiv preprint arXiv:1711.08184, 2017.