TL; DR: Conbined with wireless positioning data, we propose a novel method to boost unsupervised person re-identification performance under weak scene labeling.
Implementation of UMTF(TPAMI'22) and another variant in the Weak Scene Labeling scenario
- 2/16/2023: Still updating, please stay tuned.
Here is the weak scene labeling setting.
Here is the UMTF framework.
Here is the overview of the variant(in Chinese).
We will release the link of our dataset soon. Please contact Yiheng Liu or Qi Sun.
Run the following command to train the network
sh tools/run.sh
The evaluation results should be consistent with the table below.(in Chinese)
The evaluation results should be consistent with the table below.(in Chinese)
This project is built upon SpCL and Cluster-Contrast. We thank all the authors for their great work and repos.
If you find our code or paper useful, please cite
@ARTICLE{Liu2022UMTF,
author={Liu, Yiheng and Zhou, Wengang and Xie, Qiaokang and Li, Houqiang},
journal={IEEE Transactions on Pattern Analysis and Machine Intelligence},
title={Unsupervised Person Re-Identification with Wireless Positioning under Weak Scene Labeling},
year={2022},
volume={},
number={},
pages={1-14},
doi={10.1109/TPAMI.2022.3196364}}
}@inproceedings{Liu2020RCPM,
author = {Liu, Yiheng and Zhou, Wengang and Xi, Mao and Shen, Sanjing and Li, Houqiang},
title = {Vision Meets Wireless Positioning: Effective Person Re-Identification with Recurrent Context Propagation},
year = {2020},
booktitle = {Proceedings of the 28th ACM International Conference on Multimedia},
}