SiCL: Silhouette-Driven Contrastive Learning for Unsupervised Person Re-Identification with Clothes Change
🚀 SiCL is dedicated to an innovative task: unsupervised clothes changing person re-identification.
⭐ Within the realm of clothing changing person re-identification, SiCL proudly stands as the inaugural unsupervised methodology to attain commendable outcomes across a multitude of datasets!!!
❤️ We warmly welcome and encourage fellow researchers to engage in enlightening discussions and exchanges on this topic!!!
📚Paper Link
https://arxiv.org/abs/2305.13600
📚Dataset：
We evaluate SiCL on Six datasets:
| Dataset | Link |
|---|---|
| PRCC | https://www.isee-ai.cn/~yangqize/clothing.html |
| LTCC | https://naiq.github.io/LTCC_Perosn_ReID.html |
| Celeb-ReID | https://github.com/Huang-3/Celeb-reID |
| Celeb-ReID-Light | https://github.com/Huang-3/Celeb-reID |
| VC-Clothes | https://wanfb.github.io/dataset.html |
| DeepChange | https://github.com/PengBoXiangShang/deepchange |
💬 Remark:
In SiCL, during the training process, it needs to prepare the original dataset along with the corresponding dataset of person masks.
The mask images used in the SiCL are generated based on human parsing networks.
In this study, we employed SCHP for pedestrian silhouette extraction.
After generating the corresponding mask image dataset, you should change the dataset path in CMC.py and /utils/dataset/data/preprocessor.py.
❤️ **We extend a warm invitation to researchers to venture into exploring diverse approaches in generating pedestrian semantic information and conducting experiments on SiCL.
❤️We are greatly anticipating the researchers' invaluable contributions in terms of sharing and providing feedback on their experimental outcomes!**
💡Train:
sh run_code.sh
📢 We will release the model weight soon!