About

This project implements paper Adversarially Occluded Samples for Person Re-identification using pytorch.

Requirements

Python 2.7 (Recommend to use Anaconda for easy package management.)
Pytorch 0.3

The other packages and versions are listed in requirements.txt. You can install them by pip install -r requirements.txt.

Prepare Datasets

Create directory dataset under the project directory, then place datasets in it, as follows.

${project_dir}/dataset
    market1501
        Market-1501-v15.09.15                   # Extracted from Market-1501-v15.09.15.zip, http://www.liangzheng.org/Project/project_reid.html
    cuhk03
        cuhk03-np                               # Extracted from cuhk03-np.zip, https://pan.baidu.com/s/1RNvebTccjmmj1ig-LVjw7A
    duke
        DukeMTMC-reID                           # Extracted from DukeMTMC-reID.zip, https://github.com/layumi/DukeMTMC-reID_evaluation

Then run following command to transform datasets.

python script/dataset/transform_market1501.py
python script/dataset/transform_cuhk03.py
python script/dataset/transform_duke.py

Experiments

Step I: Train Baseline

To train Baseline on Market1501, with GPU 0, run

bash script/experiment/train.sh Baseline market1501 0

Step II: Sliding Window Occlusion

To apply sliding window occlusion with the trained Baseline model and obtain recognition probability, for Market1501, with GPU 0, run

bash script/experiment/sw_occlude.sh market1501 0

Step III: Re-train Model

To re-train the model on Market1501 with original and occluded images, with GPU 0, run

bash script/experiment/train.sh OCCLUSION_TYPE market1501 0

where OCCLUSION_TYPE should be set to No-Adversary, Random, Hard-1, or Sampling.

Citation

If you find our work useful, please kindly cite our paper:

@inproceedings{huang2018adversarially,
  title={Adversarially Occluded Samples for Person Re-Identification},
  author={Huang, Houjing and Li, Dangwei and Zhang, Zhang and Chen, Xiaotang and Huang, Kaiqi},
  booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},
  pages={5098--5107},
  year={2018}
}

huanghoujing/AOS4ReID