/ICE

Implementation for ICCV 2021 paper "ICE: Inter-instance Contrastive Encoding for Unsupervised Person Re-identification"

Primary LanguagePythonMIT LicenseMIT

ICE

This is the official PyTorch implementation of the ICCV 2021 paper ICE: Inter-instance Contrastive Encoding for Unsupervised Person Re-identification.

[Video] [Poster]

teaser

Installation

git clone https://github.com/chenhao2345/ICE
cd ICE
python setup.py develop

Prepare Datasets

Download the raw datasets DukeMTMC-reID, Market-1501, MSMT17, and then unzip them under the directory like

ICE/examples/data
├── dukemtmc-reid
│   └── DukeMTMC-reID
├── market1501
└── msmt17
    └── MSMT17_V1(or MSMT17_V2)

Training

We used 4 GPUs to train our model.

Train Market-1501:

python examples/unsupervised_train.py --dataset-target market1501

Train DukeMTMC-reID:

python examples/unsupervised_train.py --dataset-target dukemtmc-reid

Train MSMT17:

python examples/unsupervised_train.py --dataset-target msmt17

Citation

If you find this project useful, please kindly star our project and cite our paper.

@article{chen2021ice,
  title={ICE: Inter-instance Contrastive Encoding for Unsupervised Person Re-identification},
  author={Chen, Hao and Lagadec, Benoit and Bremond, Francois},
  journal={arXiv preprint arXiv:2103.16364},
  year={2021}
}