Awesome
Part-Aligned Bilinear Representations for Person Re-identification
Yumin Suh, Jingdong Wang, Siyu Tang, Tao Mei, and Kyoung Mu Lee. “Part-Aligned Bilinear Representations for Person Re-Identification”, Proceedings of the European Conference on Computer Vision (ECCV), 2018. (paper)
@InProceedings{suh_eccv18,
author = {Yumin Suh and Jingdong Wang and Siyu Tang and Tao Mei and Kyoung Mu Lee},
title = {Part-Aligned Bilinear Representations for Person Re-Identification},
booktitle = {European Conference on Computer Vision (ECCV)},
year = {2018}
}
Our paper is firstly implemented in caffe and reimplemented in pytorch. Minor details have been changed
Contact: Yumin Suh (n12345@snu.ac.kr)
Prerequisite
- Pytorch 0.4
Acknowledgement
- This code is modified from the original code https://github.com/Cysu/open_reid
Usage
-
Run
train_market1501.sh
to train a model on the Market-1501 dataset -
Run
eval_market1501.sh
to evaluate -
Run
train_dukemtmc.sh
to train a model on the DukeMTMC dataset -
Run
eval_dukemtmc.sh
to evaluate