CUPR: Contrastive Unsupervised Learning for Person Re-identification

Khadija Khaldi, Shishir Shah

Abstract

Most of the current person re-identification (Re-ID) algorithms require a large labeled training dataset to obtain better results. For example, domain adaptation-based approaches rely heavily on limited real-world data to alleviate the problem of domain shift. However, such assumptions are impractical and rarely hold, since the data is not freely accessible and require expensive annotation. To address this problem, we propose a novel pure unsupervised learning approach using contrastive learning (CUPR). Our framework is a simple iterative approach that learns strong high-level features from raw pixels using contrastive learning and then performs clustering to generate pseudo-labels. We demonstrate that CUPR outperforms the unsupervised and semi-supervised state-of-the-art methods on Market-1501 and DukeMTMC-reID datasets.

Download


Paper Citation


in Harvard Style

Khaldi K. and Shah S. (2021). CUPR: Contrastive Unsupervised Learning for Person Re-identification.In Proceedings of the 16th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 4: VISAPP, ISBN 978-989-758-488-6, pages 92-100. DOI: 10.5220/0010239900920100


in Bibtex Style

@conference{visapp21,
author={Khadija Khaldi and Shishir Shah},
title={CUPR: Contrastive Unsupervised Learning for Person Re-identification},
booktitle={Proceedings of the 16th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 4: VISAPP,},
year={2021},
pages={92-100},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0010239900920100},
isbn={978-989-758-488-6},
}


in EndNote Style

TY - CONF

JO - Proceedings of the 16th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 4: VISAPP,
TI - CUPR: Contrastive Unsupervised Learning for Person Re-identification
SN - 978-989-758-488-6
AU - Khaldi K.
AU - Shah S.
PY - 2021
SP - 92
EP - 100
DO - 10.5220/0010239900920100