Adaptive Exploration for Unsupervised Person Re-identification

Ding, Y; Fan, H; Xu, M; Yang, Y

Adaptive Exploration for Unsupervised Person Re-identification

Ding, Y Fan, H Xu, M Yang, Y

Permalink

Publisher:: Association for Computing Machinery (ACM)
Publication Type:: Journal Article
Citation:: ACM Transactions on Multimedia Computing, Communications and Applications, 2020, 16, (1), pp. 1-19
Issue Date:: 2020-03-01

Closed Access

	Filename	Description	Size
	3369393.pdf	Published version	2.36 MB	Adobe PDF	View/Open

Copyright Clearance Process

Recently Added
In Progress
Closed Access

This item is closed access and not available.

Full metadata record

Field	Value	Language
dc.contributor.author	Ding, Y
dc.contributor.author	Fan, H
dc.contributor.author	Xu, M
dc.contributor.author	Yang, Y https://orcid.org/0000-0001-5528-0546
dc.date.accessioned	2020-10-27T20:38:05Z
dc.date.available	2020-10-27T20:38:05Z
dc.date.issued	2020-03-01
dc.identifier.citation	ACM Transactions on Multimedia Computing, Communications and Applications, 2020, 16, (1), pp. 1-19
dc.identifier.issn	1551-6857
dc.identifier.issn	1551-6865
dc.identifier.uri	http://hdl.handle.net/10453/143546
dc.description.abstract	© 2020 ACM. Due to domain bias, directly deploying a deep person re-identification (re-ID) model trained on one dataset often achieves considerably poor accuracy on another dataset. In this article, we propose an Adaptive Exploration (AE) method to address the domain-shift problem for re-ID in an unsupervised manner. Specifically, in the target domain, the re-ID model is inducted to (1) maximize distances between all person images and (2) minimize distances between similar person images. In the first case, by treating each person image as an individual class, a non-parametric classifier with a feature memory is exploited to encourage person images to move far away from each other. In the second case, according to a similarity threshold, our method adaptively selects neighborhoods for each person image in the feature space. By treating these similar person images as the same class, the non-parametric classifier forces them to stay closer. However, a problem of the adaptive selection is that, when an image has too many neighborhoods, it is more likely to attract other images as its neighborhoods. As a result, a minority of images may select a large number of neighborhoods while a majority of images has only a few neighborhoods. To address this issue, we additionally integrate a balance strategy into the adaptive selection. We evaluate our methods with two protocols. The first one is called "target-only re-ID", in which only the unlabeled target data is used for training. The second one is called "domain adaptive re-ID", in which both the source data and the target data are used during training. Experimental results on large-scale re-ID datasets demonstrate the effectiveness of our method. Our code has been released at https://github.com/dyh127/Adaptive-Exploration-for-Unsupervised-Person-Re-Identification.
dc.language	en
dc.publisher	Association for Computing Machinery (ACM)
dc.relation.ispartof	ACM Transactions on Multimedia Computing, Communications and Applications
dc.relation.isbasedon	10.1145/3369393
dc.rights	info:eu-repo/semantics/restrictedAccess
dc.subject	0803 Computer Software, 0805 Distributed Computing, 0806 Information Systems
dc.subject.classification	Artificial Intelligence & Image Processing
dc.title	Adaptive Exploration for Unsupervised Person Re-identification
dc.type	Journal Article
utslib.citation.volume	16
utslib.for	0803 Computer Software
utslib.for	0805 Distributed Computing
utslib.for	0806 Information Systems
pubs.organisational-group	/University of Technology Sydney/Faculty of Engineering and Information Technology
pubs.organisational-group	/University of Technology Sydney/Strength - AAII - Australian Artificial Intelligence Institute
pubs.organisational-group	/University of Technology Sydney
utslib.copyright.status	closed_access	*
dc.date.updated	2020-10-27T20:37:58Z
pubs.issue	1
pubs.publication-status	Published
pubs.volume	16
utslib.citation.issue	1

Abstract:

© 2020 ACM. Due to domain bias, directly deploying a deep person re-identification (re-ID) model trained on one dataset often achieves considerably poor accuracy on another dataset. In this article, we propose an Adaptive Exploration (AE) method to address the domain-shift problem for re-ID in an unsupervised manner. Specifically, in the target domain, the re-ID model is inducted to (1) maximize distances between all person images and (2) minimize distances between similar person images. In the first case, by treating each person image as an individual class, a non-parametric classifier with a feature memory is exploited to encourage person images to move far away from each other. In the second case, according to a similarity threshold, our method adaptively selects neighborhoods for each person image in the feature space. By treating these similar person images as the same class, the non-parametric classifier forces them to stay closer. However, a problem of the adaptive selection is that, when an image has too many neighborhoods, it is more likely to attract other images as its neighborhoods. As a result, a minority of images may select a large number of neighborhoods while a majority of images has only a few neighborhoods. To address this issue, we additionally integrate a balance strategy into the adaptive selection. We evaluate our methods with two protocols. The first one is called "target-only re-ID", in which only the unlabeled target data is used for training. The second one is called "domain adaptive re-ID", in which both the source data and the target data are used during training. Experimental results on large-scale re-ID datasets demonstrate the effectiveness of our method. Our code has been released at https://github.com/dyh127/Adaptive-Exploration-for-Unsupervised-Person-Re-Identification.

Please use this identifier to cite or link to this item:

http://hdl.handle.net/10453/143546