Joint transfer component analysis and metric learning for person re-identification

Yunzhou Zhang, Yixiu Liu, Sonya Coleman, Jianning Chi

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

A novel and efficient metric learning strategy for person re-identification is proposed. Person re-identification is formulated as a multi-domain learning problem. The assumption that the feature distributions from different camera views are the same is overthrown in this Letter. ID-based transfer component analysis (IDB-TCA) is proposed to learn a shared subspace, in which the differences in the feature distribution between source domain and target domain are significantly reduced. Experimental evaluation on the CUHK01 dataset demonstrates that metric learning with IDB-TCA embedded outperforms state-of-art metric methods for person re-identification.
LanguageEnglish
Pages821-823
JournalElectronics Letters
Volume54
Issue number13
DOIs
Publication statusPublished - 21 Jun 2018

Fingerprint

Cameras

Keywords

  • component analysis
  • person re-identification
  • machine learning

Cite this

Zhang, Yunzhou ; Liu, Yixiu ; Coleman, Sonya ; Chi, Jianning. / Joint transfer component analysis and metric learning for person re-identification. In: Electronics Letters. 2018 ; Vol. 54, No. 13. pp. 821-823.
@article{19c0efa0a54e442589c9ac49c2a8002d,
title = "Joint transfer component analysis and metric learning for person re-identification",
abstract = "A novel and efficient metric learning strategy for person re-identification is proposed. Person re-identification is formulated as a multi-domain learning problem. The assumption that the feature distributions from different camera views are the same is overthrown in this Letter. ID-based transfer component analysis (IDB-TCA) is proposed to learn a shared subspace, in which the differences in the feature distribution between source domain and target domain are significantly reduced. Experimental evaluation on the CUHK01 dataset demonstrates that metric learning with IDB-TCA embedded outperforms state-of-art metric methods for person re-identification.",
keywords = "component analysis, person re-identification, machine learning",
author = "Yunzhou Zhang and Yixiu Liu and Sonya Coleman and Jianning Chi",
year = "2018",
month = "6",
day = "21",
doi = "10.1049/el.2018.0324",
language = "English",
volume = "54",
pages = "821--823",
journal = "Electronics Letters",
issn = "0013-5194",
number = "13",

}

Joint transfer component analysis and metric learning for person re-identification. / Zhang, Yunzhou; Liu, Yixiu; Coleman, Sonya; Chi, Jianning.

In: Electronics Letters, Vol. 54, No. 13, 21.06.2018, p. 821-823.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Joint transfer component analysis and metric learning for person re-identification

AU - Zhang, Yunzhou

AU - Liu, Yixiu

AU - Coleman, Sonya

AU - Chi, Jianning

PY - 2018/6/21

Y1 - 2018/6/21

N2 - A novel and efficient metric learning strategy for person re-identification is proposed. Person re-identification is formulated as a multi-domain learning problem. The assumption that the feature distributions from different camera views are the same is overthrown in this Letter. ID-based transfer component analysis (IDB-TCA) is proposed to learn a shared subspace, in which the differences in the feature distribution between source domain and target domain are significantly reduced. Experimental evaluation on the CUHK01 dataset demonstrates that metric learning with IDB-TCA embedded outperforms state-of-art metric methods for person re-identification.

AB - A novel and efficient metric learning strategy for person re-identification is proposed. Person re-identification is formulated as a multi-domain learning problem. The assumption that the feature distributions from different camera views are the same is overthrown in this Letter. ID-based transfer component analysis (IDB-TCA) is proposed to learn a shared subspace, in which the differences in the feature distribution between source domain and target domain are significantly reduced. Experimental evaluation on the CUHK01 dataset demonstrates that metric learning with IDB-TCA embedded outperforms state-of-art metric methods for person re-identification.

KW - component analysis

KW - person re-identification

KW - machine learning

U2 - 10.1049/el.2018.0324

DO - 10.1049/el.2018.0324

M3 - Article

VL - 54

SP - 821

EP - 823

JO - Electronics Letters

T2 - Electronics Letters

JF - Electronics Letters

SN - 0013-5194

IS - 13

ER -