Attribute-restricted latent topic model for person re-identification

Xiao Liu; Mingli Song; Qi Zhao; Dacheng Tao; Chun Chen; Jiajun Bu

doi:10.1016/j.patcog.2012.05.019

Attribute-restricted latent topic model for person re-identification

Xiao Liu, Mingli Song, Qi Zhao, Dacheng Tao, Chun Chen, Jiajun Bu

Computer Science and Engineering

Research output: Contribution to journal › Article › peer-review

52 Scopus citations

Abstract

Searching for specific persons from surveillance videos captured by different cameras, known as person re-identification, is a key yet under-addressed challenge. Difficulties arise from the large variations of human appearance in different poses, from the different camera views that may be involved, making low-level descriptor representation unreliable. In this paper, we propose a novel Attribute-Restricted Latent Topic Model (ARLTM) to encode targets into semantic topics. Compared to conventional topic models such as LDA and pLSI, ARLTM performs best by imposing semantic restrictions onto the generation of human specific attributes. We use MCMC EM for model learning. Experimental results show that our method achieves state-of-the-art performance.

Original language	English (US)
Pages (from-to)	4204-4213
Number of pages	10
Journal	Pattern Recognition
Volume	45
Issue number	12
DOIs	https://doi.org/10.1016/j.patcog.2012.05.019
State	Published - Dec 2012

Bibliographical note

Funding Information:
This work is supported by National Natural Science Foundation of China (61170142), National Key Technology R&D Program (2011BAG05B04), the Zhejiang Province Key S&T Innovation Group Project (2009R50009), and the Fundamental Research Funds for the Central Universities (2012FZA5017).

Keywords

Attribute-restricted latent topic model
Person re-identification
Semantic topic
Visual attribute

Access

10.1016/j.patcog.2012.05.019

OpenUrl availability

Full text

Cite this

@article{a49702e06a5146c694ac827a30f4ddcc,

title = "Attribute-restricted latent topic model for person re-identification",

abstract = "Searching for specific persons from surveillance videos captured by different cameras, known as person re-identification, is a key yet under-addressed challenge. Difficulties arise from the large variations of human appearance in different poses, from the different camera views that may be involved, making low-level descriptor representation unreliable. In this paper, we propose a novel Attribute-Restricted Latent Topic Model (ARLTM) to encode targets into semantic topics. Compared to conventional topic models such as LDA and pLSI, ARLTM performs best by imposing semantic restrictions onto the generation of human specific attributes. We use MCMC EM for model learning. Experimental results show that our method achieves state-of-the-art performance.",

keywords = "Attribute-restricted latent topic model, Person re-identification, Semantic topic, Visual attribute",

author = "Xiao Liu and Mingli Song and Qi Zhao and Dacheng Tao and Chun Chen and Jiajun Bu",

note = "Funding Information: This work is supported by National Natural Science Foundation of China (61170142), National Key Technology R&D Program (2011BAG05B04), the Zhejiang Province Key S&T Innovation Group Project (2009R50009), and the Fundamental Research Funds for the Central Universities (2012FZA5017). ",

year = "2012",

month = dec,

doi = "10.1016/j.patcog.2012.05.019",

language = "English (US)",

volume = "45",

pages = "4204--4213",

journal = "Pattern Recognition",

issn = "0031-3203",

publisher = "Elsevier Limited",

number = "12",

}

TY - JOUR

T1 - Attribute-restricted latent topic model for person re-identification

AU - Liu, Xiao

AU - Song, Mingli

AU - Zhao, Qi

AU - Tao, Dacheng

AU - Chen, Chun

AU - Bu, Jiajun

N1 - Funding Information: This work is supported by National Natural Science Foundation of China (61170142), National Key Technology R&D Program (2011BAG05B04), the Zhejiang Province Key S&T Innovation Group Project (2009R50009), and the Fundamental Research Funds for the Central Universities (2012FZA5017).

PY - 2012/12

Y1 - 2012/12

N2 - Searching for specific persons from surveillance videos captured by different cameras, known as person re-identification, is a key yet under-addressed challenge. Difficulties arise from the large variations of human appearance in different poses, from the different camera views that may be involved, making low-level descriptor representation unreliable. In this paper, we propose a novel Attribute-Restricted Latent Topic Model (ARLTM) to encode targets into semantic topics. Compared to conventional topic models such as LDA and pLSI, ARLTM performs best by imposing semantic restrictions onto the generation of human specific attributes. We use MCMC EM for model learning. Experimental results show that our method achieves state-of-the-art performance.

AB - Searching for specific persons from surveillance videos captured by different cameras, known as person re-identification, is a key yet under-addressed challenge. Difficulties arise from the large variations of human appearance in different poses, from the different camera views that may be involved, making low-level descriptor representation unreliable. In this paper, we propose a novel Attribute-Restricted Latent Topic Model (ARLTM) to encode targets into semantic topics. Compared to conventional topic models such as LDA and pLSI, ARLTM performs best by imposing semantic restrictions onto the generation of human specific attributes. We use MCMC EM for model learning. Experimental results show that our method achieves state-of-the-art performance.

KW - Attribute-restricted latent topic model

KW - Person re-identification

KW - Semantic topic

KW - Visual attribute

UR - http://www.scopus.com/inward/record.url?scp=84864284856&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84864284856&partnerID=8YFLogxK

U2 - 10.1016/j.patcog.2012.05.019

DO - 10.1016/j.patcog.2012.05.019

M3 - Article

AN - SCOPUS:84864284856

SN - 0031-3203

VL - 45

SP - 4204

EP - 4213

JO - Pattern Recognition

JF - Pattern Recognition

IS - 12

ER -

Attribute-restricted latent topic model for person re-identification

Abstract

Bibliographical note

Keywords

Access

OpenUrl availability

Other files and links

Fingerprint

Cite this