Feature extraction using random walks

Yue Deng; Qionghai Dai; Zengke Zhang

doi:10.1109/YCICT.2009.5382449

Feature extraction using random walks

Yue Deng^*
, Qionghai Dai
, Zengke Zhang

^*Corresponding author for this work

Tsinghua University

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

3 Scopus citations

Abstract

In this paper, a novel idea, which utilizes the metric on a graph, is proposed to extract prominent features for pattern recognition. This proposed model, called "Graphical Metrics Guided Transformation" (GMGT), aims to find projections that can preserve the original metric on the graphic domain in a new Euclidean subspace. With the functional analysis, we present the definition of the metric in the graphical domain and prove that the commute time of random walk is a metric on graphs with the help of real physical model. Furthermore, a new feature extraction algorithm based on GMGT and the commute time is proposed, and is applied to face recognition.

Original language	English
Title of host publication	Proceedings - 2009 IEEE Youth Conference on Information, Computing and Telecommunication, YC-ICT2009
Pages	498-501
Number of pages	4
DOIs	https://doi.org/10.1109/YCICT.2009.5382449
State	Published - 2009
Externally published	Yes
Event	2009 IEEE Youth Conference on Information, Computing and Telecommunication, YC-ICT2009 - Beijing, China Duration: 20 Sep 2009 → 21 Sep 2009

Publication series

Name	Proceedings - 2009 IEEE Youth Conference on Information, Computing and Telecommunication, YC-ICT2009

Conference

Conference	2009 IEEE Youth Conference on Information, Computing and Telecommunication, YC-ICT2009
Country/Territory	China
City	Beijing
Period	20/09/09 → 21/09/09

Keywords

Commute time
Random walk
Spectral graphic
Subspace learning

Access to Document

10.1109/YCICT.2009.5382449

Cite this

@inproceedings{661c100b03344a7e9cc030e4e181a767,

title = "Feature extraction using random walks",

abstract = "In this paper, a novel idea, which utilizes the metric on a graph, is proposed to extract prominent features for pattern recognition. This proposed model, called {"}Graphical Metrics Guided Transformation{"} (GMGT), aims to find projections that can preserve the original metric on the graphic domain in a new Euclidean subspace. With the functional analysis, we present the definition of the metric in the graphical domain and prove that the commute time of random walk is a metric on graphs with the help of real physical model. Furthermore, a new feature extraction algorithm based on GMGT and the commute time is proposed, and is applied to face recognition.",

keywords = "Commute time, Random walk, Spectral graphic, Subspace learning",

author = "Yue Deng and Qionghai Dai and Zengke Zhang",

year = "2009",

doi = "10.1109/YCICT.2009.5382449",

language = "英语",

isbn = "9781424450756",

series = "Proceedings - 2009 IEEE Youth Conference on Information, Computing and Telecommunication, YC-ICT2009",

pages = "498--501",

booktitle = "Proceedings - 2009 IEEE Youth Conference on Information, Computing and Telecommunication, YC-ICT2009",

note = "2009 IEEE Youth Conference on Information, Computing and Telecommunication, YC-ICT2009 ; Conference date: 20-09-2009 Through 21-09-2009",

}

Deng, Y, Dai, Q & Zhang, Z 2009, Feature extraction using random walks. in Proceedings - 2009 IEEE Youth Conference on Information, Computing and Telecommunication, YC-ICT2009., 5382449, Proceedings - 2009 IEEE Youth Conference on Information, Computing and Telecommunication, YC-ICT2009, pp. 498-501, 2009 IEEE Youth Conference on Information, Computing and Telecommunication, YC-ICT2009, Beijing, China, 20/09/09. https://doi.org/10.1109/YCICT.2009.5382449

Feature extraction using random walks. / Deng, Yue; Dai, Qionghai; Zhang, Zengke.
Proceedings - 2009 IEEE Youth Conference on Information, Computing and Telecommunication, YC-ICT2009. 2009. p. 498-501 5382449 (Proceedings - 2009 IEEE Youth Conference on Information, Computing and Telecommunication, YC-ICT2009).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Feature extraction using random walks

AU - Deng, Yue

AU - Dai, Qionghai

AU - Zhang, Zengke

PY - 2009

Y1 - 2009

N2 - In this paper, a novel idea, which utilizes the metric on a graph, is proposed to extract prominent features for pattern recognition. This proposed model, called "Graphical Metrics Guided Transformation" (GMGT), aims to find projections that can preserve the original metric on the graphic domain in a new Euclidean subspace. With the functional analysis, we present the definition of the metric in the graphical domain and prove that the commute time of random walk is a metric on graphs with the help of real physical model. Furthermore, a new feature extraction algorithm based on GMGT and the commute time is proposed, and is applied to face recognition.

AB - In this paper, a novel idea, which utilizes the metric on a graph, is proposed to extract prominent features for pattern recognition. This proposed model, called "Graphical Metrics Guided Transformation" (GMGT), aims to find projections that can preserve the original metric on the graphic domain in a new Euclidean subspace. With the functional analysis, we present the definition of the metric in the graphical domain and prove that the commute time of random walk is a metric on graphs with the help of real physical model. Furthermore, a new feature extraction algorithm based on GMGT and the commute time is proposed, and is applied to face recognition.

KW - Commute time

KW - Random walk

KW - Spectral graphic

KW - Subspace learning

UR - https://www.scopus.com/pages/publications/77950892752

U2 - 10.1109/YCICT.2009.5382449

DO - 10.1109/YCICT.2009.5382449

M3 - 会议稿件

AN - SCOPUS:77950892752

SN - 9781424450756

T3 - Proceedings - 2009 IEEE Youth Conference on Information, Computing and Telecommunication, YC-ICT2009

SP - 498

EP - 501

BT - Proceedings - 2009 IEEE Youth Conference on Information, Computing and Telecommunication, YC-ICT2009

T2 - 2009 IEEE Youth Conference on Information, Computing and Telecommunication, YC-ICT2009

Y2 - 20 September 2009 through 21 September 2009

ER -

Feature extraction using random walks

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this