Locality Regularization Embedding for face verification

Ying Han Pang, Beng Jin Teoh, Fu San Hiew

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

Graph embedding (GE) is a unified framework for dimensionality reduction techniques. GE attempts to maximally preserve data locality after embedding for face representation and classification. However, estimation of true data locality could be severely biased due to limited number of training samples, which trigger overfitting problem. In this paper, a graph embedding regularization technique is proposed to remedy this problem. The regularization model, dubbed as Locality Regularization Embedding (LRE), adopts local Laplacian matrix to restore true data locality. Based on LRE model, three dimensionality reduction techniques are proposed. Experimental results on five public benchmark face datasets such as CMU PIE, FERET, ORL, Yale and FRGC, along with Nemenyi Post-hoc statistical of significant test attest the promising performance of the proposed techniques.

Original languageEnglish
Pages (from-to)86-102
Number of pages17
JournalPattern Recognition
Volume48
Issue number1
DOIs
Publication statusPublished - 2015 Jan 1

All Science Journal Classification (ASJC) codes

  • Software
  • Signal Processing
  • Computer Vision and Pattern Recognition
  • Artificial Intelligence

Cite this

Pang, Ying Han ; Teoh, Beng Jin ; Hiew, Fu San. / Locality Regularization Embedding for face verification. In: Pattern Recognition. 2015 ; Vol. 48, No. 1. pp. 86-102.
@article{2f7290fca1974ccbb6e40052f2bc7ab5,
title = "Locality Regularization Embedding for face verification",
abstract = "Graph embedding (GE) is a unified framework for dimensionality reduction techniques. GE attempts to maximally preserve data locality after embedding for face representation and classification. However, estimation of true data locality could be severely biased due to limited number of training samples, which trigger overfitting problem. In this paper, a graph embedding regularization technique is proposed to remedy this problem. The regularization model, dubbed as Locality Regularization Embedding (LRE), adopts local Laplacian matrix to restore true data locality. Based on LRE model, three dimensionality reduction techniques are proposed. Experimental results on five public benchmark face datasets such as CMU PIE, FERET, ORL, Yale and FRGC, along with Nemenyi Post-hoc statistical of significant test attest the promising performance of the proposed techniques.",
author = "Pang, {Ying Han} and Teoh, {Beng Jin} and Hiew, {Fu San}",
year = "2015",
month = "1",
day = "1",
doi = "10.1016/j.patcog.2014.07.010",
language = "English",
volume = "48",
pages = "86--102",
journal = "Pattern Recognition",
issn = "0031-3203",
publisher = "Elsevier Limited",
number = "1",

}

Locality Regularization Embedding for face verification. / Pang, Ying Han; Teoh, Beng Jin; Hiew, Fu San.

In: Pattern Recognition, Vol. 48, No. 1, 01.01.2015, p. 86-102.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Locality Regularization Embedding for face verification

AU - Pang, Ying Han

AU - Teoh, Beng Jin

AU - Hiew, Fu San

PY - 2015/1/1

Y1 - 2015/1/1

N2 - Graph embedding (GE) is a unified framework for dimensionality reduction techniques. GE attempts to maximally preserve data locality after embedding for face representation and classification. However, estimation of true data locality could be severely biased due to limited number of training samples, which trigger overfitting problem. In this paper, a graph embedding regularization technique is proposed to remedy this problem. The regularization model, dubbed as Locality Regularization Embedding (LRE), adopts local Laplacian matrix to restore true data locality. Based on LRE model, three dimensionality reduction techniques are proposed. Experimental results on five public benchmark face datasets such as CMU PIE, FERET, ORL, Yale and FRGC, along with Nemenyi Post-hoc statistical of significant test attest the promising performance of the proposed techniques.

AB - Graph embedding (GE) is a unified framework for dimensionality reduction techniques. GE attempts to maximally preserve data locality after embedding for face representation and classification. However, estimation of true data locality could be severely biased due to limited number of training samples, which trigger overfitting problem. In this paper, a graph embedding regularization technique is proposed to remedy this problem. The regularization model, dubbed as Locality Regularization Embedding (LRE), adopts local Laplacian matrix to restore true data locality. Based on LRE model, three dimensionality reduction techniques are proposed. Experimental results on five public benchmark face datasets such as CMU PIE, FERET, ORL, Yale and FRGC, along with Nemenyi Post-hoc statistical of significant test attest the promising performance of the proposed techniques.

UR - http://www.scopus.com/inward/record.url?scp=84908023884&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84908023884&partnerID=8YFLogxK

U2 - 10.1016/j.patcog.2014.07.010

DO - 10.1016/j.patcog.2014.07.010

M3 - Article

VL - 48

SP - 86

EP - 102

JO - Pattern Recognition

JF - Pattern Recognition

SN - 0031-3203

IS - 1

ER -