Human face analysis with nonlinear manifold learning

Xiao Kan Wang, Xia Mao, Ishizuka Mitsuru

Research output: Contribution to journalArticle

6 Citations (Scopus)

Abstract

Since human face movements distribute on a nonlinear manifold, there are inherent alignment residuals brought by the global linearity hypothesis in the traditional Principal Component Analysis (PCA) based Active Appearance Models (AAM). In this paper, a famous manifold learning method, Local Linear Embedding (LLE) is improved to model human face shape space for reducing the inherent alignment residuals. The experimental results show that the method, LLE-AAM, obtains lower alignment residuals to the tiny alterations of human face and still make successful alignment when PCA-AAM failed to some large alterations. According to the statistical analysis, LLE-AAM could reduce the residual to a certain extent.

Original languageEnglish
Pages (from-to)2531-2535
Number of pages5
JournalDianzi Yu Xinxi Xuebao/Journal of Electronics and Information Technology
Volume33
Issue number10
DOIs
Publication statusPublished - 2011 Oct
Externally publishedYes

Fingerprint

Principal component analysis
Statistical methods

Keywords

  • Active Appearance Models (AAM)
  • Face recognition
  • Local Linear Embedding (LLE)
  • Nonlinear manifold learning

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Human face analysis with nonlinear manifold learning. / Wang, Xiao Kan; Mao, Xia; Mitsuru, Ishizuka.

In: Dianzi Yu Xinxi Xuebao/Journal of Electronics and Information Technology, Vol. 33, No. 10, 10.2011, p. 2531-2535.

Research output: Contribution to journalArticle

Wang, Xiao Kan ; Mao, Xia ; Mitsuru, Ishizuka. / Human face analysis with nonlinear manifold learning. In: Dianzi Yu Xinxi Xuebao/Journal of Electronics and Information Technology. 2011 ; Vol. 33, No. 10. pp. 2531-2535.
@article{aace67982ab4474c9ad1d11693dc23bf,
title = "Human face analysis with nonlinear manifold learning",
abstract = "Since human face movements distribute on a nonlinear manifold, there are inherent alignment residuals brought by the global linearity hypothesis in the traditional Principal Component Analysis (PCA) based Active Appearance Models (AAM). In this paper, a famous manifold learning method, Local Linear Embedding (LLE) is improved to model human face shape space for reducing the inherent alignment residuals. The experimental results show that the method, LLE-AAM, obtains lower alignment residuals to the tiny alterations of human face and still make successful alignment when PCA-AAM failed to some large alterations. According to the statistical analysis, LLE-AAM could reduce the residual to a certain extent.",
keywords = "Active Appearance Models (AAM), Face recognition, Local Linear Embedding (LLE), Nonlinear manifold learning",
author = "Wang, {Xiao Kan} and Xia Mao and Ishizuka Mitsuru",
year = "2011",
month = "10",
doi = "10.3724/SP.J.1146.2011.00153",
language = "English",
volume = "33",
pages = "2531--2535",
journal = "Dianzi Yu Xinxi Xuebao/Journal of Electronics and Information Technology",
issn = "1009-5896",
publisher = "Science Press",
number = "10",

}

TY - JOUR

T1 - Human face analysis with nonlinear manifold learning

AU - Wang, Xiao Kan

AU - Mao, Xia

AU - Mitsuru, Ishizuka

PY - 2011/10

Y1 - 2011/10

N2 - Since human face movements distribute on a nonlinear manifold, there are inherent alignment residuals brought by the global linearity hypothesis in the traditional Principal Component Analysis (PCA) based Active Appearance Models (AAM). In this paper, a famous manifold learning method, Local Linear Embedding (LLE) is improved to model human face shape space for reducing the inherent alignment residuals. The experimental results show that the method, LLE-AAM, obtains lower alignment residuals to the tiny alterations of human face and still make successful alignment when PCA-AAM failed to some large alterations. According to the statistical analysis, LLE-AAM could reduce the residual to a certain extent.

AB - Since human face movements distribute on a nonlinear manifold, there are inherent alignment residuals brought by the global linearity hypothesis in the traditional Principal Component Analysis (PCA) based Active Appearance Models (AAM). In this paper, a famous manifold learning method, Local Linear Embedding (LLE) is improved to model human face shape space for reducing the inherent alignment residuals. The experimental results show that the method, LLE-AAM, obtains lower alignment residuals to the tiny alterations of human face and still make successful alignment when PCA-AAM failed to some large alterations. According to the statistical analysis, LLE-AAM could reduce the residual to a certain extent.

KW - Active Appearance Models (AAM)

KW - Face recognition

KW - Local Linear Embedding (LLE)

KW - Nonlinear manifold learning

UR - http://www.scopus.com/inward/record.url?scp=80555126930&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=80555126930&partnerID=8YFLogxK

U2 - 10.3724/SP.J.1146.2011.00153

DO - 10.3724/SP.J.1146.2011.00153

M3 - Article

VL - 33

SP - 2531

EP - 2535

JO - Dianzi Yu Xinxi Xuebao/Journal of Electronics and Information Technology

JF - Dianzi Yu Xinxi Xuebao/Journal of Electronics and Information Technology

SN - 1009-5896

IS - 10

ER -