Kernel ICA: An alternative formulation and its application to face recognition

Jian Yang; Xiumei Gao; Dapeng Zhang; Jing Yu Yang

doi:10.1016/j.patcog.2005.01.023

Kernel ICA: An alternative formulation and its application to face recognition

Jian Yang, Xiumei Gao, Dapeng Zhang, Jing Yu Yang

Research output: Journal article publication › Journal article › Academic research › peer-review

112 Citations (Scopus)

Abstract

This paper formulates independent component analysis (ICA) in the kernel-inducing feature space and develops a two-phase kernel ICA algorithm: whitened kernel principal component analysis (KPCA) plus ICA. KPCA spheres data and makes the data structure become as linearly separable as possible by virtue of an implicit nonlinear mapping determined by kernel. ICA seeks the projection directions in the KPCA whitened space, making the distribution of the projected data as non-gaussian as possible. The experiment using a subset of FERET database indicates that the proposed kernel ICA method significantly outperform ICA, PCA and KPCA in terms of the total recognition rate.

Original language	English
Pages (from-to)	1784-1787
Number of pages	4
Journal	Pattern Recognition
Volume	38
Issue number	10
DOIs	https://doi.org/10.1016/j.patcog.2005.01.023
Publication status	Published - 1 Oct 2005

Keywords

Face recognition
Feature extraction
Independent component analysis (ICA)
Kernel-based methods
Principal component analysis (PCA)

ASJC Scopus subject areas

Software
Signal Processing
Computer Vision and Pattern Recognition
Artificial Intelligence

More information

10.1016/j.patcog.2005.01.023

Cite this

@article{f445a0bbd6604591a49b9ccf5840b36e,

title = "Kernel ICA: An alternative formulation and its application to face recognition",

abstract = "This paper formulates independent component analysis (ICA) in the kernel-inducing feature space and develops a two-phase kernel ICA algorithm: whitened kernel principal component analysis (KPCA) plus ICA. KPCA spheres data and makes the data structure become as linearly separable as possible by virtue of an implicit nonlinear mapping determined by kernel. ICA seeks the projection directions in the KPCA whitened space, making the distribution of the projected data as non-gaussian as possible. The experiment using a subset of FERET database indicates that the proposed kernel ICA method significantly outperform ICA, PCA and KPCA in terms of the total recognition rate.",

keywords = "Face recognition, Feature extraction, Independent component analysis (ICA), Kernel-based methods, Principal component analysis (PCA)",

author = "Jian Yang and Xiumei Gao and Dapeng Zhang and Yang, {Jing Yu}",

year = "2005",

month = oct,

day = "1",

doi = "10.1016/j.patcog.2005.01.023",

language = "English",

volume = "38",

pages = "1784--1787",

journal = "Pattern Recognition",

issn = "0031-3203",

publisher = "Elsevier Ltd",

number = "10",

}

TY - JOUR

T1 - Kernel ICA: An alternative formulation and its application to face recognition

AU - Yang, Jian

AU - Gao, Xiumei

AU - Zhang, Dapeng

AU - Yang, Jing Yu

PY - 2005/10/1

Y1 - 2005/10/1

N2 - This paper formulates independent component analysis (ICA) in the kernel-inducing feature space and develops a two-phase kernel ICA algorithm: whitened kernel principal component analysis (KPCA) plus ICA. KPCA spheres data and makes the data structure become as linearly separable as possible by virtue of an implicit nonlinear mapping determined by kernel. ICA seeks the projection directions in the KPCA whitened space, making the distribution of the projected data as non-gaussian as possible. The experiment using a subset of FERET database indicates that the proposed kernel ICA method significantly outperform ICA, PCA and KPCA in terms of the total recognition rate.

AB - This paper formulates independent component analysis (ICA) in the kernel-inducing feature space and develops a two-phase kernel ICA algorithm: whitened kernel principal component analysis (KPCA) plus ICA. KPCA spheres data and makes the data structure become as linearly separable as possible by virtue of an implicit nonlinear mapping determined by kernel. ICA seeks the projection directions in the KPCA whitened space, making the distribution of the projected data as non-gaussian as possible. The experiment using a subset of FERET database indicates that the proposed kernel ICA method significantly outperform ICA, PCA and KPCA in terms of the total recognition rate.

KW - Face recognition

KW - Feature extraction

KW - Independent component analysis (ICA)

KW - Kernel-based methods

KW - Principal component analysis (PCA)

UR - http://www.scopus.com/inward/record.url?scp=22844447009&partnerID=8YFLogxK

U2 - 10.1016/j.patcog.2005.01.023

DO - 10.1016/j.patcog.2005.01.023

M3 - Journal article

SN - 0031-3203

VL - 38

SP - 1784

EP - 1787

JO - Pattern Recognition

JF - Pattern Recognition

IS - 10

ER -

Kernel ICA: An alternative formulation and its application to face recognition

Abstract

Keywords

ASJC Scopus subject areas

More information

Other files and links

Fingerprint

Cite this