A kernelized sparsity-based approach for best spectral bands selection for face recognition

Hamdi Jamel Bouchech, Sebti Foufou, Andreas Koschan, Mongi Abidi

Research output: Contribution to journalArticle

Abstract

We study face recognition in unconstrained illumination conditions. A twofold contribution is proposed: First, the robustness of four state-of-the-art algorithms, namely Multi-block Local Binary Pattern (MBLBP), Histogram of Gabor Phase Patterns (HGPP), Local Gabor Binary Pattern Histogram Sequence (LGBPHS) and Patterns of Oriented Edge Magnitudes (POEM-WPCA) against high illumination variation is studied. Second, we propose to enhance the performance of the four mentioned algorithms, which has been drastically decreased upon the day lighted face images provided by IRIS-M3 face database. For this purpose, we use visible narrow band subspectral images selected from the mentioned database. We formulate best spectral bands selection as a pursuit optimization problem wherein the vector of weights determining the importance of each visible spectral band is supposed to be sparse, and hence can be determined by minimizing its L1-norm. Several fusing approaches are then applied on selected best spectral bands using multi-scale and multi-orientation Gabor wavelets. The results highlight further the still challenging problem of face recognition in conditions with high illumination variation, as well as the effectiveness of our subspectral images based approach with its two components; bands selection and bands fusion, to increase the accuracy of the studied algorithms by at least 14 % upon the proposed database.

Original languageEnglish (US)
Pages (from-to)8631-8654
Number of pages24
JournalMultimedia Tools and Applications
Volume74
Issue number19
DOIs
StatePublished - Oct 31 2015

Fingerprint

Face recognition
Lighting
Fusion reactions

Keywords

  • Bands selection
  • HGPP
  • IRIS-M
  • LGBPHS
  • MBLBP
  • POEM-WPCA
  • Subspectral images

ASJC Scopus subject areas

  • Software
  • Media Technology
  • Hardware and Architecture
  • Computer Networks and Communications

Cite this

A kernelized sparsity-based approach for best spectral bands selection for face recognition. / Bouchech, Hamdi Jamel; Foufou, Sebti; Koschan, Andreas; Abidi, Mongi.

In: Multimedia Tools and Applications, Vol. 74, No. 19, 31.10.2015, p. 8631-8654.

Research output: Contribution to journalArticle

Bouchech, Hamdi Jamel ; Foufou, Sebti ; Koschan, Andreas ; Abidi, Mongi. / A kernelized sparsity-based approach for best spectral bands selection for face recognition. In: Multimedia Tools and Applications. 2015 ; Vol. 74, No. 19. pp. 8631-8654.
@article{dc8b9292825f486a86545bc948e1115e,
title = "A kernelized sparsity-based approach for best spectral bands selection for face recognition",
abstract = "We study face recognition in unconstrained illumination conditions. A twofold contribution is proposed: First, the robustness of four state-of-the-art algorithms, namely Multi-block Local Binary Pattern (MBLBP), Histogram of Gabor Phase Patterns (HGPP), Local Gabor Binary Pattern Histogram Sequence (LGBPHS) and Patterns of Oriented Edge Magnitudes (POEM-WPCA) against high illumination variation is studied. Second, we propose to enhance the performance of the four mentioned algorithms, which has been drastically decreased upon the day lighted face images provided by IRIS-M3 face database. For this purpose, we use visible narrow band subspectral images selected from the mentioned database. We formulate best spectral bands selection as a pursuit optimization problem wherein the vector of weights determining the importance of each visible spectral band is supposed to be sparse, and hence can be determined by minimizing its L1-norm. Several fusing approaches are then applied on selected best spectral bands using multi-scale and multi-orientation Gabor wavelets. The results highlight further the still challenging problem of face recognition in conditions with high illumination variation, as well as the effectiveness of our subspectral images based approach with its two components; bands selection and bands fusion, to increase the accuracy of the studied algorithms by at least 14 {\%} upon the proposed database.",
keywords = "Bands selection, HGPP, IRIS-M, LGBPHS, MBLBP, POEM-WPCA, Subspectral images",
author = "Bouchech, {Hamdi Jamel} and Sebti Foufou and Andreas Koschan and Mongi Abidi",
year = "2015",
month = "10",
day = "31",
doi = "10.1007/s11042-014-2350-2",
language = "English (US)",
volume = "74",
pages = "8631--8654",
journal = "Multimedia Tools and Applications",
issn = "1380-7501",
publisher = "Springer Netherlands",
number = "19",

}

TY - JOUR

T1 - A kernelized sparsity-based approach for best spectral bands selection for face recognition

AU - Bouchech, Hamdi Jamel

AU - Foufou, Sebti

AU - Koschan, Andreas

AU - Abidi, Mongi

PY - 2015/10/31

Y1 - 2015/10/31

N2 - We study face recognition in unconstrained illumination conditions. A twofold contribution is proposed: First, the robustness of four state-of-the-art algorithms, namely Multi-block Local Binary Pattern (MBLBP), Histogram of Gabor Phase Patterns (HGPP), Local Gabor Binary Pattern Histogram Sequence (LGBPHS) and Patterns of Oriented Edge Magnitudes (POEM-WPCA) against high illumination variation is studied. Second, we propose to enhance the performance of the four mentioned algorithms, which has been drastically decreased upon the day lighted face images provided by IRIS-M3 face database. For this purpose, we use visible narrow band subspectral images selected from the mentioned database. We formulate best spectral bands selection as a pursuit optimization problem wherein the vector of weights determining the importance of each visible spectral band is supposed to be sparse, and hence can be determined by minimizing its L1-norm. Several fusing approaches are then applied on selected best spectral bands using multi-scale and multi-orientation Gabor wavelets. The results highlight further the still challenging problem of face recognition in conditions with high illumination variation, as well as the effectiveness of our subspectral images based approach with its two components; bands selection and bands fusion, to increase the accuracy of the studied algorithms by at least 14 % upon the proposed database.

AB - We study face recognition in unconstrained illumination conditions. A twofold contribution is proposed: First, the robustness of four state-of-the-art algorithms, namely Multi-block Local Binary Pattern (MBLBP), Histogram of Gabor Phase Patterns (HGPP), Local Gabor Binary Pattern Histogram Sequence (LGBPHS) and Patterns of Oriented Edge Magnitudes (POEM-WPCA) against high illumination variation is studied. Second, we propose to enhance the performance of the four mentioned algorithms, which has been drastically decreased upon the day lighted face images provided by IRIS-M3 face database. For this purpose, we use visible narrow band subspectral images selected from the mentioned database. We formulate best spectral bands selection as a pursuit optimization problem wherein the vector of weights determining the importance of each visible spectral band is supposed to be sparse, and hence can be determined by minimizing its L1-norm. Several fusing approaches are then applied on selected best spectral bands using multi-scale and multi-orientation Gabor wavelets. The results highlight further the still challenging problem of face recognition in conditions with high illumination variation, as well as the effectiveness of our subspectral images based approach with its two components; bands selection and bands fusion, to increase the accuracy of the studied algorithms by at least 14 % upon the proposed database.

KW - Bands selection

KW - HGPP

KW - IRIS-M

KW - LGBPHS

KW - MBLBP

KW - POEM-WPCA

KW - Subspectral images

UR - http://www.scopus.com/inward/record.url?scp=84940595990&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84940595990&partnerID=8YFLogxK

U2 - 10.1007/s11042-014-2350-2

DO - 10.1007/s11042-014-2350-2

M3 - Article

AN - SCOPUS:84940595990

VL - 74

SP - 8631

EP - 8654

JO - Multimedia Tools and Applications

JF - Multimedia Tools and Applications

SN - 1380-7501

IS - 19

ER -