Multilinear sparse decomposition for best spectral bands selection

Hamdi Jamel Bouchech, Sebti Foufou, Mongi Abidi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Optimal spectral bands selection is a primordial step in multispectral images based systems for face recognition. In this context, we select the best spectral bands using a multilinear sparse decomposition based approach. Multispectral images of 35 subjects presenting 25 different lengths from 480nm to 720nm and three lighting conditions: fluorescent, Halogen and Sun light are groupped in a 3-mode face tensor T of size 35x25x2 . T is then decomposed using 3-mode SVD where three mode matrices for subjects, spectral bands and illuminations are sparsely determined. The 25x25 spectral bands mode matrix defines a sparse vector for each spectral band. Spectral bands having the sparse vectors with the lowest variation with illumination are selected as the best spectral bands. Experiments on two state-of-the-art algorithms, MBLBP and HGPP, showed the effectiveness of our approach for best spectral bands selection.

Original languageEnglish (US)
Title of host publicationImage and Signal Processing - 6th International Conference, ICISP 2014, Proceedings
PublisherSpringer-Verlag
Pages384-391
Number of pages8
ISBN (Print)9783319079974
DOIs
StatePublished - Jan 1 2014
Event6th International Conference on Image and Signal Processing, ICISP 2014 - Cherbourg, France
Duration: Jun 30 2014Jul 2 2014

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8509 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other6th International Conference on Image and Signal Processing, ICISP 2014
CountryFrance
CityCherbourg
Period6/30/147/2/14

Fingerprint

Lighting
Decomposition
Decompose
Singular value decomposition
Face recognition
Sun
Multispectral Images
Tensors
Illumination
Experiments
Face Recognition
Lowest
Tensor
Face
Experiment

Keywords

  • HGPP
  • MBLBP
  • Multilinear
  • sparse
  • Spectral bands
  • Tensor

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Bouchech, H. J., Foufou, S., & Abidi, M. (2014). Multilinear sparse decomposition for best spectral bands selection. In Image and Signal Processing - 6th International Conference, ICISP 2014, Proceedings (pp. 384-391). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8509 LNCS). Springer-Verlag. https://doi.org/10.1007/978-3-319-07998-1_44

Multilinear sparse decomposition for best spectral bands selection. / Bouchech, Hamdi Jamel; Foufou, Sebti; Abidi, Mongi.

Image and Signal Processing - 6th International Conference, ICISP 2014, Proceedings. Springer-Verlag, 2014. p. 384-391 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8509 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Bouchech, HJ, Foufou, S & Abidi, M 2014, Multilinear sparse decomposition for best spectral bands selection. in Image and Signal Processing - 6th International Conference, ICISP 2014, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 8509 LNCS, Springer-Verlag, pp. 384-391, 6th International Conference on Image and Signal Processing, ICISP 2014, Cherbourg, France, 6/30/14. https://doi.org/10.1007/978-3-319-07998-1_44
Bouchech HJ, Foufou S, Abidi M. Multilinear sparse decomposition for best spectral bands selection. In Image and Signal Processing - 6th International Conference, ICISP 2014, Proceedings. Springer-Verlag. 2014. p. 384-391. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-07998-1_44
Bouchech, Hamdi Jamel ; Foufou, Sebti ; Abidi, Mongi. / Multilinear sparse decomposition for best spectral bands selection. Image and Signal Processing - 6th International Conference, ICISP 2014, Proceedings. Springer-Verlag, 2014. pp. 384-391 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{10713d3246e2421a86063b9a24482863,
title = "Multilinear sparse decomposition for best spectral bands selection",
abstract = "Optimal spectral bands selection is a primordial step in multispectral images based systems for face recognition. In this context, we select the best spectral bands using a multilinear sparse decomposition based approach. Multispectral images of 35 subjects presenting 25 different lengths from 480nm to 720nm and three lighting conditions: fluorescent, Halogen and Sun light are groupped in a 3-mode face tensor T of size 35x25x2 . T is then decomposed using 3-mode SVD where three mode matrices for subjects, spectral bands and illuminations are sparsely determined. The 25x25 spectral bands mode matrix defines a sparse vector for each spectral band. Spectral bands having the sparse vectors with the lowest variation with illumination are selected as the best spectral bands. Experiments on two state-of-the-art algorithms, MBLBP and HGPP, showed the effectiveness of our approach for best spectral bands selection.",
keywords = "HGPP, MBLBP, Multilinear, sparse, Spectral bands, Tensor",
author = "Bouchech, {Hamdi Jamel} and Sebti Foufou and Mongi Abidi",
year = "2014",
month = "1",
day = "1",
doi = "10.1007/978-3-319-07998-1_44",
language = "English (US)",
isbn = "9783319079974",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer-Verlag",
pages = "384--391",
booktitle = "Image and Signal Processing - 6th International Conference, ICISP 2014, Proceedings",

}

TY - GEN

T1 - Multilinear sparse decomposition for best spectral bands selection

AU - Bouchech, Hamdi Jamel

AU - Foufou, Sebti

AU - Abidi, Mongi

PY - 2014/1/1

Y1 - 2014/1/1

N2 - Optimal spectral bands selection is a primordial step in multispectral images based systems for face recognition. In this context, we select the best spectral bands using a multilinear sparse decomposition based approach. Multispectral images of 35 subjects presenting 25 different lengths from 480nm to 720nm and three lighting conditions: fluorescent, Halogen and Sun light are groupped in a 3-mode face tensor T of size 35x25x2 . T is then decomposed using 3-mode SVD where three mode matrices for subjects, spectral bands and illuminations are sparsely determined. The 25x25 spectral bands mode matrix defines a sparse vector for each spectral band. Spectral bands having the sparse vectors with the lowest variation with illumination are selected as the best spectral bands. Experiments on two state-of-the-art algorithms, MBLBP and HGPP, showed the effectiveness of our approach for best spectral bands selection.

AB - Optimal spectral bands selection is a primordial step in multispectral images based systems for face recognition. In this context, we select the best spectral bands using a multilinear sparse decomposition based approach. Multispectral images of 35 subjects presenting 25 different lengths from 480nm to 720nm and three lighting conditions: fluorescent, Halogen and Sun light are groupped in a 3-mode face tensor T of size 35x25x2 . T is then decomposed using 3-mode SVD where three mode matrices for subjects, spectral bands and illuminations are sparsely determined. The 25x25 spectral bands mode matrix defines a sparse vector for each spectral band. Spectral bands having the sparse vectors with the lowest variation with illumination are selected as the best spectral bands. Experiments on two state-of-the-art algorithms, MBLBP and HGPP, showed the effectiveness of our approach for best spectral bands selection.

KW - HGPP

KW - MBLBP

KW - Multilinear

KW - sparse

KW - Spectral bands

KW - Tensor

UR - http://www.scopus.com/inward/record.url?scp=84903641396&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84903641396&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-07998-1_44

DO - 10.1007/978-3-319-07998-1_44

M3 - Conference contribution

AN - SCOPUS:84903641396

SN - 9783319079974

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 384

EP - 391

BT - Image and Signal Processing - 6th International Conference, ICISP 2014, Proceedings

PB - Springer-Verlag

ER -