The GCS kernel for SVM-based image recognition

Sabri Boughorbel, Jean Philippe Tarel, François Fleuret, Nozha Boujemaa

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

6 Citations (Scopus)

Abstract

In this paper, we present a new compactly supported kernel for SVM based image recognition. This kernel which we called Geometric Compactly Supported (GCS) can be viewed as a generalization of spherical kernels to higher dimensions. The construction of the GCS kernel is based on a geometric approach using the intersection volume of two n-dimensional balls. The compactness property of the GCS kernel leads to a sparse Gram matrix which enhances computation efficiency by using sparse linear algebra algorithms. Comparisons of the GCS kernel performance, for image recognition task, with other known kernels prove the interest of this new kernel.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages595-600
Number of pages6
Volume3697 LNCS
Publication statusPublished - 2005
Externally publishedYes
Event15th International Conference on Artificial Neural Networks: Biological Inspirations - ICANN 2005 - Warsaw, Poland
Duration: 11 Sep 200515 Sep 2005

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume3697 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other15th International Conference on Artificial Neural Networks: Biological Inspirations - ICANN 2005
CountryPoland
CityWarsaw
Period11/9/0515/9/05

Fingerprint

Image recognition
Image Recognition
kernel
Linear algebra
Gram Matrix
Geometric Approach
Sparse matrix
Higher Dimensions
Compactness
n-dimensional
Ball
Intersection

ASJC Scopus subject areas

  • Computer Science(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Theoretical Computer Science

Cite this

Boughorbel, S., Tarel, J. P., Fleuret, F., & Boujemaa, N. (2005). The GCS kernel for SVM-based image recognition. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 3697 LNCS, pp. 595-600). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 3697 LNCS).

The GCS kernel for SVM-based image recognition. / Boughorbel, Sabri; Tarel, Jean Philippe; Fleuret, François; Boujemaa, Nozha.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 3697 LNCS 2005. p. 595-600 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 3697 LNCS).

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

Boughorbel, S, Tarel, JP, Fleuret, F & Boujemaa, N 2005, The GCS kernel for SVM-based image recognition. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 3697 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 3697 LNCS, pp. 595-600, 15th International Conference on Artificial Neural Networks: Biological Inspirations - ICANN 2005, Warsaw, Poland, 11/9/05.
Boughorbel S, Tarel JP, Fleuret F, Boujemaa N. The GCS kernel for SVM-based image recognition. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 3697 LNCS. 2005. p. 595-600. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
Boughorbel, Sabri ; Tarel, Jean Philippe ; Fleuret, François ; Boujemaa, Nozha. / The GCS kernel for SVM-based image recognition. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 3697 LNCS 2005. pp. 595-600 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{6d2f6970d81f4435a1032ddebb692075,
title = "The GCS kernel for SVM-based image recognition",
abstract = "In this paper, we present a new compactly supported kernel for SVM based image recognition. This kernel which we called Geometric Compactly Supported (GCS) can be viewed as a generalization of spherical kernels to higher dimensions. The construction of the GCS kernel is based on a geometric approach using the intersection volume of two n-dimensional balls. The compactness property of the GCS kernel leads to a sparse Gram matrix which enhances computation efficiency by using sparse linear algebra algorithms. Comparisons of the GCS kernel performance, for image recognition task, with other known kernels prove the interest of this new kernel.",
author = "Sabri Boughorbel and Tarel, {Jean Philippe} and Fran{\cc}ois Fleuret and Nozha Boujemaa",
year = "2005",
language = "English",
isbn = "3540287558",
volume = "3697 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "595--600",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - The GCS kernel for SVM-based image recognition

AU - Boughorbel, Sabri

AU - Tarel, Jean Philippe

AU - Fleuret, François

AU - Boujemaa, Nozha

PY - 2005

Y1 - 2005

N2 - In this paper, we present a new compactly supported kernel for SVM based image recognition. This kernel which we called Geometric Compactly Supported (GCS) can be viewed as a generalization of spherical kernels to higher dimensions. The construction of the GCS kernel is based on a geometric approach using the intersection volume of two n-dimensional balls. The compactness property of the GCS kernel leads to a sparse Gram matrix which enhances computation efficiency by using sparse linear algebra algorithms. Comparisons of the GCS kernel performance, for image recognition task, with other known kernels prove the interest of this new kernel.

AB - In this paper, we present a new compactly supported kernel for SVM based image recognition. This kernel which we called Geometric Compactly Supported (GCS) can be viewed as a generalization of spherical kernels to higher dimensions. The construction of the GCS kernel is based on a geometric approach using the intersection volume of two n-dimensional balls. The compactness property of the GCS kernel leads to a sparse Gram matrix which enhances computation efficiency by using sparse linear algebra algorithms. Comparisons of the GCS kernel performance, for image recognition task, with other known kernels prove the interest of this new kernel.

UR - http://www.scopus.com/inward/record.url?scp=33646255226&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33646255226&partnerID=8YFLogxK

M3 - Conference contribution

SN - 3540287558

SN - 9783540287551

VL - 3697 LNCS

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

SP - 595

EP - 600

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

ER -