The intermediate matching kernel for image local features

Sabri Boughorbel, Jean Philippe Tarel, Nozha Boujemaa

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

32 Citations (Scopus)

Abstract

We introduce the Intermediate Matching (IM) kernel for SVM-based object recognition. The IM kernel operates on a feature space of vector sets where each image is represented by a set of local features. Matching algorithms have proved to be efficient for such types of features. Nevertheless, kernelizing the matching for SVM does not lead to positive definite kernels. The IM kernel overcomes this drawback, as it mimics matching algorithms while being positive definite. The IM kernel introduces an intermediary set of so-called virtual local features. These select the pairs of local features to be matched. Comparisons with the Matching kernel shows that the IM kernels leads to similar performances.

Original languageEnglish
Title of host publicationProceedings of the International Joint Conference on Neural Networks, IJCNN 2005
Pages889-894
Number of pages6
Volume2
DOIs
Publication statusPublished - 2005
Externally publishedYes
EventInternational Joint Conference on Neural Networks, IJCNN 2005 - Montreal, QC, Canada
Duration: 31 Jul 20054 Aug 2005

Other

OtherInternational Joint Conference on Neural Networks, IJCNN 2005
CountryCanada
CityMontreal, QC
Period31/7/054/8/05

Fingerprint

Object recognition

ASJC Scopus subject areas

  • Software

Cite this

Boughorbel, S., Tarel, J. P., & Boujemaa, N. (2005). The intermediate matching kernel for image local features. In Proceedings of the International Joint Conference on Neural Networks, IJCNN 2005 (Vol. 2, pp. 889-894). [1555970] https://doi.org/10.1109/IJCNN.2005.1555970

The intermediate matching kernel for image local features. / Boughorbel, Sabri; Tarel, Jean Philippe; Boujemaa, Nozha.

Proceedings of the International Joint Conference on Neural Networks, IJCNN 2005. Vol. 2 2005. p. 889-894 1555970.

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

Boughorbel, S, Tarel, JP & Boujemaa, N 2005, The intermediate matching kernel for image local features. in Proceedings of the International Joint Conference on Neural Networks, IJCNN 2005. vol. 2, 1555970, pp. 889-894, International Joint Conference on Neural Networks, IJCNN 2005, Montreal, QC, Canada, 31/7/05. https://doi.org/10.1109/IJCNN.2005.1555970
Boughorbel S, Tarel JP, Boujemaa N. The intermediate matching kernel for image local features. In Proceedings of the International Joint Conference on Neural Networks, IJCNN 2005. Vol. 2. 2005. p. 889-894. 1555970 https://doi.org/10.1109/IJCNN.2005.1555970
Boughorbel, Sabri ; Tarel, Jean Philippe ; Boujemaa, Nozha. / The intermediate matching kernel for image local features. Proceedings of the International Joint Conference on Neural Networks, IJCNN 2005. Vol. 2 2005. pp. 889-894
@inproceedings{b6cbf86e27de4aa789205726b272dfb4,
title = "The intermediate matching kernel for image local features",
abstract = "We introduce the Intermediate Matching (IM) kernel for SVM-based object recognition. The IM kernel operates on a feature space of vector sets where each image is represented by a set of local features. Matching algorithms have proved to be efficient for such types of features. Nevertheless, kernelizing the matching for SVM does not lead to positive definite kernels. The IM kernel overcomes this drawback, as it mimics matching algorithms while being positive definite. The IM kernel introduces an intermediary set of so-called virtual local features. These select the pairs of local features to be matched. Comparisons with the Matching kernel shows that the IM kernels leads to similar performances.",
author = "Sabri Boughorbel and Tarel, {Jean Philippe} and Nozha Boujemaa",
year = "2005",
doi = "10.1109/IJCNN.2005.1555970",
language = "English",
isbn = "0780390482",
volume = "2",
pages = "889--894",
booktitle = "Proceedings of the International Joint Conference on Neural Networks, IJCNN 2005",

}

TY - GEN

T1 - The intermediate matching kernel for image local features

AU - Boughorbel, Sabri

AU - Tarel, Jean Philippe

AU - Boujemaa, Nozha

PY - 2005

Y1 - 2005

N2 - We introduce the Intermediate Matching (IM) kernel for SVM-based object recognition. The IM kernel operates on a feature space of vector sets where each image is represented by a set of local features. Matching algorithms have proved to be efficient for such types of features. Nevertheless, kernelizing the matching for SVM does not lead to positive definite kernels. The IM kernel overcomes this drawback, as it mimics matching algorithms while being positive definite. The IM kernel introduces an intermediary set of so-called virtual local features. These select the pairs of local features to be matched. Comparisons with the Matching kernel shows that the IM kernels leads to similar performances.

AB - We introduce the Intermediate Matching (IM) kernel for SVM-based object recognition. The IM kernel operates on a feature space of vector sets where each image is represented by a set of local features. Matching algorithms have proved to be efficient for such types of features. Nevertheless, kernelizing the matching for SVM does not lead to positive definite kernels. The IM kernel overcomes this drawback, as it mimics matching algorithms while being positive definite. The IM kernel introduces an intermediary set of so-called virtual local features. These select the pairs of local features to be matched. Comparisons with the Matching kernel shows that the IM kernels leads to similar performances.

UR - http://www.scopus.com/inward/record.url?scp=33745963554&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33745963554&partnerID=8YFLogxK

U2 - 10.1109/IJCNN.2005.1555970

DO - 10.1109/IJCNN.2005.1555970

M3 - Conference contribution

SN - 0780390482

SN - 9780780390485

VL - 2

SP - 889

EP - 894

BT - Proceedings of the International Joint Conference on Neural Networks, IJCNN 2005

ER -