Presentation 2011-12-16
Robust Image Matching by Multiple LS-SVMs
Toshitaka SHIMIZU, Hajimu KAWAKAMI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) This paper describes a method of designing robust classifiers of multi-class image data with multiple LS-SVMs. We first introduce an idea of the ECOC (error-correcting output code) to cause the multiple LS-SVMs to be robust. We then propose a learning method for the multiple LS-SVMs with considering the properties of the ECOC and a method of identifying multi-class image data with the multiple LS-SVMs in consideration of the properties inherent in the LS-SVM. To show effectiveness in the proposed methods, we made experiments of identifying 4-class image data in the presence of noise. The experimental results show the proposed methods possess robustness in the presence of noise.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Robust image recognition / LS-SVM / multiple classifier systems / error-correcting output code
Paper # PRMU2011-143
Date of Issue

Conference Information
Committee PRMU
Conference Date 2011/12/8(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Pattern Recognition and Media Understanding (PRMU)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Robust Image Matching by Multiple LS-SVMs
Sub Title (in English)
Keyword(1) Robust image recognition
Keyword(2) LS-SVM
Keyword(3) multiple classifier systems
Keyword(4) error-correcting output code
1st Author's Name Toshitaka SHIMIZU
1st Author's Affiliation Graduate School of Science and Technology, Ryukoku University()
2nd Author's Name Hajimu KAWAKAMI
2nd Author's Affiliation Faculty of Science and Technology, Ryukoku University
Date 2011-12-16
Paper # PRMU2011-143
Volume (vol) vol.111
Number (no) 353
Page pp.pp.-
#Pages 6
Date of Issue