Presentation 2010-09-05
Multiple Kernel Learning for Generic Object Recognition Using SIFT Gaussian Mixture Models
Nakamasa INOUE, Yusuke KAMISHIMA, Koichi SHINODA, Sadaoki FURUI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We propose a statistical framework for generic object recognition using SIFT Gaussian mixture models (GMMs) and multiple kernel learning. We model SIFT features extracted from images by using GMMs. RBF kernels for multiple kernel learning are generated based on the distance between SIFT GMMs. In our experiments, we show which type of regularizer (e.g. l_1, l_p, elastic-net) is optimal for combining the kernels for different features, interest regions, and vocabulary sizes. Our best result obtained by using a single kernel was 41.0% in terms of Mean AP on the PASCAL VOC 2010 dataset, and it was improved to 50.5% by using the elastic-net multiple kernel learning.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Generic object recognition / SIFT GMMs / GMM supervector / Multiple kernel learning
Paper # PRMU2010-58,IBISML2010-30
Date of Issue

Conference Information
Committee PRMU
Conference Date 2010/8/29(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) Multiple Kernel Learning for Generic Object Recognition Using SIFT Gaussian Mixture Models
Sub Title (in English)
Keyword(1) Generic object recognition
Keyword(2) SIFT GMMs
Keyword(3) GMM supervector
Keyword(4) Multiple kernel learning
1st Author's Name Nakamasa INOUE
1st Author's Affiliation Department of Computer Science, Tokyo Institute of Technology()
2nd Author's Name Yusuke KAMISHIMA
2nd Author's Affiliation Department of Computer Science, Tokyo Institute of Technology
3rd Author's Name Koichi SHINODA
3rd Author's Affiliation Department of Computer Science, Tokyo Institute of Technology
4th Author's Name Sadaoki FURUI
4th Author's Affiliation Department of Computer Science, Tokyo Institute of Technology
Date 2010-09-05
Paper # PRMU2010-58,IBISML2010-30
Volume (vol) vol.110
Number (no) 187
Page pp.pp.-
#Pages 6
Date of Issue