IEICE Technical Committee Submission System
Conference Paper's Information
Online Proceedings
[Sign in]
Tech. Rep. Archives
 Go Top Page Go Previous   [Japanese] / [English] 

Paper Abstract and Keywords
Presentation 2012-09-03 09:30
Simultaneous training of multi-class object detectors via large scale image dataset -- introduction of target specific negative classes --
Asako Kanezaki, Sho Inaba, Yoshitaka Ushiku, Yuya Yamashita, Hiroshi Muraoka, Tatsuya Harada, Yasuo Kuniyoshi (Univ. Tokyo) PRMU2012-42 IBISML2012-25
Abstract (in Japanese) (See Japanese page) 
(in English) We propose an efficient method to train multiple object detectors simultaneously using a large-scale image dataset. The one-vs-all approach that optimizes the boundary between positive samples from a target class and negative samples from the others has been the most standard approach for object detection. However, because this approach trains each object detector independently, the likelihoods are not balanced between object classes.
The proposed method combines ideas derived from both detection and classification in order to balance the scores across all object classes. We optimized the boundary between target classes and their hard-negative samples, just as in detection, while simultaneously balancing the detector likelihoods across object classes, as done in multi-class classification. We evaluated the performances on multi-class object detection using a subset of the ImageNet Large Scale Visual Recognition Challenge (ILSVRC) 2011 dataset and showed our method outperformed a de facto standard method.
Keyword (in Japanese) (See Japanese page) 
(in English) object detection / multiclass classification / large-scale image dataset / online learning / / / /  
Reference Info. IEICE Tech. Rep., vol. 112, no. 197, PRMU2012-42, pp. 105-112, Sept. 2012.
Paper # PRMU2012-42 
Date of Issue 2012-08-26 (PRMU, IBISML) 
ISSN Print edition: ISSN 0913-5685    Online edition: ISSN 2432-6380
Copyright
and
reproduction
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034)
Download PDF PRMU2012-42 IBISML2012-25

Conference Information
Committee PRMU IBISML IPSJ-CVIM  
Conference Date 2012-09-02 - 2012-09-03 
Place (in Japanese) (See Japanese page) 
Place (in English)  
Topics (in Japanese) (See Japanese page) 
Topics (in English)  
Paper Information
Registration To PRMU 
Conference Code 2012-09-PRMU-IBISML-CVIM 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) Simultaneous training of multi-class object detectors via large scale image dataset 
Sub Title (in English) introduction of target specific negative classes 
Keyword(1) object detection  
Keyword(2) multiclass classification  
Keyword(3) large-scale image dataset  
Keyword(4) online learning  
Keyword(5)  
Keyword(6)  
Keyword(7)  
Keyword(8)  
1st Author's Name Asako Kanezaki  
1st Author's Affiliation The University of Tokyo (Univ. Tokyo)
2nd Author's Name Sho Inaba  
2nd Author's Affiliation The University of Tokyo (Univ. Tokyo)
3rd Author's Name Yoshitaka Ushiku  
3rd Author's Affiliation The University of Tokyo (Univ. Tokyo)
4th Author's Name Yuya Yamashita  
4th Author's Affiliation The University of Tokyo (Univ. Tokyo)
5th Author's Name Hiroshi Muraoka  
5th Author's Affiliation The University of Tokyo (Univ. Tokyo)
6th Author's Name Tatsuya Harada  
6th Author's Affiliation The University of Tokyo (Univ. Tokyo)
7th Author's Name Yasuo Kuniyoshi  
7th Author's Affiliation The University of Tokyo (Univ. Tokyo)
8th Author's Name  
8th Author's Affiliation ()
9th Author's Name  
9th Author's Affiliation ()
10th Author's Name  
10th Author's Affiliation ()
11th Author's Name  
11th Author's Affiliation ()
12th Author's Name  
12th Author's Affiliation ()
13th Author's Name  
13th Author's Affiliation ()
14th Author's Name  
14th Author's Affiliation ()
15th Author's Name  
15th Author's Affiliation ()
16th Author's Name  
16th Author's Affiliation ()
17th Author's Name  
17th Author's Affiliation ()
18th Author's Name  
18th Author's Affiliation ()
19th Author's Name  
19th Author's Affiliation ()
20th Author's Name  
20th Author's Affiliation ()
Speaker Author-1 
Date Time 2012-09-03 09:30:00 
Presentation Time 30 minutes 
Registration for PRMU 
Paper # PRMU2012-42, IBISML2012-25 
Volume (vol) vol.112 
Number (no) no.197(PRMU), no.198(IBISML) 
Page pp.105-112 
#Pages
Date of Issue 2012-08-26 (PRMU, IBISML) 


[Return to Top Page]

[Return to IEICE Web Page]


The Institute of Electronics, Information and Communication Engineers (IEICE), Japan