IEICE Technical Committee Submission System
Conference Paper's Information
Online Proceedings
[Sign in]
... (for ESS/CS/ES/ISS)
Tech. Rep. Archives
... (for ES/CS)
 Go Top Page Go Previous   [Japanese] / [English] 

Paper Abstract and Keywords
Presentation 2017-02-18 14:10
Fine-grained Pedestrian Classification by Fusing Multiple CNN Models
Yoshihito Kokubo, Yu Wang, Jien Kato, Kenji Mase (Nagoya Univ.) PRMU2016-166 CNR2016-33
Abstract (in Japanese) (See Japanese page) 
(in English) Fine-grained categories of pedestrians lead to higher level understanding on pedestrians as well as their activities, therefore is a perspective key component in the next generation smart vehicle and surveillance systems. In this study, we propose fine-grained pedestrian classification method by integrating features obtained from partial image and whole image using Convolutional Neural Network (CNN), which has dramatically improved results in various fields in recent years. Generally, when the partial image is given as an input to the network, detailed features can be expressed for the local portion, but the entire information is missing. On the other hand, when the entire image is given as an input to the network, it is possible to consider the appearance of the entire image, but detailed features can not be acquired compared with learning using a partial image. In the proposed method, a CNN model is learned for each patch of partial images and an entire image. Next, by re-training the fusion network in which all CNN models are integrated, the proposed method can eliminate mutual disadvantages and implement more accurate classification. In the experiment, we confirmed that the proposed method exceeds the baseline accuracy.
Keyword (in Japanese) (See Japanese page) 
(in English) image classification / deep learning / / / / / /  
Reference Info. IEICE Tech. Rep., vol. 116, no. 461, PRMU2016-166, pp. 81-85, Feb. 2017.
Paper # PRMU2016-166 
Date of Issue 2017-02-11 (PRMU, CNR) 
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. (No. 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034)
Download PDF PRMU2016-166 CNR2016-33

Conference Information
Committee PRMU CNR  
Conference Date 2017-02-18 - 2017-02-19 
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 2017-02-PRMU-CNR 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) Fine-grained Pedestrian Classification by Fusing Multiple CNN Models 
Sub Title (in English)  
Keyword(1) image classification  
Keyword(2) deep learning  
Keyword(3)  
Keyword(4)  
Keyword(5)  
Keyword(6)  
Keyword(7)  
Keyword(8)  
1st Author's Name Yoshihito Kokubo  
1st Author's Affiliation Nagoya University (Nagoya Univ.)
2nd Author's Name Yu Wang  
2nd Author's Affiliation Nagoya University (Nagoya Univ.)
3rd Author's Name Jien Kato  
3rd Author's Affiliation Nagoya University (Nagoya Univ.)
4th Author's Name Kenji Mase  
4th Author's Affiliation Nagoya University (Nagoya Univ.)
5th Author's Name  
5th Author's Affiliation ()
6th Author's Name  
6th Author's Affiliation ()
7th Author's Name  
7th Author's Affiliation ()
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
Date Time 2017-02-18 14:10:00 
Presentation Time 25 
Registration for PRMU 
Paper # IEICE-PRMU2016-166,IEICE-CNR2016-33 
Volume (vol) IEICE-116 
Number (no) no.461(PRMU), no.462(CNR) 
Page pp.81-85 
#Pages IEICE-5 
Date of Issue IEICE-PRMU-2017-02-11,IEICE-CNR-2017-02-11 


[Return to Top Page]

[Return to IEICE Web Page]


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