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 2019-03-04 15:45
Variational Bayes algorithm of region base coupled MRF with hidden phase variables
Naoki Wada (Tokyo Inst. of Tech.), Masaichiro Mizumaki (JASRI), Yoshiki Seno (Saga prefectural regional industry support center), Masato Okada (The Univ. of Tokyo), Akai Ichiro (Kumamoto Univ.), Toru Aonishi (Tokyo Inst. of Tech.) NC2018-59
Abstract (in Japanese) (See Japanese page) 
(in English) There are two methods in coupled Markov Random Field(MRF) model for image segmentation: edge-based method and region-based method. Region-based method is easier to implement and more robust to the noise than edge-based method. However, region-based method is often trapped to local optima. We focus on region base coupled MRF with hidden phase variables, which is reported to be less likely to be trapped to local optima. This model has already been implemented in a LSI circuit but has not been constructed probabilistic inference algorithm and evaluated its performance. We derive fast approximate inference algorithm using variational Bayes and compare with the conventional Ising spin model. Then, we have found that the hidden phase variables model is less likely to be trapped at local optima than the conventinal method. Moreover, we compare the variational method with Markov chain Monte Carlo to verify approximation accuracy of the variational method. Furthermore, we conducted mesoscopic structure detection as an application of image segmentation. Hidden phase variables model provides us with good result in real data.
Keyword (in Japanese) (See Japanese page) 
(in English) MRF / Variational Bayes / MCMC / image segmentation / / / /  
Reference Info. IEICE Tech. Rep., vol. 118, no. 470, NC2018-59, pp. 87-92, March 2019.
Paper # NC2018-59 
Date of Issue 2019-02-25 (NC) 
ISSN Print edition: ISSN 0913-5685  Online edition: ISSN 2432-6380
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 NC2018-59

Conference Information
Committee NC MBE  
Conference Date 2019-03-04 - 2019-03-06 
Place (in Japanese) (See Japanese page) 
Place (in English) University of Electro Communications 
Topics (in Japanese) (See Japanese page) 
Topics (in English)  
Paper Information
Registration To NC 
Conference Code 2019-03-NC-MBE 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) Variational Bayes algorithm of region base coupled MRF with hidden phase variables 
Sub Title (in English)  
Keyword(1) MRF  
Keyword(2) Variational Bayes  
Keyword(3) MCMC  
Keyword(4) image segmentation  
1st Author's Name Naoki Wada  
1st Author's Affiliation Tokyo Institute of Technology (Tokyo Inst. of Tech.)
2nd Author's Name Masaichiro Mizumaki  
2nd Author's Affiliation Japan Synchrotron Radiation Research Institute (JASRI)
3rd Author's Name Yoshiki Seno  
3rd Author's Affiliation Saga prefectural regional industry support center (Saga prefectural regional industry support center)
4th Author's Name Masato Okada  
4th Author's Affiliation The University of Tokyo (The Univ. of Tokyo)
5th Author's Name Akai Ichiro  
5th Author's Affiliation Kumamoto University (Kumamoto Univ.)
6th Author's Name Toru Aonishi  
6th Author's Affiliation Tokyo Institute of Technology (Tokyo Inst. of Tech.)
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 ()
Date Time 2019-03-04 15:45:00 
Presentation Time 25 
Registration for NC 
Paper # IEICE-NC2018-59 
Volume (vol) IEICE-118 
Number (no) no.470 
Page pp.87-92 
#Pages IEICE-6 
Date of Issue IEICE-NC-2019-02-25 

[Return to Top Page]

[Return to IEICE Web Page]

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