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 2022-01-26 10:13
[Short Paper] Abnormality Detection for Covid-19 Chest CT Images by Dimensionality Reduction Based on Contrastive Learning
Hiroki Tobise, Kugler Mauricio, Tatsuya Yokota (NITech), Masahiro Hashimoto (Keio Univ.), Yoshito Otake (NAIST), Toshiaki Akashi (Juntendo Univ.), Akinobu Shimizu (TUAT), Hidekata Hontani (NITech) MI2021-53
Abstract (in Japanese) (See Japanese page) 
(in English) In this article, we propose a method that detects anomaly regions in chest CT images for the aid of Covid-19 diagnosis. Employing an approach for constructing a 1-class classifier based on the probability distribution of patch images of normal cases, we can relax the unbalance of the training data between different classes. The probability distribution should be estimated not in the patch image space but in a low-dimensional space in which we can estimate the similarity between patch images by referring to the Euclid distance between them. We therefore employ a contrastive-loss-based self-supervised learning method for the dimensionality reduction. The contrastive-loss is useful for realizing the projection invariant to the operations defined by users. We obtain a projection of patch images that is invariant against translation and flipping. Some experimental results are reported in this presentation.
Keyword (in Japanese) (See Japanese page) 
(in English) chest CT images / Covid-19 / anomaly detection / neural network / Contrastive Learning / Normalizing Flows / /  
Reference Info. IEICE Tech. Rep., vol. 121, no. 347, MI2021-53, pp. 41-42, Jan. 2022.
Paper # MI2021-53 
Date of Issue 2022-01-18 (MI) 
ISSN 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 MI2021-53

Conference Information
Committee MI  
Conference Date 2022-01-25 - 2022-01-27 
Place (in Japanese) (See Japanese page) 
Place (in English) Online 
Topics (in Japanese) (See Japanese page) 
Topics (in English)  
Paper Information
Registration To MI 
Conference Code 2022-01-MI 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) Abnormality Detection for Covid-19 Chest CT Images by Dimensionality Reduction Based on Contrastive Learning 
Sub Title (in English)  
Keyword(1) chest CT images  
Keyword(2) Covid-19  
Keyword(3) anomaly detection  
Keyword(4) neural network  
Keyword(5) Contrastive Learning  
Keyword(6) Normalizing Flows  
Keyword(7)  
Keyword(8)  
1st Author's Name Hiroki Tobise  
1st Author's Affiliation Nagoya Institute of Technology (NITech)
2nd Author's Name Kugler Mauricio  
2nd Author's Affiliation Nagoya Institute of Technology (NITech)
3rd Author's Name Tatsuya Yokota  
3rd Author's Affiliation Nagoya Institute of Technology (NITech)
4th Author's Name Masahiro Hashimoto  
4th Author's Affiliation Keio University (Keio Univ.)
5th Author's Name Yoshito Otake  
5th Author's Affiliation Nara Institute of Science and Technology (NAIST)
6th Author's Name Toshiaki Akashi  
6th Author's Affiliation Juntendo University (Juntendo Univ.)
7th Author's Name Akinobu Shimizu  
7th Author's Affiliation Tokyo University of Agriculture and Technology (TUAT)
8th Author's Name Hidekata Hontani  
8th Author's Affiliation Nagoya Institute of Technology (NITech)
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 2022-01-26 10:13:00 
Presentation Time 13 minutes 
Registration for MI 
Paper # MI2021-53 
Volume (vol) vol.121 
Number (no) no.347 
Page pp.41-42 
#Pages
Date of Issue 2022-01-18 (MI) 


[Return to Top Page]

[Return to IEICE Web Page]


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