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 2020-03-05 16:45
[Poster Presentation] Detecting Adversarial Examples Based on Sensitivities to Lossy Compression Algorithms
Akinori Higashi, Minoru Kuribayashi, Nobuo Funabiki (Okayama Univ.), Huy Hong Nguyen, Isao Echizen (NII) EMM2019-123
Abstract (in Japanese) (See Japanese page) 
(in English) The adversarial examples are created by adding small perturbations to an input image for misleading an CNN-based image classifier into a wrong class, and it is difficult for humans to visually recognize the perturbation. In CSS 2019, we focused on the property that the discrimination class fluctuates by removing the tiny noise added to the adversarial examples using the lossy compression. And we proposed a method to detect the adversarial examples by counting the number of changes in the output of CNN classifier when compression quality is gradually reduced. In this study, we examined how the detecting accuracy of adversarial examples is affected by increasing the interval of changing compression quality in order to reduce the computational complexity. As a result, it was found that it is possible to detect the adversarial examples with high accuracy from the observation of a few patterns. As the scaling down operation removes the entropy in an input image, the effects of the scaling down and then scaling up are similar to that of lossy compression. Hence, we gradually change the scaling down parameter and count the number of changes in the output of CNN classifier. It is confirmed from experiments that the detection accuracy is also comparable to the case of JPEG compression.
Keyword (in Japanese) (See Japanese page) 
(in English) Adversarial Example / Lossy Compression / CNN / Image Classifier / / / /  
Reference Info. IEICE Tech. Rep., vol. 119, no. 463, EMM2019-123, pp. 113-116, March 2020.
Paper # EMM2019-123 
Date of Issue 2020-02-27 (EMM) 
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 EMM2019-123

Conference Information
Committee EMM  
Conference Date 2020-03-05 - 2020-03-06 
Place (in Japanese) (See Japanese page) 
Place (in English)  
Topics (in Japanese) (See Japanese page) 
Topics (in English) Image and Sound Quality, Metrics for Perception and Recognition, Human Auditory and Visual System, etc. 
Paper Information
Registration To EMM 
Conference Code 2020-03-EMM 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) Detecting Adversarial Examples Based on Sensitivities to Lossy Compression Algorithms 
Sub Title (in English)  
Keyword(1) Adversarial Example  
Keyword(2) Lossy Compression  
Keyword(3) CNN  
Keyword(4) Image Classifier  
Keyword(5)  
Keyword(6)  
Keyword(7)  
Keyword(8)  
1st Author's Name Akinori Higashi  
1st Author's Affiliation Okayama University (Okayama Univ.)
2nd Author's Name Minoru Kuribayashi  
2nd Author's Affiliation Okayama University (Okayama Univ.)
3rd Author's Name Nobuo Funabiki  
3rd Author's Affiliation Okayama University (Okayama Univ.)
4th Author's Name Huy Hong Nguyen  
4th Author's Affiliation National Institute of Informatics (NII)
5th Author's Name Isao Echizen  
5th Author's Affiliation National Institute of Informatics (NII)
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 Author-1 
Date Time 2020-03-05 16:45:00 
Presentation Time 60 minutes 
Registration for EMM 
Paper # EMM2019-123 
Volume (vol) vol.119 
Number (no) no.463 
Page pp.113-116 
#Pages
Date of Issue 2020-02-27 (EMM) 


[Return to Top Page]

[Return to IEICE Web Page]


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