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 2021-11-15 09:00
[Poster Presentation] A consideration of training datasets for universal detectors of CNN-generated images
Miki Tanaka, Hitoshi Kiya (Tokyo Metro. Univ.) EA2021-32 EMM2021-59
Abstract (in Japanese) (See Japanese page) 
(in English) Recent rapid advances in convolutional neural networks (CNNs) have made manipulating and generating images easy, so synthetic media have been spread rapidly. In contrast, CNN-generated images have generated a new problem that is to make more complicated to judge images to be either real images or ones generated with CNNs. In this paper, we aim to improve a universal detector of CNN-generated images. It is difficult to train a detector by using images generated from all CNN-generation models because there are a variety of CNN-generated models. In this paper, CNN-generated models are classified in accordance with the properties that existing CNN models have, and a universal detector is then trained by using images selected from each class. In addition, from an ensemble of RGB images and spectrum images is also proposed to enhance the performance of the proposed universal detector.
Keyword (in Japanese) (See Japanese page) 
(in English) GAN / CNN / checkerboard artifact / fake image / / / /  
Reference Info. IEICE Tech. Rep., vol. 121, no. 247, EMM2021-59, pp. 31-36, Nov. 2021.
Paper # EMM2021-59 
Date of Issue 2021-11-08 (EA, EMM) 
ISSN 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 EA2021-32 EMM2021-59

Conference Information
Committee EMM EA ASJ-H  
Conference Date 2021-11-15 - 2021-11-16 
Place (in Japanese) (See Japanese page) 
Place (in English) Online 
Topics (in Japanese) (See Japanese page) 
Topics (in English) [Beginners Session] Engineering/Electro Acoustics, Content Processing, Digital Watermarking, Psychological and Physiological Acoustics, and Related Topics 
Paper Information
Registration To EMM 
Conference Code 2021-11-EMM-EA-H 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) A consideration of training datasets for universal detectors of CNN-generated images 
Sub Title (in English)  
Keyword(1) GAN  
Keyword(2) CNN  
Keyword(3) checkerboard artifact  
Keyword(4) fake image  
1st Author's Name Miki Tanaka  
1st Author's Affiliation Tokyo Metropolitan University (Tokyo Metro. Univ.)
2nd Author's Name Hitoshi Kiya  
2nd Author's Affiliation Tokyo Metropolitan University (Tokyo Metro. Univ.)
3rd Author's Name  
3rd Author's Affiliation ()
4th Author's Name  
4th Author's Affiliation ()
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 Author-1 
Date Time 2021-11-15 09:00:00 
Presentation Time 120 minutes 
Registration for EMM 
Paper # EA2021-32, EMM2021-59 
Volume (vol) vol.121 
Number (no) no.246(EA), no.247(EMM) 
Page pp.31-36 
Date of Issue 2021-11-08 (EA, EMM) 

[Return to Top Page]

[Return to IEICE Web Page]

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