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 2019-06-21 13:30
An image pre-transformation method to suppress recognition errors in highly compressed images
Satoshi Suzuki, Motohiro Takagi, Kazuya Hayase, Takayuki Onishi, Atsushi Shimizu (NTT) IE2019-18
Abstract (in Japanese) (See Japanese page) 
(in English) In image recognition, it is desirable to input an original image because image distortion leads to lower accuracy.
However, when the image is compressed with high compression rate the accuracy may be degraded due to the loss of high frequency information and coding artifacts.
This paper describes a method for pre-transforming images before image compression.
It aims to preserve image recognition accuracy using deep neural network techniques, even for highly compressed images.
A deep encoder-decoder network model is used as the pre-transformation model.
The model is learnt with a new loss function that combines recognition loss and the loss that increases the spatial correlation.
Compared with the original image the image transformed with the proposed method reduces the bitrate while maintaining equivalent accuracy.
Keyword (in Japanese) (See Japanese page) 
(in English) Image compression / image recognition / deep neural network / image pre-transformation / / / /  
Reference Info. IEICE Tech. Rep., vol. 119, no. 95, IE2019-18, pp. 7-12, June 2019.
Paper # IE2019-18 
Date of Issue 2019-06-14 (IE) 
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 IE2019-18

Conference Information
Committee IE  
Conference Date 2019-06-21 - 2019-06-21 
Place (in Japanese) (See Japanese page) 
Place (in English)  
Topics (in Japanese) (See Japanese page) 
Topics (in English) Image processing, coding, etc. 
Paper Information
Registration To IE 
Conference Code 2019-06-IE 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) An image pre-transformation method to suppress recognition errors in highly compressed images 
Sub Title (in English)  
Keyword(1) Image compression  
Keyword(2) image recognition  
Keyword(3) deep neural network  
Keyword(4) image pre-transformation  
1st Author's Name Satoshi Suzuki  
1st Author's Affiliation Nippon Telegraph and Telephone Corporation (NTT)
2nd Author's Name Motohiro Takagi  
2nd Author's Affiliation Nippon Telegraph and Telephone Corporation (NTT)
3rd Author's Name Kazuya Hayase  
3rd Author's Affiliation Nippon Telegraph and Telephone Corporation (NTT)
4th Author's Name Takayuki Onishi  
4th Author's Affiliation Nippon Telegraph and Telephone Corporation (NTT)
5th Author's Name Atsushi Shimizu  
5th Author's Affiliation Nippon Telegraph and Telephone Corporation (NTT)
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 ()
Date Time 2019-06-21 13:30:00 
Presentation Time 30 
Registration for IE 
Paper # IEICE-IE2019-18 
Volume (vol) IEICE-119 
Number (no) no.95 
Page pp.7-12 
#Pages IEICE-6 
Date of Issue IEICE-IE-2019-06-14 

[Return to Top Page]

[Return to IEICE Web Page]

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