Information: Join today and make your research activities more affordable! Technical workshop participation fees and annual registration fees are available at member rates.
Notice: [Important] Announcement of Changes to Registration Fee Payment and Manuscript Upload Procedures for IEICE Technical Meetings
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 2023-05-18 13:30
Image Harmonization Using Diffusion Model for Perceptual Quality Improvement
Taito Naruki, Norimichi Ukita (TTI) PRMU2023-1
Abstract (in Japanese) (See Japanese page) 
(in English) Image harmonization is the task of eliminating the discomfort of color tones that occurs when images are composited. However, conventional image harmonization datasets were not able to take into account the perceptual aspect of discomfort, and models were learned and evaluated only by the difference from the GT. For this reason, even if the generated image was not uncomfortable, if the difference from the GT was large, it was treated as an insufficiently trained image.

Recently, the diffusion model has been attracting attention. The diffusion model generates images by gradually removing noise from a noisy image over multiple steps, and is known for its higher quality and diversity of generated images compared to other generative models due to this gradually restoring generative process and stable objective function. In addition, the method called DPS applies an inverse problem to the diffusion model to restore an image in zero shots such that it satisfies a given condition.

In the proposed method, the image is generated by applying constraints based on image gradients and VGG features in DPS, which results in image harmonization with high quality while preserving foreground textures. Furthermore, by gradually increasing the constraints at each step of the reverse process, the restoration is performed with higher perceptual quality.
Keyword (in Japanese) (See Japanese page) 
(in English) Image harmonization / Diffusion model / Inverse problem / / / / /  
Reference Info. IEICE Tech. Rep., vol. 123, no. 30, PRMU2023-1, pp. 1-5, May 2023.
Paper # PRMU2023-1 
Date of Issue 2023-05-11 (PRMU) 
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 PRMU2023-1

Conference Information
Committee PRMU IPSJ-CVIM  
Conference Date 2023-05-18 - 2023-05-19 
Place (in Japanese) (See Japanese page) 
Place (in English)  
Topics (in Japanese) (See Japanese page) 
Topics (in English)  
Paper Information
Registration To PRMU 
Conference Code 2023-05-PRMU-CVIM 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) Image Harmonization Using Diffusion Model for Perceptual Quality Improvement 
Sub Title (in English)  
Keyword(1) Image harmonization  
Keyword(2) Diffusion model  
Keyword(3) Inverse problem  
Keyword(4)  
Keyword(5)  
Keyword(6)  
Keyword(7)  
Keyword(8)  
1st Author's Name Taito Naruki  
1st Author's Affiliation Toyota Technological Institute (TTI)
2nd Author's Name Norimichi Ukita  
2nd Author's Affiliation Toyota Technological Institute (TTI)
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 ()
21st Author's Name  
21st Author's Affiliation ()
22nd Author's Name  
22nd Author's Affiliation ()
23rd Author's Name  
23rd Author's Affiliation ()
24th Author's Name  
24th Author's Affiliation ()
25th Author's Name  
25th Author's Affiliation ()
26th Author's Name / /
26th Author's Affiliation ()
()
27th Author's Name / /
27th Author's Affiliation ()
()
28th Author's Name / /
28th Author's Affiliation ()
()
29th Author's Name / /
29th Author's Affiliation ()
()
30th Author's Name / /
30th Author's Affiliation ()
()
31st Author's Name / /
31st Author's Affiliation ()
()
32nd Author's Name / /
32nd Author's Affiliation ()
()
33rd Author's Name / /
33rd Author's Affiliation ()
()
34th Author's Name / /
34th Author's Affiliation ()
()
35th Author's Name / /
35th Author's Affiliation ()
()
36th Author's Name / /
36th Author's Affiliation ()
()
Speaker Author-1 
Date Time 2023-05-18 13:30:00 
Presentation Time 15 minutes 
Registration for PRMU 
Paper # PRMU2023-1 
Volume (vol) vol.123 
Number (no) no.30 
Page pp.1-5 
#Pages 5 
Date of Issue 2023-05-11 (PRMU) 

[Return to Top Page]

[Return to IEICE Web Page]
The Institute of Electronics, Information and Communication Engineers (IEICE), Japan