Proceedings of the 2nd International Conference on Intelligent Design and Innovative Technology (ICIDIT 2023)

<Previous Article In Volume
Next Article In Volume>

Adaptive Convolution Kernel for Painterly Image Harmonization

Authors
Xiao Zhang1, *, Yun Jiang1, Shanshan Wang1
1School of Artificial Intelligence, Chongqing Technology and Business University, Chongqing, China
*Corresponding author. Email: zhangxiao@ctbu.edu.cn
Corresponding Author
Xiao Zhang
Available Online 10 October 2023.
DOI
10.2991/978-94-6463-266-8_52How to use a DOI?
Keywords
image synthesis; image harmonization; painterly image harmonization
Abstract

Image synthesis is a technique for cutting out foreground images and pasting them onto another image. To make the synthesized image harmonious, it is necessary to adjust the color and light of the foreground image so that it blends smoothly with the background image and has a consistent color. When the background image is an artistic style image, this process is called painterly image harmonization. This method can adjust the style of the foreground image so that it is compatible with the background image, producing a visually harmonious composite image while preserving the content of the foreground image. Existing painterly image harmonization methods have relied heavily on AdaIN methods, which often focus on color harmony but ignore the local structure information in the style image. We propose a new method based on dynamic convolution kernel for painterly image harmonization, which can dynamically generate convolution kernel to accommodate different style images during inference. Our method can effectively perceive the spatial structural elements of style images and generate more aesthetically pleasing composite images than AdaIN-based painterly image harmonization methods.

Copyright
© 2024 The Author(s)
Open Access
Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

<Previous Article In Volume
Next Article In Volume>
Volume Title
Proceedings of the 2nd International Conference on Intelligent Design and Innovative Technology (ICIDIT 2023)
Series
Atlantis Highlights in Intelligent Systems
Publication Date
10 October 2023
ISBN
978-94-6463-266-8
ISSN
2589-4919
DOI
10.2991/978-94-6463-266-8_52How to use a DOI?
Copyright
© 2024 The Author(s)
Open Access
Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Cite this article

risenwbib
TY  - CONF
AU  - Xiao Zhang
AU  - Yun Jiang
AU  - Shanshan Wang
PY  - 2023
DA  - 2023/10/10
TI  - Adaptive Convolution Kernel for Painterly Image Harmonization
BT  - Proceedings of the 2nd International Conference on Intelligent Design and Innovative Technology (ICIDIT 2023)
PB  - Atlantis Press
SP  - 478
EP  - 484
SN  - 2589-4919
UR  - https://doi.org/10.2991/978-94-6463-266-8_52
DO  - 10.2991/978-94-6463-266-8_52
ID  - Zhang2023
ER  -