Proceedings of the International Workshop on Advances in Deep Learning for Image Analysis and Computer Vision (IWADIC 2025)

<Previous Article In Volume
Next Article In Volume>

Character Image Editing via Segment-Anything Model and In-Context Edit Integration

Authors
Zhuoran Jia1, *
1School of AI and Advanced Computing, Xi’an Jiaotong–Liverpool University, Suzhou, Jiangsu Province, China
*Corresponding author. Email: Zhuoran.Jia23@student.xjtlu.edu.cn
Corresponding Author
Zhuoran Jia
Available Online 24 April 2026.
DOI
10.2991/978-94-6239-648-7_29How to use a DOI?
Keywords
Image Editing; Image Segmentation; Natural Language Prompts; Image Style Transformation
Abstract

In recent years, the development of computer vision tasks for image segmentation has become relatively mature. Image editing tasks based on instructions can achieve powerful image modification by natural language prompts. However, existing instruction-based image editing driven by natural language or short instructions encounters instability and difficulties in maintaining the overall image style when dealing with complex prompt scenarios, especially in image editing tasks mainly featuring human figures. This paper, based on the image editing model ICEdit as the inpainting backend and combined with the segmentation model Segment-Anything Model (SAM), constructs a two-stage workflow of “mask generation → local editing” to improve the overall stability and accuracy of image editing tasks in special scenarios. Precise image segmentation can provide more accurate modification parts for preprocessing masks, thereby enhancing the over-all stability and accuracy of image editing tasks in special scenarios. This paper presents the visual results generated as a comparison under the same text input conditions, demonstrating more text-input-compliant image editing results compared to the original model and providing new ideas for the processing of such special images.

Copyright
© 2026 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 International Workshop on Advances in Deep Learning for Image Analysis and Computer Vision (IWADIC 2025)
Series
Advances in Computer Science Research
Publication Date
24 April 2026
ISBN
978-94-6239-648-7
ISSN
2352-538X
DOI
10.2991/978-94-6239-648-7_29How to use a DOI?
Copyright
© 2026 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  - Zhuoran Jia
PY  - 2026
DA  - 2026/04/24
TI  - Character Image Editing via Segment-Anything Model and In-Context Edit Integration
BT  - Proceedings of the International Workshop on Advances in Deep Learning for Image Analysis and Computer Vision (IWADIC 2025)
PB  - Atlantis Press
SP  - 260
EP  - 267
SN  - 2352-538X
UR  - https://doi.org/10.2991/978-94-6239-648-7_29
DO  - 10.2991/978-94-6239-648-7_29
ID  - Jia2026
ER  -