IPSM-GAN: A Generative Adversarial Network for Shadow Removal Guided by Mixed Shadow Masks

DOI

10.2991/978-94-6463-300-9_82How to use a DOI?

Keywords

Shadow removal; generative adversarial network; deep learning

Abstract

Recently, thanks to the rapid development of deep learning, there are many methods to remove shadows in images by using generative adversarial networks. Most of them can learn the relationship between different domains, like shadow and shadow-free areas, to transform the shadow areas into areas with no shadow. However, due to inaccurate shadow shapes or masks obtained, these methods cannot lead to a better performance in the shadow image and even create more artifacts. To solve these problems, the authors propose IPSM-GAN, a new framework that learns to remove shadows in images by formulating cycle-consistency constraints and the guidance of mixed shadow masks. The mixed shadow mask generation method can accurately capture the shape of shadows. Also, the method can be a guide to the learning of the framework, which makes IPSM-GAN achieve better performance in removing shadows. Extensive experimental results verify the effectiveness of the proposed method, which can provide some new insights into the research field of shadow removal.

Copyright

© 2023 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)

TY  - CONF
AU  - Chuang Xie
PY  - 2023
DA  - 2023/11/27
TI  - IPSM-GAN: A Generative Adversarial Network for Shadow Removal Guided by Mixed Shadow Masks
BT  - Proceedings of the 2023 International Conference on Image, Algorithms and Artificial Intelligence (ICIAAI 2023)
PB  - Atlantis Press
SP  - 788
EP  - 799
SN  - 2352-538X
UR  - https://doi.org/10.2991/978-94-6463-300-9_82
DO  - 10.2991/978-94-6463-300-9_82
ID  - Xie2023
ER  -