Mask-shadownet: Toward shadow removal via masked adaptive instance normalization
Publication Type
Journal Article
Publication Date
1-2021
Abstract
Shadow removal is an important yet challenging task in image processing and computer vision. Existing methods are limited in extracting good global features due to the interference of shadow. And also, most of them ignore a fact that features inside and outside the shaded area should be treated disparately because of different semantics or materials. In this letter, we propose a novel deep neural network Mask-ShadowNet for shadow removal. The core of our approach is a well-designed masked adaptive instance normalization (MAdaIN) mechanism with embedded aligners that serves two goals: 1) producing hidden features that considering an illumination consistency of different regions. 2) treating the feature statistics of shadow and non-shadow areas discriminately based on the shadow mask. Experimental results demonstrate that the proposed model outperforms the state-of-the-art on the ISTD benchmark. Our code is available in https://github.com/penguinbing/Mask-ShadowNet.
Keywords
Lighting, Feature extraction, Training, Neural networks, task analysis, Predictive models, Adaptation models, Deep neural network, shadow removal, masked adaptive instance normalization
Discipline
Information Security
Research Areas
Information Systems and Management
Publication
IEEE Signal Processing Letters
Volume
28
First Page
957
Last Page
961
ISSN
1070-9908
Identifier
10.1109/LSP.2021.3074082
Publisher
Institute of Electrical and Electronics Engineers
Citation
HE, Shengfeng; PENG, Bing; DONG, Junyu; and DU, Yong.
Mask-shadownet: Toward shadow removal via masked adaptive instance normalization. (2021). IEEE Signal Processing Letters. 28, 957-961.
Available at: https://ink.library.smu.edu.sg/sis_research/7874
Additional URL
https://doi.org/10.1109/LSP.2021.3074082