Mask-shadownet: Toward shadow removal via masked adaptive instance normalization
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 disp...
Saved in:
Main Authors: | , , , |
---|---|
Format: | text |
Language: | English |
Published: |
Institutional Knowledge at Singapore Management University
2021
|
Subjects: | |
Online Access: | https://ink.library.smu.edu.sg/sis_research/8380 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Singapore Management University |
Language: | English |
id |
sg-smu-ink.sis_research-9383 |
---|---|
record_format |
dspace |
spelling |
sg-smu-ink.sis_research-93832023-12-12T07:48:03Z Mask-shadownet: Toward shadow removal via masked adaptive instance normalization HE, Shengfeng PENG, Bing DONG, Junyu DU, Yong 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. 2021-04-19T07:00:00Z text https://ink.library.smu.edu.sg/sis_research/8380 info:doi/10.1109/LSP.2021.3074082 Research Collection School Of Computing and Information Systems eng Institutional Knowledge at Singapore Management University Aligners Global feature Image processing and computer vision Shadow mask Shadow removal State of the art Databases and Information Systems Theory and Algorithms |
institution |
Singapore Management University |
building |
SMU Libraries |
continent |
Asia |
country |
Singapore Singapore |
content_provider |
SMU Libraries |
collection |
InK@SMU |
language |
English |
topic |
Aligners Global feature Image processing and computer vision Shadow mask Shadow removal State of the art Databases and Information Systems Theory and Algorithms |
spellingShingle |
Aligners Global feature Image processing and computer vision Shadow mask Shadow removal State of the art Databases and Information Systems Theory and Algorithms HE, Shengfeng PENG, Bing DONG, Junyu DU, Yong Mask-shadownet: Toward shadow removal via masked adaptive instance normalization |
description |
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. |
format |
text |
author |
HE, Shengfeng PENG, Bing DONG, Junyu DU, Yong |
author_facet |
HE, Shengfeng PENG, Bing DONG, Junyu DU, Yong |
author_sort |
HE, Shengfeng |
title |
Mask-shadownet: Toward shadow removal via masked adaptive instance normalization |
title_short |
Mask-shadownet: Toward shadow removal via masked adaptive instance normalization |
title_full |
Mask-shadownet: Toward shadow removal via masked adaptive instance normalization |
title_fullStr |
Mask-shadownet: Toward shadow removal via masked adaptive instance normalization |
title_full_unstemmed |
Mask-shadownet: Toward shadow removal via masked adaptive instance normalization |
title_sort |
mask-shadownet: toward shadow removal via masked adaptive instance normalization |
publisher |
Institutional Knowledge at Singapore Management University |
publishDate |
2021 |
url |
https://ink.library.smu.edu.sg/sis_research/8380 |
_version_ |
1787136846806908928 |