FRANet: A Feature Refinement Attention Network for SAR Image Denoising

Since synthetic aperture radar (SAR) images have complex noise and have no clean reference images, SAR image denoising is very challenging. With the development of deep learning, several denoising algorithms based on deep learning are proposed to achieve a better SAR image denoising effect. However,...

Full description

Saved in:
Bibliographic Details
Main Authors: Shuaiqi Liu, Yu Lei, Qi Hu, Ming Liu, Bing Li, Weiming Hu, Yu-Dong Zhang
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10979213/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850268419560046592
author Shuaiqi Liu
Yu Lei
Qi Hu
Ming Liu
Bing Li
Weiming Hu
Yu-Dong Zhang
author_facet Shuaiqi Liu
Yu Lei
Qi Hu
Ming Liu
Bing Li
Weiming Hu
Yu-Dong Zhang
author_sort Shuaiqi Liu
collection DOAJ
description Since synthetic aperture radar (SAR) images have complex noise and have no clean reference images, SAR image denoising is very challenging. With the development of deep learning, several denoising algorithms based on deep learning are proposed to achieve a better SAR image denoising effect. However, most networks are prone to gradient disappearance and explosion in the training process. The deep network model will produce an excessive amount of computation. The denoising time is also too long. Since most of the denoising algorithms based on deep learning use simulated images for model training, it is difficult to effectively suppress speckle noise in the real SAR image while a balance between denoising and detail preservation cannot be achieved. To address the mentioned problems, we propose a novel feature refinement attention network named FRANet. In FRANet, a feature refinement network is first used to refine the input noise image to extract more useful features while accelerating network training. Second, a feature attention encoder–decoder network is constructed for deep feature extraction. This network uses an asymmetric encoder–decoder structure to expand the receptive field, which can improve the information extraction ability and reduce the number of parameters effectively. Finally, the final denoised SAR image is obtained by global residual learning. Compared with other denoising algorithms, the proposed algorithm can achieve better results in denoising performance and running time.
format Article
id doaj-art-5c07b3199eac47f1bf7e16b2322f38c7
institution OA Journals
issn 1939-1404
2151-1535
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
spelling doaj-art-5c07b3199eac47f1bf7e16b2322f38c72025-08-20T01:53:30ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-0118123431236310.1109/JSTARS.2025.356484610979213FRANet: A Feature Refinement Attention Network for SAR Image DenoisingShuaiqi Liu0https://orcid.org/0000-0001-7520-8226Yu Lei1https://orcid.org/0000-0003-3392-9441Qi Hu2https://orcid.org/0000-0002-7495-1066Ming Liu3https://orcid.org/0009-0002-9258-0737Bing Li4https://orcid.org/0000-0002-5888-6735Weiming Hu5https://orcid.org/0000-0001-9237-8825Yu-Dong Zhang6https://orcid.org/0000-0002-4870-1493College of Electronic and Information Engineering, Machine Vision Engineering Research Center of Hebei Province, Hebei University, Baoding, ChinaCollege of Electronic and Information Engineering, Machine Vision Engineering Research Center of Hebei Province, Hebei University, Baoding, ChinaCollege of Electronic and Information Engineering, Machine Vision Engineering Research Center of Hebei Province, Hebei University, Baoding, ChinaEducation and Teaching Research and Teacher Training Promotion Center, Hebei University, Baoding, ChinaNational Laboratory of Pattern Recognition (NLPR), Institute of Automation, Chinese Academy of Sciences, Beijing, ChinaNational Laboratory of Pattern Recognition (NLPR), Institute of Automation, Chinese Academy of Sciences, Beijing, ChinaSchool of Computing and Mathematical Science, University of Leicester, Leicester, U.K.Since synthetic aperture radar (SAR) images have complex noise and have no clean reference images, SAR image denoising is very challenging. With the development of deep learning, several denoising algorithms based on deep learning are proposed to achieve a better SAR image denoising effect. However, most networks are prone to gradient disappearance and explosion in the training process. The deep network model will produce an excessive amount of computation. The denoising time is also too long. Since most of the denoising algorithms based on deep learning use simulated images for model training, it is difficult to effectively suppress speckle noise in the real SAR image while a balance between denoising and detail preservation cannot be achieved. To address the mentioned problems, we propose a novel feature refinement attention network named FRANet. In FRANet, a feature refinement network is first used to refine the input noise image to extract more useful features while accelerating network training. Second, a feature attention encoder–decoder network is constructed for deep feature extraction. This network uses an asymmetric encoder–decoder structure to expand the receptive field, which can improve the information extraction ability and reduce the number of parameters effectively. Finally, the final denoised SAR image is obtained by global residual learning. Compared with other denoising algorithms, the proposed algorithm can achieve better results in denoising performance and running time.https://ieeexplore.ieee.org/document/10979213/Attention encoder–decoder networkdeep learningfeature refinementimage denoisingsynthetic aperture radar (SAR)
spellingShingle Shuaiqi Liu
Yu Lei
Qi Hu
Ming Liu
Bing Li
Weiming Hu
Yu-Dong Zhang
FRANet: A Feature Refinement Attention Network for SAR Image Denoising
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Attention encoder–decoder network
deep learning
feature refinement
image denoising
synthetic aperture radar (SAR)
title FRANet: A Feature Refinement Attention Network for SAR Image Denoising
title_full FRANet: A Feature Refinement Attention Network for SAR Image Denoising
title_fullStr FRANet: A Feature Refinement Attention Network for SAR Image Denoising
title_full_unstemmed FRANet: A Feature Refinement Attention Network for SAR Image Denoising
title_short FRANet: A Feature Refinement Attention Network for SAR Image Denoising
title_sort franet a feature refinement attention network for sar image denoising
topic Attention encoder–decoder network
deep learning
feature refinement
image denoising
synthetic aperture radar (SAR)
url https://ieeexplore.ieee.org/document/10979213/
work_keys_str_mv AT shuaiqiliu franetafeaturerefinementattentionnetworkforsarimagedenoising
AT yulei franetafeaturerefinementattentionnetworkforsarimagedenoising
AT qihu franetafeaturerefinementattentionnetworkforsarimagedenoising
AT mingliu franetafeaturerefinementattentionnetworkforsarimagedenoising
AT bingli franetafeaturerefinementattentionnetworkforsarimagedenoising
AT weiminghu franetafeaturerefinementattentionnetworkforsarimagedenoising
AT yudongzhang franetafeaturerefinementattentionnetworkforsarimagedenoising