SCAT: Shift Channel Attention Transformer for Remote Sensing Image Super-Resolution

The quadratic increase in computational complexity caused by global receptive fields has been a persistent challenge when applying Transformer-based methods in remote sensing image super-resolution (RSISR), involving high-resolution images. Channel attention (CA)-based Transformers offer an efficien...

Full description

Saved in:
Bibliographic Details
Main Authors: Yingdong Kang, Xuemin Zhang, Shaoju Wang, Guang Jin
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/10886926/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850155042452013056
author Yingdong Kang
Xuemin Zhang
Shaoju Wang
Guang Jin
author_facet Yingdong Kang
Xuemin Zhang
Shaoju Wang
Guang Jin
author_sort Yingdong Kang
collection DOAJ
description The quadratic increase in computational complexity caused by global receptive fields has been a persistent challenge when applying Transformer-based methods in remote sensing image super-resolution (RSISR), involving high-resolution images. Channel attention (CA)-based Transformers offer an efficient approach with linear complexity by computing self-attention across the channel dimension. However, current CA-based Transformers suffer from performance degradation due to two main issues: constrained receptive field in the channel dimension caused by the multihead strategy and insufficient feature diversity resulting from the small size of attention matrices. To address these drawbacks, a novel shift channel attention Transformer (SCAT) is proposed for RSISR in this article. The core innovation of SCAT lies in its shift channel attention block (SCAB), which expands the receptive field by facilitating cross-head communication through a shift channel strategy. This design enables parallel computation of self-attention across multiple heads while ensuring robust cross-channel connections, thereby enhancing the global context modeling capabilities of the network. In addition, an attention supplementation module using depthwise convolution (DWC) is incorporated into SCAB to improve feature diversity. Finally, the proposed gated feedforward network utilizes the gating mechanism and DWC to effectively control the information flow and extract complementary spatial details. In the experiments, the effectiveness of these proposed modules is verified, and the SCAT model demonstrated superior performance in terms of quantitative and qualitative compared to several state-of-the-art RSISR methods.
format Article
id doaj-art-c0a94e84c6d94105bb3bd62e2c1bf9e3
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-c0a94e84c6d94105bb3bd62e2c1bf9e32025-08-20T02:25:04ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-0118103371034710.1109/JSTARS.2025.353759310886926SCAT: Shift Channel Attention Transformer for Remote Sensing Image Super-ResolutionYingdong Kang0https://orcid.org/0009-0001-3465-4366Xuemin Zhang1https://orcid.org/0009-0000-5583-0982Shaoju Wang2https://orcid.org/0009-0005-5730-1597Guang Jin3School of Remote Sensing and Information Engineering, Wuhan University, Wuhan, ChinaSchool of Remote Sensing and Information Engineering, Wuhan University, Wuhan, ChinaSchool of Remote Sensing and Information Engineering, Wuhan University, Wuhan, ChinaSchool of Remote Sensing and Information Engineering, Wuhan University, Wuhan, ChinaThe quadratic increase in computational complexity caused by global receptive fields has been a persistent challenge when applying Transformer-based methods in remote sensing image super-resolution (RSISR), involving high-resolution images. Channel attention (CA)-based Transformers offer an efficient approach with linear complexity by computing self-attention across the channel dimension. However, current CA-based Transformers suffer from performance degradation due to two main issues: constrained receptive field in the channel dimension caused by the multihead strategy and insufficient feature diversity resulting from the small size of attention matrices. To address these drawbacks, a novel shift channel attention Transformer (SCAT) is proposed for RSISR in this article. The core innovation of SCAT lies in its shift channel attention block (SCAB), which expands the receptive field by facilitating cross-head communication through a shift channel strategy. This design enables parallel computation of self-attention across multiple heads while ensuring robust cross-channel connections, thereby enhancing the global context modeling capabilities of the network. In addition, an attention supplementation module using depthwise convolution (DWC) is incorporated into SCAB to improve feature diversity. Finally, the proposed gated feedforward network utilizes the gating mechanism and DWC to effectively control the information flow and extract complementary spatial details. In the experiments, the effectiveness of these proposed modules is verified, and the SCAT model demonstrated superior performance in terms of quantitative and qualitative compared to several state-of-the-art RSISR methods.https://ieeexplore.ieee.org/document/10886926/Deep learningoptical remote sensingsuper-resolution (SR)transformers
spellingShingle Yingdong Kang
Xuemin Zhang
Shaoju Wang
Guang Jin
SCAT: Shift Channel Attention Transformer for Remote Sensing Image Super-Resolution
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Deep learning
optical remote sensing
super-resolution (SR)
transformers
title SCAT: Shift Channel Attention Transformer for Remote Sensing Image Super-Resolution
title_full SCAT: Shift Channel Attention Transformer for Remote Sensing Image Super-Resolution
title_fullStr SCAT: Shift Channel Attention Transformer for Remote Sensing Image Super-Resolution
title_full_unstemmed SCAT: Shift Channel Attention Transformer for Remote Sensing Image Super-Resolution
title_short SCAT: Shift Channel Attention Transformer for Remote Sensing Image Super-Resolution
title_sort scat shift channel attention transformer for remote sensing image super resolution
topic Deep learning
optical remote sensing
super-resolution (SR)
transformers
url https://ieeexplore.ieee.org/document/10886926/
work_keys_str_mv AT yingdongkang scatshiftchannelattentiontransformerforremotesensingimagesuperresolution
AT xueminzhang scatshiftchannelattentiontransformerforremotesensingimagesuperresolution
AT shaojuwang scatshiftchannelattentiontransformerforremotesensingimagesuperresolution
AT guangjin scatshiftchannelattentiontransformerforremotesensingimagesuperresolution