Loading…

ACT-SR: Aggregation Connection Transformer for Remote Sensing Image Super-Resolution

Recently, Transformer-based methods have shown impressive performances in remote sensing image superresolution (RSISR). However, the application of Transformer in RSISR frequently results in artifacts and the loss of image detail due to limited information transmission pathways and the constraints o...

Full description

Saved in:

Bibliographic Details
Published in:	IEEE journal of selected topics in applied earth observations and remote sensing 2024-11, p.1-12
Main Authors:	Kang, Yingdong, Wang, Xinyu, Zhang, Xuemin, Wang, Shaoju, Jin, Guang
Format:	Article
Language:	English
Subjects:	Convolution Deep learning Feature extraction Image reconstruction Image restoration Interpolation Logic gates optical remote sensing Remote sensing Super-resolution Transformer Transformers Visualization
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

cited_by
cites
container_end_page	12
container_issue
container_start_page	1
container_title	IEEE journal of selected topics in applied earth observations and remote sensing
container_volume
creator	Kang, Yingdong Wang, Xinyu Zhang, Xuemin Wang, Shaoju Jin, Guang
description	Recently, Transformer-based methods have shown impressive performances in remote sensing image superresolution (RSISR). However, the application of Transformer in RSISR frequently results in artifacts and the loss of image detail due to limited information transmission pathways and the constraints of uni-dimensional self-attention mechanisms. To solve these problems, an Aggregation Connection Transformer (ACT-SR) is proposed for RSISR. ACT-SR employs an advanced attention mechanism designed to enrich information transmission across spatial and channel dimensions, thereby enlarging the receptive fields for more accurate feature extraction. A core component of ACT-SR is the novel aggregation connection attention block, which effectively captures spatial similarities and channel importance, aggregating this information through a combination of series and parallel connections for enhanced feature representation. Furthermore, a new gated feed-forward network is introduced to enhance the nonlinear mapping capabilities of the Transformer and control the information flow through the network. In addition, ACT-SR integrates a shifted windows scheme alongside interpolation residual calculation to facilitate efficient detail recovery and artifact elimination. Experimental results confirm the effectiveness of the proposed modules, with ACT-SR outperforming several state-of-the-art RSISR methods in both objective metrics and visual quality
doi_str_mv	10.1109/JSTARS.2024.3506717
format	article
fullrecord	<record><control><sourceid>crossref_ieee_</sourceid><recordid>TN_cdi_ieee_primary_10767430</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>10767430</ieee_id><sourcerecordid>10_1109_JSTARS_2024_3506717</sourcerecordid><originalsourceid>FETCH-LOGICAL-c650-4d9bc9f6188e41a07cd76087bd98c96311421b74d53c096594faef765c6cb51d3</originalsourceid><addsrcrecordid>eNpNkMtOg0AYhSdGE7H6BLrgBabOz9wYd4R4qWliAuzJMPwQTIFmpl349hbbhatzzuI7i4-QR2BrAGaeP8sqK8p1whKx5pIpDfqKRAlIoCC5vCYRGG4oCCZuyV0I34ypRBsekSrLK1oWL3HW9x57exjmKc7naUL3Vytvp9DNfkQfnyIucJwPGJc4hWHq481o-9M67tHTAsO8Oy7UPbnp7C7gwyVXpHp7rfIPuv163-TZljolGRWtaZzpFKQpCrBMu1YrluqmNakzigOIBBotWskdM0oa0VnstJJOuUZCy1eEn2-dn0Pw2NV7P4zW_9TA6sVLffZSL17qi5cT9XSmBkT8R2ilBWf8F598X2s</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>ACT-SR: Aggregation Connection Transformer for Remote Sensing Image Super-Resolution</title><source>Alma/SFX Local Collection</source><creator>Kang, Yingdong ; Wang, Xinyu ; Zhang, Xuemin ; Wang, Shaoju ; Jin, Guang</creator><creatorcontrib>Kang, Yingdong ; Wang, Xinyu ; Zhang, Xuemin ; Wang, Shaoju ; Jin, Guang</creatorcontrib><description>Recently, Transformer-based methods have shown impressive performances in remote sensing image superresolution (RSISR). However, the application of Transformer in RSISR frequently results in artifacts and the loss of image detail due to limited information transmission pathways and the constraints of uni-dimensional self-attention mechanisms. To solve these problems, an Aggregation Connection Transformer (ACT-SR) is proposed for RSISR. ACT-SR employs an advanced attention mechanism designed to enrich information transmission across spatial and channel dimensions, thereby enlarging the receptive fields for more accurate feature extraction. A core component of ACT-SR is the novel aggregation connection attention block, which effectively captures spatial similarities and channel importance, aggregating this information through a combination of series and parallel connections for enhanced feature representation. Furthermore, a new gated feed-forward network is introduced to enhance the nonlinear mapping capabilities of the Transformer and control the information flow through the network. In addition, ACT-SR integrates a shifted windows scheme alongside interpolation residual calculation to facilitate efficient detail recovery and artifact elimination. Experimental results confirm the effectiveness of the proposed modules, with ACT-SR outperforming several state-of-the-art RSISR methods in both objective metrics and visual quality</description><identifier>ISSN: 1939-1404</identifier><identifier>EISSN: 2151-1535</identifier><identifier>DOI: 10.1109/JSTARS.2024.3506717</identifier><identifier>CODEN: IJSTHZ</identifier><language>eng</language><publisher>IEEE</publisher><subject>Convolution ; Deep learning ; Feature extraction ; Image reconstruction ; Image restoration ; Interpolation ; Logic gates ; optical remote sensing ; Remote sensing ; Super-resolution ; Transformer ; Transformers ; Visualization</subject><ispartof>IEEE journal of selected topics in applied earth observations and remote sensing, 2024-11, p.1-12</ispartof><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>314,780,784,27923,27924</link.rule.ids></links><search><creatorcontrib>Kang, Yingdong</creatorcontrib><creatorcontrib>Wang, Xinyu</creatorcontrib><creatorcontrib>Zhang, Xuemin</creatorcontrib><creatorcontrib>Wang, Shaoju</creatorcontrib><creatorcontrib>Jin, Guang</creatorcontrib><title>ACT-SR: Aggregation Connection Transformer for Remote Sensing Image Super-Resolution</title><title>IEEE journal of selected topics in applied earth observations and remote sensing</title><addtitle>JSTARS</addtitle><description>Recently, Transformer-based methods have shown impressive performances in remote sensing image superresolution (RSISR). However, the application of Transformer in RSISR frequently results in artifacts and the loss of image detail due to limited information transmission pathways and the constraints of uni-dimensional self-attention mechanisms. To solve these problems, an Aggregation Connection Transformer (ACT-SR) is proposed for RSISR. ACT-SR employs an advanced attention mechanism designed to enrich information transmission across spatial and channel dimensions, thereby enlarging the receptive fields for more accurate feature extraction. A core component of ACT-SR is the novel aggregation connection attention block, which effectively captures spatial similarities and channel importance, aggregating this information through a combination of series and parallel connections for enhanced feature representation. Furthermore, a new gated feed-forward network is introduced to enhance the nonlinear mapping capabilities of the Transformer and control the information flow through the network. In addition, ACT-SR integrates a shifted windows scheme alongside interpolation residual calculation to facilitate efficient detail recovery and artifact elimination. Experimental results confirm the effectiveness of the proposed modules, with ACT-SR outperforming several state-of-the-art RSISR methods in both objective metrics and visual quality</description><subject>Convolution</subject><subject>Deep learning</subject><subject>Feature extraction</subject><subject>Image reconstruction</subject><subject>Image restoration</subject><subject>Interpolation</subject><subject>Logic gates</subject><subject>optical remote sensing</subject><subject>Remote sensing</subject><subject>Super-resolution</subject><subject>Transformer</subject><subject>Transformers</subject><subject>Visualization</subject><issn>1939-1404</issn><issn>2151-1535</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>ESBDL</sourceid><recordid>eNpNkMtOg0AYhSdGE7H6BLrgBabOz9wYd4R4qWliAuzJMPwQTIFmpl349hbbhatzzuI7i4-QR2BrAGaeP8sqK8p1whKx5pIpDfqKRAlIoCC5vCYRGG4oCCZuyV0I34ypRBsekSrLK1oWL3HW9x57exjmKc7naUL3Vytvp9DNfkQfnyIucJwPGJc4hWHq481o-9M67tHTAsO8Oy7UPbnp7C7gwyVXpHp7rfIPuv163-TZljolGRWtaZzpFKQpCrBMu1YrluqmNakzigOIBBotWskdM0oa0VnstJJOuUZCy1eEn2-dn0Pw2NV7P4zW_9TA6sVLffZSL17qi5cT9XSmBkT8R2ilBWf8F598X2s</recordid><startdate>20241125</startdate><enddate>20241125</enddate><creator>Kang, Yingdong</creator><creator>Wang, Xinyu</creator><creator>Zhang, Xuemin</creator><creator>Wang, Shaoju</creator><creator>Jin, Guang</creator><general>IEEE</general><scope>97E</scope><scope>ESBDL</scope><scope>RIA</scope><scope>RIE</scope><scope>AAYXX</scope><scope>CITATION</scope></search><sort><creationdate>20241125</creationdate><title>ACT-SR: Aggregation Connection Transformer for Remote Sensing Image Super-Resolution</title><author>Kang, Yingdong ; Wang, Xinyu ; Zhang, Xuemin ; Wang, Shaoju ; Jin, Guang</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c650-4d9bc9f6188e41a07cd76087bd98c96311421b74d53c096594faef765c6cb51d3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Convolution</topic><topic>Deep learning</topic><topic>Feature extraction</topic><topic>Image reconstruction</topic><topic>Image restoration</topic><topic>Interpolation</topic><topic>Logic gates</topic><topic>optical remote sensing</topic><topic>Remote sensing</topic><topic>Super-resolution</topic><topic>Transformer</topic><topic>Transformers</topic><topic>Visualization</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Kang, Yingdong</creatorcontrib><creatorcontrib>Wang, Xinyu</creatorcontrib><creatorcontrib>Zhang, Xuemin</creatorcontrib><creatorcontrib>Wang, Shaoju</creatorcontrib><creatorcontrib>Jin, Guang</creatorcontrib><collection>IEEE All-Society Periodicals Package (ASPP) 2005-present</collection><collection>IEEE Xplore Open Access Journals</collection><collection>IEEE All-Society Periodicals Package (ASPP) 1998-Present</collection><collection>IEEE Xplore</collection><collection>CrossRef</collection><jtitle>IEEE journal of selected topics in applied earth observations and remote sensing</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Kang, Yingdong</au><au>Wang, Xinyu</au><au>Zhang, Xuemin</au><au>Wang, Shaoju</au><au>Jin, Guang</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>ACT-SR: Aggregation Connection Transformer for Remote Sensing Image Super-Resolution</atitle><jtitle>IEEE journal of selected topics in applied earth observations and remote sensing</jtitle><stitle>JSTARS</stitle><date>2024-11-25</date><risdate>2024</risdate><spage>1</spage><epage>12</epage><pages>1-12</pages><issn>1939-1404</issn><eissn>2151-1535</eissn><coden>IJSTHZ</coden><abstract>Recently, Transformer-based methods have shown impressive performances in remote sensing image superresolution (RSISR). However, the application of Transformer in RSISR frequently results in artifacts and the loss of image detail due to limited information transmission pathways and the constraints of uni-dimensional self-attention mechanisms. To solve these problems, an Aggregation Connection Transformer (ACT-SR) is proposed for RSISR. ACT-SR employs an advanced attention mechanism designed to enrich information transmission across spatial and channel dimensions, thereby enlarging the receptive fields for more accurate feature extraction. A core component of ACT-SR is the novel aggregation connection attention block, which effectively captures spatial similarities and channel importance, aggregating this information through a combination of series and parallel connections for enhanced feature representation. Furthermore, a new gated feed-forward network is introduced to enhance the nonlinear mapping capabilities of the Transformer and control the information flow through the network. In addition, ACT-SR integrates a shifted windows scheme alongside interpolation residual calculation to facilitate efficient detail recovery and artifact elimination. Experimental results confirm the effectiveness of the proposed modules, with ACT-SR outperforming several state-of-the-art RSISR methods in both objective metrics and visual quality</abstract><pub>IEEE</pub><doi>10.1109/JSTARS.2024.3506717</doi><tpages>12</tpages><oa>free_for_read</oa></addata></record>
fulltext	fulltext
identifier	ISSN: 1939-1404
ispartof	IEEE journal of selected topics in applied earth observations and remote sensing, 2024-11, p.1-12
issn	1939-1404 2151-1535
language	eng
recordid	cdi_ieee_primary_10767430
source	Alma/SFX Local Collection
subjects	Convolution Deep learning Feature extraction Image reconstruction Image restoration Interpolation Logic gates optical remote sensing Remote sensing Super-resolution Transformer Transformers Visualization
title	ACT-SR: Aggregation Connection Transformer for Remote Sensing Image Super-Resolution
url	http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-13T09%3A57%3A40IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-crossref_ieee_&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=ACT-SR:%20Aggregation%20Connection%20Transformer%20for%20Remote%20Sensing%20Image%20Super-Resolution&rft.jtitle=IEEE%20journal%20of%20selected%20topics%20in%20applied%20earth%20observations%20and%20remote%20sensing&rft.au=Kang,%20Yingdong&rft.date=2024-11-25&rft.spage=1&rft.epage=12&rft.pages=1-12&rft.issn=1939-1404&rft.eissn=2151-1535&rft.coden=IJSTHZ&rft_id=info:doi/10.1109/JSTARS.2024.3506717&rft_dat=%3Ccrossref_ieee_%3E10_1109_JSTARS_2024_3506717%3C/crossref_ieee_%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-c650-4d9bc9f6188e41a07cd76087bd98c96311421b74d53c096594faef765c6cb51d3%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_id=info:pmid/&rft_ieee_id=10767430&rfr_iscdi=true