CAT-EDNet: Cross-Attention Transformer-Based Encoder-Decoder Network for Salient Defect Detection of Strip Steel Surface

The morphologies of various surface defects on strip steel suffer from oil stain, water drops, steel textures, and erratic illumination. It is still challenging to recognize defect boundary precisely from cluttered backgrounds. This article emphasizes such a fact that skip connections between encode...

Full description

Saved in:

Bibliographic Details
Published in:	IEEE transactions on instrumentation and measurement Vol. 71; pp. 1 - 13
Main Authors:	Luo, Qiwu, Su, Jiaojiao, Yang, Chunhua, Gui, Weihua, Silven, Olli, Liu, Li
Format:	Journal Article
Language:	English
Published:	New York IEEE 2022 The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Subjects:	Automated visual inspection (AVI) Coders Decoding encoder–decoder network (EDNet) Entropy (Information theory) Feature extraction Inspection salient detection Steel steel strip Strip steel Strips Surface defects transformer Transformers Visualization Water drops
ISSN:	0018-9456, 1557-9662
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Abstract	The morphologies of various surface defects on strip steel suffer from oil stain, water drops, steel textures, and erratic illumination. It is still challenging to recognize defect boundary precisely from cluttered backgrounds. This article emphasizes such a fact that skip connections between encoder and decoder are not equally effective, attempts to adaptively allocate the aggregation weights that represent differentiated information entropy values in channelwise, by importing a stack of cross-attention transformer (CAT) into the encoder-decoder network (EDNet). Besides, a cross-attention refinement module (CARM) is constructed closely after the decoder to further optimize the coarse saliency maps. This newly nominated CAT-EDNet can well address the semantic gap issue among the multiscale features for its multihead attention structure. The CAT-EDNet performs best on insuring defect integrity and maintaining defect boundary details when compared with 12 state-of-the-arts, and the detection efficiency is at 28 fps even under the noise interfered scenario.
AbstractList	The morphologies of various surface defects on strip steel suffer from oil stain, water drops, steel textures, and erratic illumination. It is still challenging to recognize defect boundary precisely from cluttered backgrounds. This article emphasizes such a fact that skip connections between encoder and decoder are not equally effective, attempts to adaptively allocate the aggregation weights that represent differentiated information entropy values in channelwise, by importing a stack of cross-attention transformer (CAT) into the encoder-decoder network (EDNet). Besides, a cross-attention refinement module (CARM) is constructed closely after the decoder to further optimize the coarse saliency maps. This newly nominated CAT-EDNet can well address the semantic gap issue among the multiscale features for its multihead attention structure. The CAT-EDNet performs best on insuring defect integrity and maintaining defect boundary details when compared with 12 state-of-the-arts, and the detection efficiency is at 28 fps even under the noise interfered scenario.
Author	Gui, Weihua Silven, Olli Liu, Li Su, Jiaojiao Luo, Qiwu Yang, Chunhua
Author_xml	– sequence: 1 givenname: Qiwu orcidid: 0000-0003-2822-5538 surname: Luo fullname: Luo, Qiwu organization: School of Automation, Central South University, Changsha, China – sequence: 2 givenname: Jiaojiao orcidid: 0000-0002-2115-0978 surname: Su fullname: Su, Jiaojiao organization: School of Automation, Central South University, Changsha, China – sequence: 3 givenname: Chunhua orcidid: 0000-0003-2550-1509 surname: Yang fullname: Yang, Chunhua email: ychh@csu.edu.cn organization: School of Automation, Central South University, Changsha, China – sequence: 4 givenname: Weihua orcidid: 0000-0002-5337-6445 surname: Gui fullname: Gui, Weihua organization: School of Automation, Central South University, Changsha, China – sequence: 5 givenname: Olli orcidid: 0000-0002-2661-804X surname: Silven fullname: Silven, Olli organization: Center for Machine Vision and Signal Analysis (CMVS), University of Oulu, Oulu, Finland – sequence: 6 givenname: Li orcidid: 0000-0002-2011-2873 surname: Liu fullname: Liu, Li organization: College of System Engineering, National University of Defense Technology, Changsha, China
BookMark	eNp9kEFPwyAYhomZidv0buKFxHMn0EKLt9lNXTL1sHpuKPtIOrsyKYv676Vu8eDBC28I7_N94RmhQWtbQOiSkgmlRN4Ui6cJI4xNYio4S8kJGlLO00gKwQZoSAjNIplwcYZGXbchhKQiSYfoM58W0Xz2DP4W5852XTT1Hlpf2xYXTrWdsW4LLrpTHazxvNV2HW4z-EkcsA_r3nAo4ZVq6gDiGRjQffgQ_Rhr8Mq7ehdOgAav9s4oDefo1Kimg4tjjtHr_bzIH6Ply8Miny4jzST1kQHDFYnXFQhdVZkgoCGhcZaxJDMMjACdSFVxphIS3kSSpMZIKSRnlIKp4jG6PszdOfu-h86XG7t3bVhZMsF5FhwJGlrk0NK9Awem3Ll6q9xXSUnZ-y2D37L3Wx79BkT8QXTtVf9j71Td_AdeHcAaAH73yJSnMmbxN-Pgil4
CODEN	IEIMAO
CitedBy_id	crossref_primary_10_1109_TASE_2023_3307588 crossref_primary_10_1109_TIM_2023_3307753 crossref_primary_10_1007_s00371_024_03442_y crossref_primary_10_1088_1361_6501_adc1f3 crossref_primary_10_1109_TIM_2025_3551901 crossref_primary_10_1109_JSEN_2022_3229031 crossref_primary_10_1109_TIM_2024_3470998 crossref_primary_10_1109_TITS_2023_3293822 crossref_primary_10_3390_s23063216 crossref_primary_10_3390_s25072043 crossref_primary_10_3390_asi7010011 crossref_primary_10_1109_TIM_2025_3558829 crossref_primary_10_1109_JSEN_2024_3407874 crossref_primary_10_1109_TIM_2024_3504562 crossref_primary_10_1007_s11042_024_18684_0 crossref_primary_10_1109_TASE_2024_3457829 crossref_primary_10_1109_TIM_2023_3336452 crossref_primary_10_1109_TIM_2022_3214605 crossref_primary_10_1109_TII_2023_3348835 crossref_primary_10_1109_TIM_2023_3246519 crossref_primary_10_1109_TGRS_2024_3383649
Cites_doi	10.1109/TIM.2019.2963555 10.1109/CVPR.2014.39 10.1109/CVPR.2019.00766 10.1609/aaai.v36i3.20144 10.1109/CVPR.2009.5206596 10.1109/TIM.2012.2218677 10.1109/CVPR.2017.189 10.1109/CVPR.2019.00403 10.1109/ICCV.2019.00736 10.1007/978-3-319-50835-1_22 10.1109/CVPR46437.2021.00865 10.1109/CVPR46437.2021.00681 10.1109/ICCV.2015.123 10.1109/CVPR.2015.7299184 10.1109/CVPR42600.2020.00285 10.1109/ICCV.2017.487 10.1109/ICCV.2019.00887 10.1007/978-3-319-75238-9_6 10.1109/TIP.2021.3072811 10.1016/j.patcog.2020.107404 10.1109/CVPR.2018.00326 10.1109/TIM.2019.2915404 10.24963/ijcai.2018/97 10.1109/CVPR.2019.00320 10.1109/CVPR.2019.00404 10.1109/TIP.2020.2965989 10.1109/TPAMI.2016.2562626 10.1109/ACSSC.2003.1292216 10.1109/CVPR.2012.6247743 10.1007/978-3-030-58452-8_13 10.1109/ICCV.2017.433 10.1109/TIM.2020.3030167 10.1109/CVPR.2018.00187 10.1109/TII.2019.2958826 10.1109/TNNLS.2022.3227717 10.1109/ICCV48922.2021.00683 10.1007/s10479-005-5724-z 10.1109/TIM.2018.2852918 10.1109/TIM.2020.3002277
ContentType	Journal Article
Copyright	Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2022
Copyright_xml	– notice: Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2022
DBID	97E RIA RIE AAYXX CITATION 7SP 7U5 8FD L7M
DOI	10.1109/TIM.2022.3165270
DatabaseName	IEEE Xplore (IEEE) IEEE All-Society Periodicals Package (ASPP) 1998–Present IEEE Electronic Library (IEL) CrossRef Electronics & Communications Abstracts Solid State and Superconductivity Abstracts Technology Research Database Advanced Technologies Database with Aerospace
DatabaseTitle	CrossRef Solid State and Superconductivity Abstracts Technology Research Database Advanced Technologies Database with Aerospace Electronics & Communications Abstracts
DatabaseTitleList	Solid State and Superconductivity Abstracts
Database_xml	– sequence: 1 dbid: RIE name: IEEE Electronic Library (IEL) url: https://ieeexplore.ieee.org/ sourceTypes: Publisher
DeliveryMethod	fulltext_linktorsrc
Discipline	Engineering Physics
EISSN	1557-9662
EndPage	13
ExternalDocumentID	10_1109_TIM_2022_3165270 9757932
Genre	orig-research
GrantInformation_xml	– fundername: Science and Technology Innovation Program of Hunan Province grantid: 2021RC3019; 2021RC1001 funderid: 10.13039/501100019081 – fundername: Hunan Provincial Natural Science Foundation grantid: 2021JJ20078 funderid: 10.13039/501100004735 – fundername: National Natural Science Foundation of China grantid: 61973323; 6201101509 funderid: 10.13039/501100001809
GroupedDBID	-~X 0R~ 29I 4.4 5GY 5VS 6IK 85S 8WZ 97E A6W AAJGR AARMG AASAJ AAWTH ABAZT ABQJQ ABVLG ACGFO ACIWK ACNCT AENEX AETIX AGQYO AGSQL AHBIQ AI. AIBXA AKJIK AKQYR ALLEH ALMA_UNASSIGNED_HOLDINGS ATWAV BEFXN BFFAM BGNUA BKEBE BPEOZ CS3 DU5 EBS EJD F5P HZ~ H~9 IAAWW IBMZZ ICLAB IDIHD IFIPE IFJZH IPLJI JAVBF LAI M43 O9- OCL P2P RIA RIE RNS TN5 TWZ VH1 VJK AAYXX CITATION 7SP 7U5 8FD L7M
ID	FETCH-LOGICAL-c291t-fef5a03dbe6cbb860ece41388248f2ef6ec49ab52a400ec6447ff99695211efb3
IEDL.DBID	RIE
ISICitedReferencesCount	36
ISICitedReferencesURI	http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000794225000017&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN	0018-9456
IngestDate	Mon Jun 30 10:12:54 EDT 2025 Sat Nov 29 04:38:21 EST 2025 Tue Nov 18 22:11:17 EST 2025 Wed Aug 27 02:40:27 EDT 2025
IsPeerReviewed	true
IsScholarly	true
Language	English
License	https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html https://doi.org/10.15223/policy-029 https://doi.org/10.15223/policy-037
LinkModel	DirectLink
MergedId	FETCHMERGED-LOGICAL-c291t-fef5a03dbe6cbb860ece41388248f2ef6ec49ab52a400ec6447ff99695211efb3
Notes	ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14
ORCID	0000-0003-2822-5538 0000-0002-2661-804X 0000-0002-5337-6445 0000-0003-2550-1509 0000-0002-2115-0978 0000-0002-2011-2873
PQID	2655855761
PQPubID	85462
PageCount	13
ParticipantIDs	crossref_primary_10_1109_TIM_2022_3165270 crossref_citationtrail_10_1109_TIM_2022_3165270 proquest_journals_2655855761 ieee_primary_9757932
PublicationCentury	2000
PublicationDate	20220000 2022-00-00 20220101
PublicationDateYYYYMMDD	2022-01-01
PublicationDate_xml	– year: 2022 text: 20220000
PublicationDecade	2020
PublicationPlace	New York
PublicationPlace_xml	– name: New York
PublicationTitle	IEEE transactions on instrumentation and measurement
PublicationTitleAbbrev	TIM
PublicationYear	2022
Publisher	IEEE The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Publisher_xml	– name: IEEE – name: The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
References	ref13 Vaswani (ref15); 30 ref35 ref12 ref34 ref37 ref14 ref36 ref31 ref30 ref11 ref33 ref10 ref2 ref1 ref17 Chen (ref18) 2021 ref39 ref16 Xie (ref19) 2021 ref24 ref45 ref26 ref25 ref20 ref42 ref41 ref22 ref44 ref21 ref43 Islam (ref23) ref28 ref27 ref29 ref8 ref7 Nagendar (ref38) ref9 ref4 ref3 ref6 ref5 ref40 Tieleman (ref32) 2012; 4
References_xml	– ident: ref2 doi: 10.1109/TIM.2019.2963555 – ident: ref34 doi: 10.1109/CVPR.2014.39 – ident: ref26 doi: 10.1109/CVPR.2019.00766 – ident: ref27 doi: 10.1609/aaai.v36i3.20144 – ident: ref4 doi: 10.1109/CVPR.2009.5206596 – ident: ref7 doi: 10.1109/TIM.2012.2218677 – ident: ref22 doi: 10.1109/CVPR.2017.189 – volume: 4 start-page: 26 issue: 2 year: 2012 ident: ref32 article-title: RMSprop: Divide the gradient by a running average of its recent magnitude publication-title: Neural Netw. Mach. Learn. – ident: ref41 doi: 10.1109/CVPR.2019.00403 – ident: ref10 doi: 10.1109/ICCV.2019.00736 – ident: ref29 doi: 10.1007/978-3-319-50835-1_22 – ident: ref13 doi: 10.1109/CVPR46437.2021.00865 – ident: ref17 doi: 10.1109/CVPR46437.2021.00681 – ident: ref31 doi: 10.1109/ICCV.2015.123 – ident: ref9 doi: 10.1109/CVPR.2015.7299184 – ident: ref43 doi: 10.1109/CVPR42600.2020.00285 – volume: 30 start-page: 5998 volume-title: Proc. NIPS ident: ref15 article-title: Attention is all you need – ident: ref33 doi: 10.1109/ICCV.2017.487 – ident: ref42 doi: 10.1109/ICCV.2019.00887 – volume-title: Proc. BMVC ident: ref38 article-title: Neuro-IoU: Learning a surrogate loss for semantic segmentation – ident: ref37 doi: 10.1007/978-3-319-75238-9_6 – ident: ref11 doi: 10.1109/TIP.2021.3072811 – ident: ref12 doi: 10.1016/j.patcog.2020.107404 – ident: ref39 doi: 10.1109/CVPR.2018.00326 – ident: ref6 doi: 10.1109/TIM.2019.2915404 – ident: ref36 doi: 10.24963/ijcai.2018/97 – ident: ref44 doi: 10.1109/CVPR.2019.00320 – ident: ref40 doi: 10.1109/CVPR.2019.00404 – ident: ref45 doi: 10.1109/TIP.2020.2965989 – ident: ref8 doi: 10.1109/TPAMI.2016.2562626 – ident: ref30 doi: 10.1109/ACSSC.2003.1292216 – ident: ref35 doi: 10.1109/CVPR.2012.6247743 – ident: ref20 doi: 10.1007/978-3-030-58452-8_13 – ident: ref24 doi: 10.1109/ICCV.2017.433 – ident: ref1 doi: 10.1109/TIM.2020.3030167 – year: 2021 ident: ref19 article-title: SegFormer: Simple and efficient design for semantic segmentation with transformers publication-title: arXiv:2105.15203 – start-page: 1 volume-title: Proc. BMVC ident: ref23 article-title: Salient object detection using a context-aware refinement network – ident: ref25 doi: 10.1109/CVPR.2018.00187 – ident: ref5 doi: 10.1109/TII.2019.2958826 – ident: ref16 doi: 10.1109/TNNLS.2022.3227717 – ident: ref21 doi: 10.1109/ICCV48922.2021.00683 – ident: ref28 doi: 10.1007/s10479-005-5724-z – year: 2021 ident: ref18 article-title: TransUNet: Transformers make strong encoders for medical image segmentation publication-title: arXiv:2102.04306 – ident: ref3 doi: 10.1109/TIM.2018.2852918 – ident: ref14 doi: 10.1109/TIM.2020.3002277
SSID	ssj0007647
Score	2.497521
Snippet	The morphologies of various surface defects on strip steel suffer from oil stain, water drops, steel textures, and erratic illumination. It is still...
SourceID	proquest crossref ieee
SourceType	Aggregation Database Enrichment Source Index Database Publisher
StartPage	1
SubjectTerms	Automated visual inspection (AVI) Coders Decoding encoder–decoder network (EDNet) Entropy (Information theory) Feature extraction Inspection salient detection Steel steel strip Strip steel Strips Surface defects transformer Transformers Visualization Water drops
Title	CAT-EDNet: Cross-Attention Transformer-Based Encoder-Decoder Network for Salient Defect Detection of Strip Steel Surface
URI	https://ieeexplore.ieee.org/document/9757932 https://www.proquest.com/docview/2655855761
Volume	71
WOSCitedRecordID	wos000794225000017&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText	1
inHoldings	1
isFullTextHit
isPrint
journalDatabaseRights	– providerCode: PRVIEE databaseName: IEEE Electronic Library (IEL) customDbUrl: eissn: 1557-9662 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0007647 issn: 0018-9456 databaseCode: RIE dateStart: 19630101 isFulltext: true titleUrlDefault: https://ieeexplore.ieee.org/ providerName: IEEE
link	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1RSyMxEB56onA-qKceVr0jD_ciGLvNdrMb32pbuXu4ctAKvi1JdgIHspXtVvz5TrLbUvAQDpZNHpIQ-JKZb5LJDMAPiQq1TguuhTJ8gCbhOo4FVyYj9ppY47AIySbS6TR7fFR_OnC9eQuDiMH5DG98NdzlFwu78kdlPZUmtJxI4H5KU9m81dpI3VQOmviYfdrAxArWV5KR6s1__SZDUAiyT2UifFriLRUUcqq8E8RBu9wf_t-8juCgZZFs2MD-BTpYHsP-VmzBY9gLvp12eQKvo-GcT8ZTrG_ZyM-ID-u68XJk8zVvxYrfkUIr2KT0r9wrPsZQsmnjJ86oEZsRaaeObIzeCYSKOjhylWzh2Kwm8UN_xCc2W1VOWzyFh_vJfPSTtwkXuBWqX3OHLtFRXBiU1phMRmiRlByR8EHmBDqJdqC0SYSmnY-WqFTqHBlMijhAH52Jv8JOuSjxDFjso86gMWEITV_kMmviQsnMxjp2XeitMchtG43cJ8V4yoNVEqmcUMs9anmLWheuNj2em0gcH7Q98Sht2rUAdeFyDXPebtVlLmRCJhOZXf3zf_e6gM9-7Obc5RJ26mqF32DXvtR_l9X3sArfAJXp28o
linkProvider	IEEE
linkToHtml	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3fa9swED5Kt7H2Yd3alWZtNz30ZTAtjvxTfcuSlJa2ZhAP-mYk-QSD4hTHKf3ze5KdUNgYDIzlB8kSfNLdd9LpDuAsQYlKpRVXQmoeoY65CkPBpc6IvcZGW6x8sok0z7O7O_lzC75t7sIgonc-w-_u05_lVwuzcltlQ5nGNJ1I4L6Ko0gE3W2tjdxNk6iLkDmiJUy8YH0oGchhcXVLpqAQZKEmsXCJiV8oIZ9V5Q9R7PXLxd7_jew9vOt5JBt3wH-ALaz3YfdFdMF9eOO9O83yAJ4m44LPpjm252ziRsTHbdv5ObJizVyx4T9IpVVsVrt77g2foi9Z3nmKM6rE5kTbqSGbonMDoaL1rlw1W1g2b0kA0Rvxns1XjVUGP8Kvi1kxueR9ygVuhBy13KKNVRBWGhOjdZYEaJDUHNHwKLMCbYImkkrHQtHaR0NkKrWWTCZJLGCEVoeHsF0vajwCFrq4M6i1_4WiJ7CZ0WElk8yEKrQDGK4xKE0fj9ylxbgvvV0SyJJQKx1qZY_aAL5uWjx0sTj-UffAobSp1wM0gJM1zGW_WJelSGIymsjwGn36e6sv8PayuL0pb67y62PYcf10uzAnsN02KzyF1-ax_b1sPvsZ-Qwq698R
openUrl	ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=CAT-EDNet%3A+Cross-Attention+Transformer-Based+Encoder-Decoder+Network+for+Salient+Defect+Detection+of+Strip+Steel+Surface&rft.jtitle=IEEE+transactions+on+instrumentation+and+measurement&rft.au=Luo%2C+Qiwu&rft.au=Su%2C+Jiaojiao&rft.au=Yang%2C+Chunhua&rft.au=Gui%2C+Weihua&rft.date=2022&rft.pub=IEEE&rft.issn=0018-9456&rft.volume=71&rft.spage=1&rft.epage=13&rft_id=info:doi/10.1109%2FTIM.2022.3165270&rft.externalDocID=9757932
thumbnail_l	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0018-9456&client=summon
thumbnail_m	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0018-9456&client=summon
thumbnail_s	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0018-9456&client=summon