SGCNet: Multi-task Self-Gated Convolution based UNet with Multi Scale Convolution and Global Attention Module

SGCNet: Multi-task Self-Gated Convolution based UNet with Multi Scale Convolution and Global Attention Module | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article SGCNet: Multi-task Self-Gated Convolution based UNet with Multi Scale Convolution and Global Attention Module Xiaohu Zhang, Haifeng Huang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3811411/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Concrete cracks pose significant safety hazards to concrete buildings, and semantic segmentation models based on deep learning has achieved the state-of-arts results in concrete crack detection. However, there are some limitations in these models. Firstly, stains in crack images would inter-fere the extraction of crack features. Secondly, current crack detection models usually use convo-lutional kernels of uniform size, which could not adapt cracks with varying scales. Thirdly, due to the diversity and irregularity of cracks' edges, existing semantic segmentation models could not effectively segment these edges. Finally, current mainstream model only use serial stacked convo-lution operation, which is limited to extract global features. To solve these problems above, a Self-Gated Convolution (SGC) based U-Net model with Multi Scale Convolution Module (MSCM) and Global Attention Module (GAM) named as SGCNet is proposed by us. In this model, SGC is used to remove interference features and MSCM is used for extract features of different scales with dilated convolutions. Additionally, edge detection of cracks is used as the second task to en-hance the edge feature extraction ability of the model. Finally, GAM is used to extract global con-textual information. Experimental results show that the SGCNet proposed by us could achieve the accuracy of 0.942, 0.836 and 0.991 on the Cracktree200, CRACK500 and CFD datasets respectively, which is higher than that of the traditional models. Artificial Intelligence Machine Learning Crack Detection Image Segmentation Application Crack Segmentation Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-3811411","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":263840078,"identity":"0372d102-7670-4c71-becd-83da60eaceb9","order_by":0,"name":"Xiaohu Zhang","email":"","orcid":"","institution":"Sun Yat-sen University","correspondingAuthor":false,"prefix":"","firstName":"Xiaohu","middleName":"","lastName":"Zhang","suffix":""},{"id":263840079,"identity":"ea34af03-f579-44d3-bb3f-31706c2cc5e7","order_by":1,"name":"Haifeng Huang","email":"data:image/png;base64,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","orcid":"","institution":"Sun Yat-sen University","correspondingAuthor":true,"prefix":"","firstName":"Haifeng","middleName":"","lastName":"Huang","suffix":""}],"badges":[],"createdAt":"2023-12-27 08:59:12","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3811411/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3811411/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":49178801,"identity":"5008dd67-810d-492b-abd5-13dec886090e","added_by":"auto","created_at":"2024-01-04 13:37:27","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":675654,"visible":true,"origin":"","legend":"","description":"","filename":"SGCNetMultitaskSelfGatedConvolutionbasedUNetwithMultiScaleConvolutionandGlobalAttentionModule.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3811411/v1_covered_9dcdf350-e5f8-4be5-a3d2-a1e5b982dab6.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"SGCNet: Multi-task Self-Gated Convolution based UNet with Multi Scale Convolution and Global Attention Module","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Artificial Intelligence, Machine Learning, Crack Detection, Image Segmentation Application, Crack Segmentation","lastPublishedDoi":"10.21203/rs.3.rs-3811411/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3811411/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eConcrete cracks pose significant safety hazards to concrete buildings, and semantic segmentation models based on deep learning has achieved the state-of-arts results in concrete crack detection. However, there are some limitations in these models. Firstly, stains in crack images would inter-fere the extraction of crack features. Secondly, current crack detection models usually use convo-lutional kernels of uniform size, which could not adapt cracks with varying scales. Thirdly, due to the diversity and irregularity of cracks' edges, existing semantic segmentation models could not effectively segment these edges. Finally, current mainstream model only use serial stacked convo-lution operation, which is limited to extract global features. To solve these problems above, a Self-Gated Convolution (SGC) based U-Net model with Multi Scale Convolution Module (MSCM) and Global Attention Module (GAM) named as SGCNet is proposed by us. In this model, SGC is used to remove interference features and MSCM is used for extract features of different scales with dilated convolutions. Additionally, edge detection of cracks is used as the second task to en-hance the edge feature extraction ability of the model. Finally, GAM is used to extract global con-textual information. Experimental results show that the SGCNet proposed by us could achieve the accuracy of 0.942, 0.836 and 0.991 on the Cracktree200, CRACK500 and CFD datasets respectively, which is higher than that of the traditional models.\u003c/p\u003e","manuscriptTitle":"SGCNet: Multi-task Self-Gated Convolution based UNet with Multi Scale Convolution and Global Attention Module","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-01-01 06:50:11","doi":"10.21203/rs.3.rs-3811411/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"c911318e-9889-45ef-9d84-f9dfc385fe95","owner":[],"postedDate":"January 1st, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-01-09T16:15:05+00:00","versionOfRecord":[],"versionCreatedAt":"2024-01-01 06:50:11","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-3811411","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3811411","identity":"rs-3811411","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}