U2BIL:A Two-phase Class Separation Method for Unbalanced Tunnel Defects via Class Incremental Learning

U2BIL:A Two-phase Class Separation Method for Unbalanced Tunnel Defects via Class Incremental Learning | 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 U2BIL:A Two-phase Class Separation Method for Unbalanced Tunnel Defects via Class Incremental Learning Cai Yiwei, Gao Xinwen, Yang Yumeng, Feng Xinyang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4936065/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 24 Dec, 2024 Read the published version in Signal, Image and Video Processing → Version 1 posted 5 You are reading this latest preprint version Abstract With more and more tunnel excavation projects, tunnel defects occur frequently. To reduce the cost of manual overhaul and maintenance, machine vision inspection methods are gradually emerging, but the traditional machine learning inspection methods cannot satisfy the continuous learning to update the knowledge of the prototype class without repeated access to the defect data. Incremental learning, as a continuous learning paradigm, can accomplish this task and alleviate the catastrophic forgetting phenomenon, but traditional incremental learning is difficult to fit the extremely unbalanced data distribution of the generalized realistic distribution. In this paper, based on the extremely unbalanced data distribution of tunnels, we propose a two-phase learning paradigm applicable to incremental learning to mitigate the overfitting of header data when the classes are unbalanced and incorporate the DR loss under the ETF classifier to ensure the maximum symmetric separation of the interclass prototypes under the classifier space. The proposed method in this paper obtains better performance under tunnel data compared to other incremental SOTAS. Incremental Learning Tunnel Defects Imbalanced Learning Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 24 Dec, 2024 Read the published version in Signal, Image and Video Processing → Version 1 posted Editorial decision: Revision requested 22 Aug, 2024 Reviewers invited by journal 22 Aug, 2024 Editor assigned by journal 19 Aug, 2024 Submission checks completed at journal 19 Aug, 2024 First submitted to journal 19 Aug, 2024 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-4936065","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":343931760,"identity":"40f28b50-876e-4cfe-9e3f-b39b52c7ecb3","order_by":0,"name":"Cai Yiwei","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAyklEQVRIiWNgGAWjYBACCSBmbGBgkAOzGNhI0GJMupbEBqK1SM5If/hxZltd+vzZPQYMH8oOM/DPbsCvRVoiIVlyY9vh3A13zhgwzjh3mEHizgH8WuQkEg5IPmw7kLtBIseAmbftMIOBRAIhLYnNPx8CHSY/A6jlLzFapCWS2YAOY05guAHUwkiMFsmeZ2yWQC8YbriRVnCw51w6j8QNAlokjqc/vtlTVicvPyN544MfZdZy/DMIaEEBB4CYhwT1o2AUjIJRMApwAQA3nUJ0d0hAeAAAAABJRU5ErkJggg==","orcid":"","institution":"Shanghai University","correspondingAuthor":true,"prefix":"","firstName":"Cai","middleName":"","lastName":"Yiwei","suffix":""},{"id":343931761,"identity":"d0c03c5b-0bbf-4e93-9b0b-f6766f8f14a8","order_by":1,"name":"Gao Xinwen","email":"","orcid":"","institution":"Shanghai University","correspondingAuthor":false,"prefix":"","firstName":"Gao","middleName":"","lastName":"Xinwen","suffix":""},{"id":343931762,"identity":"c950011c-3216-49f2-97db-9e007b89584f","order_by":2,"name":"Yang Yumeng","email":"","orcid":"","institution":"Shanghai University","correspondingAuthor":false,"prefix":"","firstName":"Yang","middleName":"","lastName":"Yumeng","suffix":""},{"id":343931763,"identity":"d4dd1509-9c6b-481d-9ecc-5daad2c65408","order_by":3,"name":"Feng Xinyang","email":"","orcid":"","institution":"Shanghai University","correspondingAuthor":false,"prefix":"","firstName":"Feng","middleName":"","lastName":"Xinyang","suffix":""}],"badges":[],"createdAt":"2024-08-19 05:42:11","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4936065/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4936065/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s11760-024-03651-x","type":"published","date":"2024-12-24T15:57:20+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":72640538,"identity":"96c0ef97-a7ca-4208-9100-55ab227f2dac","added_by":"auto","created_at":"2024-12-30 16:06:35","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":705018,"visible":true,"origin":"","legend":"","description":"","filename":"ArticleTitle6.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4936065/v1_covered_499bd0e2-7edd-4878-88c8-b3e95a33d353.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"U2BIL:A Two-phase Class Separation Method for Unbalanced Tunnel Defects via Class Incremental Learning","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"signal-image-and-video-processing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"sivp","sideBox":"Learn more about [Signal, Image and Video Processing](http://link.springer.com/journal/11760)","snPcode":"11760","submissionUrl":"https://submission.nature.com/new-submission/11760/3","title":"Signal, Image and Video Processing","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Incremental Learning, Tunnel Defects, Imbalanced Learning","lastPublishedDoi":"10.21203/rs.3.rs-4936065/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4936065/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"With more and more tunnel excavation projects, tunnel defects occur frequently. To reduce the cost of manual overhaul and maintenance, machine vision inspection methods are gradually emerging, but the traditional machine learning inspection methods cannot satisfy the continuous learning to update the knowledge of the prototype class without repeated access to the defect data. Incremental learning, as a continuous learning paradigm, can accomplish this task and alleviate the catastrophic forgetting phenomenon, but traditional incremental learning is difficult to fit the extremely unbalanced data distribution of the generalized realistic distribution. In this paper, based on the extremely unbalanced data distribution of tunnels, we propose a two-phase learning paradigm applicable to incremental learning to mitigate the overfitting of header data when the classes are unbalanced and incorporate the DR loss under the ETF classifier to ensure the maximum symmetric separation of the interclass prototypes under the classifier space. The proposed method in this paper obtains better performance under tunnel data compared to other incremental SOTAS.","manuscriptTitle":"U2BIL:A Two-phase Class Separation Method for Unbalanced Tunnel Defects via Class Incremental Learning","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-09-16 18:48:41","doi":"10.21203/rs.3.rs-4936065/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2024-08-23T02:06:33+00:00","index":"","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-08-23T02:05:38+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-08-19T15:24:01+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-08-19T15:22:50+00:00","index":"","fulltext":""},{"type":"submitted","content":"Signal, Image and Video Processing","date":"2024-08-19T05:40:53+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"signal-image-and-video-processing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"sivp","sideBox":"Learn more about [Signal, Image and Video Processing](http://link.springer.com/journal/11760)","snPcode":"11760","submissionUrl":"https://submission.nature.com/new-submission/11760/3","title":"Signal, Image and Video Processing","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"a3278cf8-aed5-4712-89e2-0183dcafaffb","owner":[],"postedDate":"September 16th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2024-12-30T16:00:33+00:00","versionOfRecord":{"articleIdentity":"rs-4936065","link":"https://doi.org/10.1007/s11760-024-03651-x","journal":{"identity":"signal-image-and-video-processing","isVorOnly":false,"title":"Signal, Image and Video Processing"},"publishedOn":"2024-12-24 15:57:20","publishedOnDateReadable":"December 24th, 2024"},"versionCreatedAt":"2024-09-16 18:48:41","video":"","vorDoi":"10.1007/s11760-024-03651-x","vorDoiUrl":"https://doi.org/10.1007/s11760-024-03651-x","workflowStages":[]},"version":"v1","identity":"rs-4936065","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4936065","identity":"rs-4936065","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}