Fed-MSVT: Federated Multi-Scale Vision Transformer with Adaptive Client Aggregation for Industrial Defect Detection

Fed-MSVT: Federated Multi-Scale Vision Transformer with Adaptive Client Aggregation for Industrial Defect Detection | 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 Article Fed-MSVT: Federated Multi-Scale Vision Transformer with Adaptive Client Aggregation for Industrial Defect Detection Sailaja Are, N. Gopala Krishna, Jagadeesh Thati, Kistam Gopi This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6201071/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 01 Oct, 2025 Read the published version in Scientific Reports → Version 1 posted 11 You are reading this latest preprint version Abstract Defect detection in industrial applications is essential for maintaining product quality and operational efficiency. However, traditional deep learning methods require centralized data collection, raising privacy concerns and limiting adaptability in distributed manufacturing environments. To overcome these challenges, we propose Fed-MSVT, a Federated Multi-Scale Vision Transformer with Adaptive Client Aggregation for industrial defect detection. Our approach leverages multi-scale Vision Transformers (MSVTs) to capture both fine-grained local defects and global structural patterns, enhancing detection accuracy across diverse defect types. Unlike conventional federated learning models, we introduce an Adaptive Client Aggregation (ACA) mechanism that dynamically assigns weights to client models based on data quality, domain shift, and consistency. Additionally, a Contrastive Feature Alignment (CFA) module mitigates inter-client domain discrepancies, improving generalization. Evaluations on multiple defect datasets demonstrate superior accuracy, robustness, and scalability compared to existing approaches, enabling real-time, privacy-preserving, and adaptive defect detection for smart manufacturing systems Physical sciences/Engineering Physical sciences/Engineering/Electrical and electronic engineering Physical sciences/Engineering/Energy infrastructure Industrial defect detection federated learning multi-scale Vision Transformer (MSVT) adaptive client aggregation (ACA) contrastive feature alignment (CFA) smart manufacturing Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 01 Oct, 2025 Read the published version in Scientific Reports → Version 1 posted Editorial decision: Revision requested 25 Jun, 2025 Reviews received at journal 24 Jun, 2025 Reviewers agreed at journal 16 Jun, 2025 Reviews received at journal 18 May, 2025 Reviewers agreed at journal 17 May, 2025 Reviewers agreed at journal 20 Apr, 2025 Reviewers invited by journal 18 Mar, 2025 Editor assigned by journal 18 Mar, 2025 Editor invited by journal 18 Mar, 2025 Submission checks completed at journal 14 Mar, 2025 First submitted to journal 11 Mar, 2025 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-6201071","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":430677393,"identity":"d86ce892-e823-43ed-91de-41333e4695c1","order_by":0,"name":"Sailaja Are","email":"","orcid":"","institution":"Tirumala Engineering College","correspondingAuthor":false,"prefix":"","firstName":"Sailaja","middleName":"","lastName":"Are","suffix":""},{"id":430677395,"identity":"a6d8cdb5-03f2-4dba-b46e-8198232f4ccb","order_by":1,"name":"N. Gopala Krishna","email":"","orcid":"","institution":"Tirumala Engineering College","correspondingAuthor":false,"prefix":"","firstName":"N.","middleName":"Gopala","lastName":"Krishna","suffix":""},{"id":430677397,"identity":"b3732852-22b9-4442-9652-496ddc1dc9d4","order_by":2,"name":"Jagadeesh Thati","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABBklEQVRIiWNgGAWjYBACxgYgwQPEBhDSBiTWeIA4LWxA8oBBGlgMrxYwQGhhOAwWwKuFuf34xQ9vKmzszeV7D37+UHDebm37YaAtNTbROB3Wk1MsOedMWuLONr5kiQMGt5O3nUkEajmWltuA0y85CdK8bYcTDI7xGIC1mB0AamFsOIxbS/+b5N+8//7bA7UY/zhgcC7Z7PxDAlpmpB+T5gWq2XCMxwxoywE7sxuEbJnxhs1yzrHkxA3HcswszhgkJ5jdANqSgMcvhv3pj2+8qbGzNzh8xvhGxR87e7Pz6Q8ffKixwa2lgccARSARrDIBh3IQkGdgf4AiYI9H8SgYBaNgFIxQAADX3mnfbAZwHAAAAABJRU5ErkJggg==","orcid":"","institution":"Tirumala Engineering College","correspondingAuthor":true,"prefix":"","firstName":"Jagadeesh","middleName":"","lastName":"Thati","suffix":""},{"id":430677399,"identity":"5b113e70-07b2-4370-9b0b-2752945726f2","order_by":3,"name":"Kistam Gopi","email":"","orcid":"","institution":"Tirumala Engineering College","correspondingAuthor":false,"prefix":"","firstName":"Kistam","middleName":"","lastName":"Gopi","suffix":""}],"badges":[],"createdAt":"2025-03-11 07:53:27","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6201071/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6201071/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41598-025-12390-z","type":"published","date":"2025-10-01T15:57:14+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":92883700,"identity":"e202611c-905a-4f92-8670-ea62bb09c07b","added_by":"auto","created_at":"2025-10-06 16:08:10","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":818650,"visible":true,"origin":"","legend":"","description":"","filename":"RevisedManuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6201071/v1_covered_781fed0b-e07a-4e5e-a948-d7e8c3188bb6.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Fed-MSVT: Federated Multi-Scale Vision Transformer with Adaptive Client Aggregation for Industrial Defect Detection","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":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Industrial defect detection, federated learning, multi-scale Vision Transformer (MSVT), adaptive client aggregation (ACA), contrastive feature alignment (CFA), smart manufacturing","lastPublishedDoi":"10.21203/rs.3.rs-6201071/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6201071/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eDefect detection in industrial applications is essential for maintaining product quality and operational efficiency. However, traditional deep learning methods require centralized data collection, raising privacy concerns and limiting adaptability in distributed manufacturing environments. To overcome these challenges, we propose Fed-MSVT, a Federated Multi-Scale Vision Transformer with Adaptive Client Aggregation for industrial defect detection. Our approach leverages multi-scale Vision Transformers (MSVTs) to capture both fine-grained local defects and global structural patterns, enhancing detection accuracy across diverse defect types. Unlike conventional federated learning models, we introduce an Adaptive Client Aggregation (ACA) mechanism that dynamically assigns weights to client models based on data quality, domain shift, and consistency. Additionally, a Contrastive Feature Alignment (CFA) module mitigates inter-client domain discrepancies, improving generalization. Evaluations on multiple defect datasets demonstrate superior accuracy, robustness, and scalability compared to existing approaches, enabling real-time, privacy-preserving, and adaptive defect detection for smart manufacturing systems\u003c/p\u003e","manuscriptTitle":"Fed-MSVT: Federated Multi-Scale Vision Transformer with Adaptive Client Aggregation for Industrial Defect Detection","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-03-20 03:57:06","doi":"10.21203/rs.3.rs-6201071/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-06-25T05:45:14+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-06-24T08:04:17+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"202951940794562041061952001049677728267","date":"2025-06-16T07:44:29+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-05-18T06:27:04+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"225191099188072547532738012283778360526","date":"2025-05-17T13:54:44+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"279333190575843508613115907334400802222","date":"2025-04-20T05:55:15+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-03-18T05:49:33+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-03-18T05:43:01+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-03-18T04:21:46+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-03-15T02:14:22+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-03-11T07:50:12+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"1d65fa47-cb3c-46f8-af92-cfeb6154454a","owner":[],"postedDate":"March 20th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":45871300,"name":"Physical sciences/Engineering"},{"id":45871301,"name":"Physical sciences/Engineering/Electrical and electronic engineering"},{"id":45871302,"name":"Physical sciences/Engineering/Energy infrastructure"}],"tags":[],"updatedAt":"2025-10-06T16:00:32+00:00","versionOfRecord":{"articleIdentity":"rs-6201071","link":"https://doi.org/10.1038/s41598-025-12390-z","journal":{"identity":"scientific-reports","isVorOnly":false,"title":"Scientific Reports"},"publishedOn":"2025-10-01 15:57:14","publishedOnDateReadable":"October 1st, 2025"},"versionCreatedAt":"2025-03-20 03:57:06","video":"","vorDoi":"10.1038/s41598-025-12390-z","vorDoiUrl":"https://doi.org/10.1038/s41598-025-12390-z","workflowStages":[]},"version":"v1","identity":"rs-6201071","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6201071","identity":"rs-6201071","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}