A Transformer-Driven Clustering Framework for Image-Based Document Segregation of OCR-Extracted Data

A Transformer-Driven Clustering Framework for Image-Based Document Segregation of OCR-Extracted Data | 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 A Transformer-Driven Clustering Framework for Image-Based Document Segregation of OCR-Extracted Data Sahaya Beni Prathiba, Dhanalakshmi Ranganayakulu, Vijay Arunachalam, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7372555/v1 This work is licensed under a CC BY 4.0 License Status: Under Revision Version 1 posted 11 You are reading this latest preprint version Abstract The rapid increase in image-based documents across industries such as healthcare, law, and government underscores the need for efficient techniques to organize and extract meaningful insights from unstructured datasets. Traditional methods, including manual sorting and rule-based clustering, fail to effectively handle large-scale, noisy, and heterogeneous datasets, highlighting a significant research gap. To address this, we propose the Enhancing Document Segregation (EDS) model, a framework designed to cluster image-based datasets using a combination of Optical Character Recognition (OCR), semantic analysis, and advanced clustering algorithms. The EDS pipeline extracts text from images via OCR, preprocesses the data to eliminate noise, and generates embeddings using transformer-based models to capture semantic relationships. These embeddings are clustered using K-means, DBSCAN, Gaussian Mixture Models, and agglomerative clustering techniques to verify changes in variable data. Empirical analysis demonstrates the robustness of the EDS model in improving clustering accuracy and efficiency, particularly in noisy and complex datasets. Integrating theoretical foundations with practical clustering methodologies ensures the EDS model delivers a scalable solution for real-world challenges, enhancing document organization and retrieval in critical domains. Physical sciences/Engineering Physical sciences/Mathematics and computing Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Revision Version 1 posted Editorial decision: Revision requested 08 Dec, 2025 Reviewers agreed at journal 09 Nov, 2025 Reviews received at journal 30 Oct, 2025 Reviewers agreed at journal 30 Oct, 2025 Reviews received at journal 25 Oct, 2025 Reviewers agreed at journal 20 Oct, 2025 Reviewers invited by journal 03 Sep, 2025 Editor assigned by journal 31 Aug, 2025 Editor invited by journal 29 Aug, 2025 Submission checks completed at journal 28 Aug, 2025 First submitted to journal 28 Aug, 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-7372555","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":509872698,"identity":"50d095e7-c74c-46c1-9581-e3adf96cd8ea","order_by":0,"name":"Sahaya Beni Prathiba","email":"","orcid":"","institution":"Vellore Institute of Technology","correspondingAuthor":false,"prefix":"","firstName":"Sahaya","middleName":"Beni","lastName":"Prathiba","suffix":""},{"id":509872699,"identity":"e6be514e-a776-4eb2-969f-d9d28b59df2c","order_by":1,"name":"Dhanalakshmi Ranganayakulu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA/ElEQVRIiWNgGAWjYDCCGyCigMGAQYKB4QADgw0DG0RcgoAWA7iWNAY2NlK0AMFhBpg1OAHf7eZnD34Y2BnLz+5OPMC453xin3wD44cfDBZ5uLRI3jlmbthjkGxmcOfshgMMz24ntrExMEv2MEgU49JicCPBTILHgNnGQCIXqOXAbWOQX6SBfklswKkl/ZvkH4N6G/kZYC3nQFqYf+PXkmMmzWNw2IzhBljLATmgFja8tkjeyCmTljE4bmwA0pJwIBmoJbHNsscAtxa+G+nbJN9UVBvOn5G7+cOHA3Y88s2HD9/4UVGHUwsqSACTjA2gmBoFo2AUjIJRQAEAAILuVDZfgU46AAAAAElFTkSuQmCC","orcid":"","institution":"Vellore Institute of Technology","correspondingAuthor":true,"prefix":"","firstName":"Dhanalakshmi","middleName":"","lastName":"Ranganayakulu","suffix":""},{"id":509872700,"identity":"a8cbe693-5095-46cc-ae45-c10ae60894b6","order_by":2,"name":"Vijay Arunachalam","email":"","orcid":"","institution":"Vellore Institute of Technology","correspondingAuthor":false,"prefix":"","firstName":"Vijay","middleName":"","lastName":"Arunachalam","suffix":""},{"id":509872701,"identity":"e148104b-7987-4b19-b05a-9f006e952ced","order_by":3,"name":"Uodit Vishvaa","email":"","orcid":"","institution":"Vellore Institute of Technology","correspondingAuthor":false,"prefix":"","firstName":"Uodit","middleName":"","lastName":"Vishvaa","suffix":""},{"id":509872703,"identity":"580f49cc-12a6-468e-850e-96d1381dc9fa","order_by":4,"name":"Veera Karthick","email":"","orcid":"","institution":"Vellore Institute of Technology","correspondingAuthor":false,"prefix":"","firstName":"Veera","middleName":"","lastName":"Karthick","suffix":""},{"id":509872705,"identity":"d06989ae-885d-415d-886f-644459109df2","order_by":5,"name":"Suriya Priya R Asaithambi","email":"","orcid":"","institution":"National University of Singapore","correspondingAuthor":false,"prefix":"","firstName":"Suriya","middleName":"Priya R","lastName":"Asaithambi","suffix":""}],"badges":[],"createdAt":"2025-08-14 10:16:24","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7372555/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7372555/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":90978988,"identity":"e1d23800-10f7-40f6-8b60-b36b1d827cba","added_by":"auto","created_at":"2025-09-10 09:05:13","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":962323,"visible":true,"origin":"","legend":"","description":"","filename":"ScientificReportsEDSOCRFinal.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7372555/v1_covered_ac11fb1a-cf42-4495-8bf9-d65ecc269a2d.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"A Transformer-Driven Clustering Framework for Image-Based Document Segregation of OCR-Extracted Data","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"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":"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":"","lastPublishedDoi":"10.21203/rs.3.rs-7372555/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7372555/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"The rapid increase in image-based documents across industries such as healthcare, law, and government underscores the need for efficient techniques to organize and extract meaningful insights from unstructured datasets. Traditional methods, including manual sorting and rule-based clustering, fail to effectively handle large-scale, noisy, and heterogeneous datasets, highlighting a significant research gap. To address this, we propose the Enhancing Document Segregation (EDS) model, a framework designed to cluster image-based datasets using a combination of Optical Character Recognition (OCR), semantic analysis, and advanced clustering algorithms. The EDS pipeline extracts text from images via OCR, preprocesses the data to eliminate noise, and generates embeddings using transformer-based models to capture semantic relationships. These embeddings are clustered using K-means, DBSCAN, Gaussian Mixture Models, and agglomerative clustering techniques to verify changes in variable data. Empirical analysis demonstrates the robustness of the EDS model in improving clustering accuracy and efficiency, particularly in noisy and complex datasets. Integrating theoretical foundations with practical clustering methodologies ensures the EDS model delivers a scalable solution for real-world challenges, enhancing document organization and retrieval in critical domains.","manuscriptTitle":"A Transformer-Driven Clustering Framework for Image-Based Document Segregation of OCR-Extracted Data","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-10 08:57:08","doi":"10.21203/rs.3.rs-7372555/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-12-08T08:07:11+00:00","index":"","fulltext":""},{"type":"reviewerAgreed","content":"157443216733337724145860632439406634039","date":"2025-11-09T06:20:49+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-30T13:27:11+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"116562572346552491507727464897781919946","date":"2025-10-30T11:42:56+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-26T03:10:05+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"131061599769486641496238344627777088725","date":"2025-10-21T02:21:40+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-09-03T19:45:20+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-08-31T23:02:37+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-08-29T11:02:07+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-08-28T05:49:50+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-08-28T05:48:38+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":"6a29662f-266f-4e1c-bb51-4e57a56fb636","owner":[],"postedDate":"September 10th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"in-revision","subjectAreas":[{"id":54156586,"name":"Physical sciences/Engineering"},{"id":54156587,"name":"Physical sciences/Mathematics and computing"}],"tags":[],"updatedAt":"2026-02-24T06:10:36+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-10 08:57:08","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7372555","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7372555","identity":"rs-7372555","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}