MultiLingual Scene Text Detection via Group-Specific Models

MultiLingual Scene Text Detection via Group-Specific Models | 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 MultiLingual Scene Text Detection via Group-Specific Models Jhonatas Conceição, Manuel Córdova, Allan Pinto, Ricardo da S. Torres, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8117789/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 10 You are reading this latest preprint version Abstract Multilingual scene text detection has received significant attention in recent years. The challenge of this task is to design text detectors capable of handling a wide range of variability, such as font size, font style, color, complex background, and the presence of multilingual text in the same scene. Unlike current approaches that rely on a single model to detect text from all languages, we propose a group-specific modeling strategy for multilingual text detection. Our method clusters languages with similar visual structural characteristics, trains dedicated detectors for each group, and then fuses their results through a general object detector. In order to evaluate and compare the proposed methodology against state-of-the-art methods, the MLT-2019 dataset was used. The experiments demonstrated the effectiveness of our approach, outperforming the general single-model approach used widely in the literature by at least 6.06 percentage points when evaluated per language. Additionally, our approach surpassed the state-of-the-art method in terms of F1-macro on MLT-2019 and is the best-performing method in four out of seven languages: Arabic, Bangla, Hindi, and Chinese. Text detection scene text multilingual convolutional neural network group-specific models Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 08 Apr, 2026 Reviews received at journal 02 Apr, 2026 Reviewers agreed at journal 30 Mar, 2026 Reviews received at journal 20 Jan, 2026 Reviewers agreed at journal 20 Jan, 2026 Reviewers agreed at journal 19 Jan, 2026 Reviewers invited by journal 17 Dec, 2025 Editor assigned by journal 19 Nov, 2025 Submission checks completed at journal 16 Nov, 2025 First submitted to journal 14 Nov, 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-8117789","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":561811637,"identity":"5bb457fb-62ec-4c1c-a72f-cb7f05105da3","order_by":0,"name":"Jhonatas Conceição","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA90lEQVRIiWNgGAWjYBACxmbmg49/8NjIMRwA8yXAJDNeLe1sycYMMmnGxGth4Ocxk2awOZzYcABJEK8W5magloKctPS+44cfMPPusJDnn3b2AHPhHnx+YSu2nnHGJnfmmTQDZt4zEoYzbuclMM94hk8L88YbvD1puRsO5DAw87ZJJDDczjFg5jmATwuDgQTvv8PpBuffQLTIE9bCYiTNw3M4weAG1BYDwlrYkg1n8KQZzrzxzODg3DYJw41AvxyegUeLYf/hgw8+8NjI851PfvjgbVudvNzt3IOPC/BpaUDiQNXxwBjYgTwWMR58GkbBKBgFo2AEAgCWOVIXl8wPagAAAABJRU5ErkJggg==","orcid":"","institution":"State University of Campinas","correspondingAuthor":true,"prefix":"","firstName":"Jhonatas","middleName":"","lastName":"Conceição","suffix":""},{"id":561811640,"identity":"3a7a5a9c-817a-4956-9b61-3ddaf8c65c09","order_by":1,"name":"Manuel Córdova","email":"","orcid":"","institution":"Wageningen University \u0026 Research","correspondingAuthor":false,"prefix":"","firstName":"Manuel","middleName":"","lastName":"Córdova","suffix":""},{"id":561811642,"identity":"90fe25ce-eef5-4dc4-ab45-8b0dfe404087","order_by":2,"name":"Allan Pinto","email":"","orcid":"","institution":"Brazilian Center for Research in Energy and Materials","correspondingAuthor":false,"prefix":"","firstName":"Allan","middleName":"","lastName":"Pinto","suffix":""},{"id":561811644,"identity":"a42b011e-96ce-46ed-8ce1-226091ee320a","order_by":3,"name":"Ricardo da S. Torres","email":"","orcid":"","institution":"Wageningen University \u0026 Research","correspondingAuthor":false,"prefix":"","firstName":"Ricardo","middleName":"da S.","lastName":"Torres","suffix":""},{"id":561811646,"identity":"49b7ebb2-7a19-4af4-b931-a006715ef758","order_by":4,"name":"Helio Pedrini","email":"","orcid":"","institution":"State University of Campinas","correspondingAuthor":false,"prefix":"","firstName":"Helio","middleName":"","lastName":"Pedrini","suffix":""}],"badges":[],"createdAt":"2025-11-14 19:23:15","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8117789/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8117789/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":98576601,"identity":"e9027cd1-110c-4191-b8b9-f69df8aeb909","added_by":"auto","created_at":"2025-12-19 07:30:11","extension":"json","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6172,"visible":true,"origin":"","legend":"","description":"","filename":"ca982138a477453e99f3f7ed29239581.json","url":"https://assets-eu.researchsquare.com/files/rs-8117789/v1/aae08cb7ba5f2c66b223a957.json"},{"id":98627702,"identity":"4754f7f4-842c-40fc-be56-7458d6d18a0f","added_by":"auto","created_at":"2025-12-19 17:10:34","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":6863206,"visible":true,"origin":"","legend":"","description":"","filename":"MultiligualTextDetectionIJDAR2026.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8117789/v1_covered_940d072b-cd4f-4b4c-85d2-c3553f8b6774.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"MultiLingual Scene Text Detection via Group-Specific Models","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":"international-journal-on-document-analysis-and-recognition-ijdar","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"ijda","sideBox":"Learn more about [International Journal on Document Analysis and Recognition (IJDAR)](http://link.springer.com/journal/10032)","snPcode":"10032","submissionUrl":"https://submission.nature.com/new-submission/10032/3","title":"International Journal on Document Analysis and Recognition (IJDAR)","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Text detection, scene text, multilingual, convolutional neural network, group-specific models","lastPublishedDoi":"10.21203/rs.3.rs-8117789/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8117789/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Multilingual scene text detection has received significant attention in recent years. The challenge of this task is to design text detectors capable of handling a wide range of variability, such as font size, font style, color, complex background, and the presence of multilingual text in the same scene. Unlike current approaches that rely on a single model to detect text from all languages, we propose a group-specific modeling strategy for multilingual text detection. Our method clusters languages with similar visual structural characteristics, trains dedicated detectors for each group, and then fuses their results through a general object detector. In order to evaluate and compare the proposed methodology against state-of-the-art methods, the MLT-2019 dataset was used. The experiments demonstrated the effectiveness of our approach, outperforming the general single-model approach used widely in the literature by at least 6.06 percentage points when evaluated per language. Additionally, our approach surpassed the state-of-the-art method in terms of F1-macro on MLT-2019 and is the best-performing method in four out of seven languages: Arabic, Bangla, Hindi, and Chinese.","manuscriptTitle":"MultiLingual Scene Text Detection via Group-Specific Models","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-19 07:30:07","doi":"10.21203/rs.3.rs-8117789/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-04-08T07:00:29+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-02T18:51:11+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"292701193720830884870366256893683987159","date":"2026-03-30T14:20:49+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-01-20T21:58:37+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"337168204660645764131055232366866190852","date":"2026-01-20T20:25:06+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"62889627534655823445653178228469599135","date":"2026-01-19T17:44:35+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-12-17T11:47:18+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-11-19T05:25:03+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-11-17T02:58:19+00:00","index":"","fulltext":""},{"type":"submitted","content":"International Journal on Document Analysis and Recognition (IJDAR)","date":"2025-11-14T19:14:03+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"international-journal-on-document-analysis-and-recognition-ijdar","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"ijda","sideBox":"Learn more about [International Journal on Document Analysis and Recognition (IJDAR)](http://link.springer.com/journal/10032)","snPcode":"10032","submissionUrl":"https://submission.nature.com/new-submission/10032/3","title":"International Journal on Document Analysis and Recognition (IJDAR)","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"7e9180c8-c104-450d-ba4c-0781d38ce9ec","owner":[],"postedDate":"December 19th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2025-12-19T07:30:07+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-19 07:30:07","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8117789","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8117789","identity":"rs-8117789","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}