Benchmarking OCR and Vision-Language Models for Turkish Text Recognition: A Comprehensive Evaluation Using Synthetic Data

preprint OA: closed
Full text JSON View at publisher
Full text 27,058 characters · extracted from preprint-html · click to expand
Benchmarking OCR and Vision-Language Models for Turkish Text Recognition: A Comprehensive Evaluation Using Synthetic 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 Research Article Benchmarking OCR and Vision-Language Models for Turkish Text Recognition: A Comprehensive Evaluation Using Synthetic Data Yasin Yılmaz, Erol Görkem Hanoğlu, Ayşe Gül Özkan, Kasım Öztoprak This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7797886/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 11 You are reading this latest preprint version Abstract Purpose: We present the first systematic benchmark evaluation of Optical Character Recognition (OCR) and Vision-Language Models (VLMs) for Turkish text recognition, addressing a critical gap in low-resource language processing. Turkish, with its agglutinative structure and unique characters (ç, ğ, ı, İ, ö, ş, ü), poses challenges for models trained on high-resource languages such as English. Methods: We developed a synthetic Turkish dataset of 6,600 images spanning three main text types: printed, handwritten, scene text. The dataset includes variations such as the presence of Turkish characters, effects of word length, sentence versus word recognition, and various distortion types (rotation, resolution, noise, and blur).Our evaluation compares three different model categories: traditional OCR systems, open-source VLMs, and commercial VLMs. Results: The results show that modern VLMs significantly outperform traditional OCR approaches, with GPT-4o and Qwen2.5-VL models demonstrating superior performance. Notably, images containing Turkish-specific characters posed significant challenges for all models, with only GPT-4o maintaining stable performance. This highlights the critical impact of training dataset composition on multilingual performance. While the agglutinative word structure did not significantly affect recognition accuracy, handwritten text recognition remains a persistent challenge across all evaluated systems. Conclusion: The open-source Qwen2.5-VL model achieved comparable performance to the commercial GPT-4o despite having fewer parameters, showing strong potential as a computationally efficient alternative. This benchmark study establishes a standardized evaluation framework for Turkish text recognition research. To support future research in this domain, we publicly release the synthetic dataset, enabling reproducible research in low-resource language text recognition. OCR VLMs Turkish Text Recognition Synthetic Dataset Full Text Additional Declarations No competing interests reported. Supplementary Files ESM1.zip Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 18 Mar, 2026 Reviews received at journal 08 Jan, 2026 Reviews received at journal 06 Jan, 2026 Reviewers agreed at journal 18 Nov, 2025 Reviewers agreed at journal 14 Nov, 2025 Reviewers agreed at journal 13 Nov, 2025 Reviewers agreed at journal 12 Nov, 2025 Reviewers invited by journal 11 Nov, 2025 Editor assigned by journal 09 Oct, 2025 Submission checks completed at journal 08 Oct, 2025 First submitted to journal 07 Oct, 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-7797886","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":527097106,"identity":"6fe122c3-f96e-4ecf-9557-1d610e9a0ee6","order_by":0,"name":"Yasin Yılmaz","email":"","orcid":"","institution":"Konya Food and Agriculture University","correspondingAuthor":false,"prefix":"","firstName":"Yasin","middleName":"","lastName":"Yılmaz","suffix":""},{"id":527097107,"identity":"f569e89d-dbc3-44c4-9427-93d27ef7ea8e","order_by":1,"name":"Erol Görkem Hanoğlu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA30lEQVRIiWNgGAWjYFACHjApw8DeAKQMLIjXwsPAcwCkRYIULRIJIJoILfz8aw9+Lqio5TG4+fzqhh8FEgz87d0JeLVIzniXLD3jzHEeg9s5ZTd7gA6TOHN2A14tBjfOGEjzth0DaUm7wQPUYiCRi1+L/Y0zxr/BWm6eSbv5hxgtBvw9ZkBbangMbrAfu02ULRI3eMysec4c4JE8k8N2W8ZAgoegX/j7zxjf5qmok+M7fvzZzTd/bOT423vxa4FGx2Eg5jEAsXjwKwdbcwBE1gEx+wPCqkfBKBgFo2BEAgDpDkc57YKpNAAAAABJRU5ErkJggg==","orcid":"","institution":"Konya Food and Agriculture University","correspondingAuthor":true,"prefix":"","firstName":"Erol","middleName":"Görkem","lastName":"Hanoğlu","suffix":""},{"id":527097108,"identity":"610af6f7-052c-4adb-a0dd-9166d0fd6a9a","order_by":2,"name":"Ayşe Gül Özkan","email":"","orcid":"","institution":"Konya Food and Agriculture University","correspondingAuthor":false,"prefix":"","firstName":"Ayşe","middleName":"Gül","lastName":"Özkan","suffix":""},{"id":527097109,"identity":"c04c1a73-5c88-4067-8532-74552ad8c9ba","order_by":3,"name":"Kasım Öztoprak","email":"","orcid":"","institution":"Konya Food and Agriculture University","correspondingAuthor":false,"prefix":"","firstName":"Kasım","middleName":"","lastName":"Öztoprak","suffix":""}],"badges":[],"createdAt":"2025-10-07 09:23:22","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7797886/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7797886/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":93452040,"identity":"b98c1c2a-f80b-47df-b25b-46fdf2a09ef1","added_by":"auto","created_at":"2025-10-14 03:49:55","extension":"json","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6348,"visible":true,"origin":"","legend":"","description":"","filename":"364f5a831d044f66a9747ab8534904d7.json","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/b470c4f572980942523a1959.json"},{"id":93452043,"identity":"535503b2-abaa-4d17-b920-d896f9611958","added_by":"auto","created_at":"2025-10-14 03:49:55","extension":"xml","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":115425,"visible":true,"origin":"","legend":"","description":"","filename":"364f5a831d044f66a9747ab8534904d71enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/87dc364cf196b8f026a70a33.xml"},{"id":93452721,"identity":"a37cebd1-57f6-452f-b092-b5b7691ae7cf","added_by":"auto","created_at":"2025-10-14 03:57:55","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":635455,"visible":true,"origin":"","legend":"","description":"","filename":"BenchmarkingOCRandVisionLanguageModelsforTurkishTextRecognitionAComprehensiveEvaluationUsingSyntheticData.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/3a566cfa34681bc297a8056b.pdf"},{"id":93452720,"identity":"2465b3dc-f6fc-40eb-a573-22df88ef8460","added_by":"auto","created_at":"2025-10-14 03:57:55","extension":"zip","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":103605,"visible":true,"origin":"","legend":"","description":"","filename":"ESM1.zip","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/06fcd41626b0b9c488c29569.zip"},{"id":93452045,"identity":"74353be6-54d9-4ad9-878e-d04083c6d4bb","added_by":"auto","created_at":"2025-10-14 03:49:55","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":59478,"visible":true,"origin":"","legend":"","description":"","filename":"aysegulozkan.png","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/17e07f836f18ca93f6e7b399.png"},{"id":93452719,"identity":"70336de0-182a-4bd8-87c1-8c8f4c6d95ae","added_by":"auto","created_at":"2025-10-14 03:57:55","extension":"pdf","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":68661,"visible":true,"origin":"","legend":"","description":"","filename":"editorletter.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/a2894d068186b8871868a065.pdf"},{"id":93452042,"identity":"0a79a852-e5bc-4294-b28c-72d3d055060d","added_by":"auto","created_at":"2025-10-14 03:49:55","extension":"eps","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2890,"visible":true,"origin":"","legend":"","description":"","filename":"empty.eps","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/94ff14d63e08503023e734c1.eps"},{"id":93452047,"identity":"376f7592-cadb-480c-a102-5c10e1ec5909","added_by":"auto","created_at":"2025-10-14 03:49:55","extension":"jpeg","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":12736,"visible":true,"origin":"","legend":"","description":"","filename":"erolgorkemhanoglu.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/d68b51717e1d9879c4a1e83b.jpeg"},{"id":93452068,"identity":"698a72d8-4310-4569-99a7-160094b07306","added_by":"auto","created_at":"2025-10-14 03:49:55","extension":"eps","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":91593,"visible":true,"origin":"","legend":"","description":"","filename":"fig.eps","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/9e8dfb9b4b80fdd8467352c3.eps"},{"id":93453841,"identity":"1ec83a40-af6c-48b9-ada3-37475be87323","added_by":"auto","created_at":"2025-10-14 04:13:55","extension":"jpg","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":13492,"visible":true,"origin":"","legend":"","description":"","filename":"kasimoztoprak.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/8cb306650d41a95761d11ae8.jpg"},{"id":93452069,"identity":"be9b67b3-abc0-4b67-b67f-f4693e2e6a81","added_by":"auto","created_at":"2025-10-14 03:49:57","extension":"bst","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":146013,"visible":true,"origin":"","legend":"","description":"","filename":"snapacite.bst","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/a25d955bd5bfdd1a35234271.bst"},{"id":93453842,"identity":"037f5570-7f0f-4ac2-8cda-f7a6ad7267e9","added_by":"auto","created_at":"2025-10-14 04:13:55","extension":"bst","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":29828,"visible":true,"origin":"","legend":"","description":"","filename":"snaps.bst","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/173c300eb75c16e0657a744b.bst"},{"id":93452061,"identity":"28aa570f-7238-4b3d-8274-26718a5275dd","added_by":"auto","created_at":"2025-10-14 03:49:55","extension":"pdf","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":421391,"visible":true,"origin":"","legend":"","description":"","filename":"snarticle.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/577662d5f72fc04d999159e9.pdf"},{"id":93452051,"identity":"bc620066-6be1-4fb8-8832-67400df166cf","added_by":"auto","created_at":"2025-10-14 03:49:55","extension":"bst","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":35515,"visible":true,"origin":"","legend":"","description":"","filename":"snbasic.bst","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/1310ebd4679fd62e8037878a.bst"},{"id":93452723,"identity":"9840cb34-1d20-41a3-ae16-1c5fcf6be860","added_by":"auto","created_at":"2025-10-14 03:57:55","extension":"bst","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":33968,"visible":true,"origin":"","legend":"","description":"","filename":"snchicago.bst","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/dd95981f9de5cd7c1d45e15b.bst"},{"id":93453295,"identity":"e4257c3a-7053-4173-b6c4-e5e4871609f3","added_by":"auto","created_at":"2025-10-14 04:05:55","extension":"cls","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":55857,"visible":true,"origin":"","legend":"","description":"","filename":"snjnl.cls","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/6c7fb446fadba13546678b21.cls"},{"id":93452058,"identity":"dd038dff-c160-46e8-8f93-ca1eb86e0a9c","added_by":"auto","created_at":"2025-10-14 03:49:55","extension":"bst","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":64023,"visible":true,"origin":"","legend":"","description":"","filename":"snmathphysay.bst","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/ce9e2796fcf6895870ccb1b4.bst"},{"id":93452726,"identity":"a5b9e36d-4b8b-4f1f-a4ad-e12e8e3033bb","added_by":"auto","created_at":"2025-10-14 03:57:55","extension":"bst","order_by":25,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":64166,"visible":true,"origin":"","legend":"","description":"","filename":"snmathphysnum.bst","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/4c7174c8141dc414c18696a9.bst"},{"id":93453297,"identity":"e9f4c12f-6a11-4276-a60b-26460ad862c7","added_by":"auto","created_at":"2025-10-14 04:05:55","extension":"bst","order_by":26,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":37333,"visible":true,"origin":"","legend":"","description":"","filename":"snnature.bst","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/a93cb98f3386cda39c89dfdf.bst"},{"id":93452048,"identity":"d480215d-7a49-4560-bd71-145019f147d4","added_by":"auto","created_at":"2025-10-14 03:49:55","extension":"bst","order_by":27,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":39951,"visible":true,"origin":"","legend":"","description":"","filename":"snvancouveray.bst","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/13a33b41ee6c7d428adc21c1.bst"},{"id":93452056,"identity":"26f9025e-f7ef-4608-b360-a21f16602a24","added_by":"auto","created_at":"2025-10-14 03:49:55","extension":"bst","order_by":28,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":40758,"visible":true,"origin":"","legend":"","description":"","filename":"snvancouvernum.bst","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/866caba337c555411ace8415.bst"},{"id":93452730,"identity":"3c37cde5-9dd7-4af4-99f7-ca9580a12bd1","added_by":"auto","created_at":"2025-10-14 03:57:55","extension":"pdf","order_by":29,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":418495,"visible":true,"origin":"","legend":"","description":"","filename":"usermanual.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/aebfc38571172d8016a4f75e.pdf"},{"id":93453298,"identity":"c3caaeb9-e655-47f8-8541-174020c34b21","added_by":"auto","created_at":"2025-10-14 04:05:55","extension":"jpg","order_by":30,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":76625,"visible":true,"origin":"","legend":"","description":"","filename":"yasinyilmaz.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/06ebd1a089cbe6b056ad453e.jpg"},{"id":93453300,"identity":"0a3f16b9-4183-43dd-bfcc-1ab6b1852263","added_by":"auto","created_at":"2025-10-14 04:05:55","extension":"png","order_by":31,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":51373,"visible":true,"origin":"","legend":"","description":"","filename":"Onlineaysegulozkan.png","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/6739ea5c6e34be145357db97.png"},{"id":93452732,"identity":"3151efb3-d413-46ca-8da6-c9e44d021a17","added_by":"auto","created_at":"2025-10-14 03:57:55","extension":"png","order_by":32,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":69164,"visible":true,"origin":"","legend":"","description":"","filename":"Onlineerolgorkemhanoglu.png","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/a178258d540241f206ce93ad.png"},{"id":93452729,"identity":"d411b62c-3e25-4660-95c8-e889a40c39a5","added_by":"auto","created_at":"2025-10-14 03:57:55","extension":"png","order_by":33,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":57675,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinekasimoztoprak.png","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/4b3a357a9f83daa4807d546b.png"},{"id":93452066,"identity":"7ca74b6f-9679-4241-9e6c-ff2f8e80efbe","added_by":"auto","created_at":"2025-10-14 03:49:55","extension":"png","order_by":34,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":192669,"visible":true,"origin":"","legend":"","description":"","filename":"Onlineyasinyilmaz.png","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/64d6e80d7c8f82e76b2087c1.png"},{"id":93452067,"identity":"e0b6a88f-f46d-4206-8ecf-5de1e3af122a","added_by":"auto","created_at":"2025-10-14 03:49:55","extension":"xml","order_by":35,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":130725,"visible":true,"origin":"","legend":"","description":"","filename":"364f5a831d044f66a9747ab8534904d71structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/e7ea647b428f61f4af3f8ad6.xml"},{"id":93452064,"identity":"0606c57a-0487-4c5b-b25d-9134ce3f5296","added_by":"auto","created_at":"2025-10-14 03:49:55","extension":"html","order_by":36,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":133542,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/32c6081d98cd2bf9d858551c.html"},{"id":93454062,"identity":"0028c0eb-6706-456e-b8a6-666dc6a7e386","added_by":"auto","created_at":"2025-10-14 04:21:56","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":653586,"visible":true,"origin":"","legend":"","description":"","filename":"BenchmarkingOCRandVisionLanguageModelsforTurkishTextRecognitionAComprehensiveEvaluationUsingSyntheticData.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1_covered_fc4b87c4-b45f-44a9-ba3e-14cda6f3f3e5.pdf"},{"id":93452041,"identity":"3c17beb6-dc97-43e0-8be5-971953e48775","added_by":"auto","created_at":"2025-10-14 03:49:55","extension":"zip","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":103605,"visible":true,"origin":"","legend":"","description":"","filename":"ESM1.zip","url":"https://assets-eu.researchsquare.com/files/rs-7797886/v1/b52cb1fb86bd33c48626542b.zip"}],"financialInterests":"No competing interests reported.","formattedTitle":"Benchmarking OCR and Vision-Language Models for Turkish Text Recognition: A Comprehensive Evaluation Using Synthetic Data","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"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":"OCR, VLMs, Turkish Text Recognition, Synthetic Dataset","lastPublishedDoi":"10.21203/rs.3.rs-7797886/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7797886/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003ePurpose:\u003c/h2\u003e\u003cp\u003eWe present the first systematic benchmark evaluation of Optical Character Recognition (OCR) and Vision-Language Models (VLMs) for Turkish text recognition, addressing a critical gap in low-resource language processing. Turkish, with its agglutinative structure and unique characters (\u0026ccedil;, ğ, ı, İ, \u0026ouml;, ş, \u0026uuml;), poses challenges for models trained on high-resource languages such as English.\u003c/p\u003e\u003ch2\u003eMethods:\u003c/h2\u003e\u003cp\u003eWe developed a synthetic Turkish dataset of 6,600 images spanning three main text types: printed, handwritten, scene text. The dataset includes variations such as the presence of Turkish characters, effects of word length, sentence versus word recognition, and various distortion types (rotation, resolution, noise, and blur).Our evaluation compares three different model categories: traditional OCR systems, open-source VLMs, and commercial VLMs.\u003c/p\u003e\u003ch2\u003eResults:\u003c/h2\u003e\u003cp\u003eThe results show that modern VLMs significantly outperform traditional OCR approaches, with GPT-4o and Qwen2.5-VL models demonstrating superior performance. Notably, images containing Turkish-specific characters posed significant challenges for all models, with only GPT-4o maintaining stable performance. This highlights the critical impact of training dataset composition on multilingual performance. While the agglutinative word structure did not significantly affect recognition accuracy, handwritten text recognition remains a persistent challenge across all evaluated systems.\u003c/p\u003e\u003ch2\u003eConclusion:\u003c/h2\u003e\u003cp\u003eThe open-source Qwen2.5-VL model achieved comparable performance to the commercial GPT-4o despite having fewer parameters, showing strong potential as a computationally efficient alternative. This benchmark study establishes a standardized evaluation framework for Turkish text recognition research. To support future research in this domain, we publicly release the synthetic dataset, enabling reproducible research in low-resource language text recognition.\u003c/p\u003e","manuscriptTitle":"Benchmarking OCR and Vision-Language Models for Turkish Text Recognition: A Comprehensive Evaluation Using Synthetic Data","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-14 03:49:50","doi":"10.21203/rs.3.rs-7797886/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-03-19T02:01:54+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-01-09T04:01:25+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-01-06T11:29:04+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"226373759614334580681891770583844678234","date":"2025-11-18T13:13:08+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"240463666651013659749841748691322358103","date":"2025-11-14T05:46:41+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"76362316088674966657816684044760045096","date":"2025-11-13T08:48:09+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"101934192878012919000242992797690874797","date":"2025-11-12T21:38:11+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-11-11T15:18:13+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-10-09T12:03:00+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-10-09T02:19:09+00:00","index":"","fulltext":""},{"type":"submitted","content":"International Journal on Document Analysis and Recognition (IJDAR)","date":"2025-10-07T09:13:06+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":"cf658c42-573b-4a0b-b036-10ea9a4abe61","owner":[],"postedDate":"October 14th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-03T20:38:14+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-14 03:49:50","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7797886","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7797886","identity":"rs-7797886","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}

Text is read by the "Ask this paper" AI Q&A widget below. Extraction quality varies by source — PMC NXML preserves structure cleanly, OA-HTML may include some navigation residue, and OA-PDF can have broken hyphenation. The publisher copy (via DOI) is the canonical version.

My notes (saved in your browser only)

Ask this paper AI returns verbatim quotes from the full text · source: preprint-html

Answers must be backed by verbatim quotes from this paper's full text. Hallucinated quotes are dropped automatically; if no verbatim passage answers the question, we say so. How this works

Citation neighborhood (no data yet)

We don't have any in-corpus citations linked to this paper yet. This is a recent paper (2025) — citers typically take a year or two to land, and the OpenAlex reference graph may still be filling in.

Source provenance

europepmc
last seen: 2026-05-20T01:45:00.602351+00:00