Scene Text Recognition via Alternating Hierarchical-Global Attention in Encoder-Only Transformers

Scene Text Recognition via Alternating Hierarchical-Global Attention in Encoder-Only Transformers | 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 Scene Text Recognition via Alternating Hierarchical-Global Attention in Encoder-Only Transformers Shashank B N, S. Nagesh Bhattu, Sri Phani Krishna K This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7585117/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 4 You are reading this latest preprint version Abstract Scene Text Recognition (STR) plays a pivotal role in various real-world applications, such as autonomous driving, document analysis, and assistive technologies, by enabling machines to extract and interpret text from natural images. Despite significant advancements in deep learning, existing transformer-based STR models, often suffer from high computational costs, large memory requirements, and slow inference due to the autoregressive nature of decoder-based architectures. This paper proposes two model variants, FasterViTSTR and DualFasterViTSTR, to address these challenges. FasterViTSTR is built upon the FasterViT architecture, modified to an encoder-only design for efficient STR. FasterViTSTR utilizes hierarchical attention to capture both local and global features, enhancing the model's ability to process variable-length sequences and complex spatial layouts. DualFasterViTSTR introduces a novel alternating hybrid attention mechanism, alternating between hierarchical and global attention layers, not only enhances performance but also improves efficiency. The proposed models are evaluated on multiple STR benchmark datasets, where the FasterViTSTR V2 variant achieves up to a 0.9% improvement in word accuracy over existing baselines, while the DualFasterViTSTR V2 variant delivers up to a 2.11% improvement, along with reduced inference time, FLOPs, and parameter count compared to FasterViTSTR. The implementation is made available in the repository : github.com/ShashNagendra/STR-via-Alternating-Hierarchical-Global-Attention . Attention Scene Text Recognition Vision Transformers Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviewers invited by journal 20 Sep, 2025 Editor assigned by journal 11 Sep, 2025 Submission checks completed at journal 11 Sep, 2025 First submitted to journal 10 Sep, 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-7585117","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":523105088,"identity":"5c772b95-dca8-4f07-a29c-5d713166f71b","order_by":0,"name":"Shashank B N","email":"","orcid":"","institution":"National Institute of Technology Andhra Pradesh","correspondingAuthor":false,"prefix":"","firstName":"Shashank","middleName":"B","lastName":"N","suffix":""},{"id":523105090,"identity":"8eba2b21-4374-47ad-ad54-6b77657498a9","order_by":1,"name":"S. Nagesh Bhattu","email":"data:image/png;base64,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","orcid":"","institution":"National Institute of Technology Andhra Pradesh","correspondingAuthor":true,"prefix":"","firstName":"S.","middleName":"Nagesh","lastName":"Bhattu","suffix":""},{"id":523105092,"identity":"20d7c763-8334-42a6-aae0-ba938e81f660","order_by":2,"name":"Sri Phani Krishna K","email":"","orcid":"","institution":"National Institute of Technology Andhra Pradesh","correspondingAuthor":false,"prefix":"","firstName":"Sri","middleName":"Phani Krishna","lastName":"K","suffix":""}],"badges":[],"createdAt":"2025-09-10 17:08:27","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7585117/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7585117/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":92578340,"identity":"2fe3e2fa-f627-4240-8cd9-87cbc0917613","added_by":"auto","created_at":"2025-10-01 08:45:23","extension":"json","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5273,"visible":true,"origin":"","legend":"","description":"","filename":"64d3de10c7704d089e53c448b06622db.json","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/1f64813d21ecbc468eea0979.json"},{"id":92578349,"identity":"dfb2fc27-5a0c-4a2c-9102-b51e9edf2413","added_by":"auto","created_at":"2025-10-01 08:45:23","extension":"xml","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":124502,"visible":true,"origin":"","legend":"","description":"","filename":"64d3de10c7704d089e53c448b06622db1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/f115831c1797f514410235c6.xml"},{"id":92579202,"identity":"240ba598-b51a-4ab1-a7a0-3fad025db902","added_by":"auto","created_at":"2025-10-01 09:01:23","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":406504,"visible":true,"origin":"","legend":"","description":"","filename":"Paper6SignalImageandVideoProcessing1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/edb6679408a484239ee91e19.pdf"},{"id":92578726,"identity":"bc01c67d-7a72-4cc5-816a-e6cc4f411970","added_by":"auto","created_at":"2025-10-01 08:53:23","extension":"png","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":136098,"visible":true,"origin":"","legend":"","description":"","filename":"Picture1ViTSTR.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/b1a7f5f76dcd4986d8d8b5ad.png"},{"id":92578341,"identity":"ae5ea6f3-fc5d-42f8-8405-81547642deb4","added_by":"auto","created_at":"2025-10-01 08:45:23","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":33070,"visible":true,"origin":"","legend":"","description":"","filename":"STRFramework1.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/e93be9893f7bd75479fa947e.png"},{"id":92578342,"identity":"4e60fad9-046a-462d-98db-8b83c43b2562","added_by":"auto","created_at":"2025-10-01 08:45:23","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":26097,"visible":true,"origin":"","legend":"","description":"","filename":"STRFramework2.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/eb59a09d60775e74d99c44fc.png"},{"id":92578348,"identity":"35ec734f-1ad3-440c-a8c7-ecdd1e165f04","added_by":"auto","created_at":"2025-10-01 08:45:23","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":178510,"visible":true,"origin":"","legend":"","description":"","filename":"VANSTR.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/93c8cae05d3818927b92dc7d.png"},{"id":92578344,"identity":"e14de2bb-9d1e-4108-b9a5-759e24c73852","added_by":"auto","created_at":"2025-10-01 08:45:23","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":146557,"visible":true,"origin":"","legend":"","description":"","filename":"VANSTR2.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/b154a9fec4f52689f363f406.png"},{"id":92578730,"identity":"7a9a10ac-281d-4a4b-8c20-bf4e3370366c","added_by":"auto","created_at":"2025-10-01 08:53:23","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":251291,"visible":true,"origin":"","legend":"","description":"","filename":"almostfinal.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/c40fab940d1a7fe6bf27d10e.png"},{"id":92579203,"identity":"2d3eb35a-ce52-477a-b366-a5bf9b189c45","added_by":"auto","created_at":"2025-10-01 09:01:23","extension":"pdf","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":69686,"visible":true,"origin":"","legend":"","description":"","filename":"coverletter.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/2df7937fba5be546bd3bdec1.pdf"},{"id":92578346,"identity":"8e9a43f9-3704-4d10-b43f-087e91f53012","added_by":"auto","created_at":"2025-10-01 08:45:23","extension":"eps","order_by":10,"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-7585117/v1/dee3a0db6145f281430437de.eps"},{"id":92578351,"identity":"5b54dad4-a78c-4a39-bbeb-b552969600bf","added_by":"auto","created_at":"2025-10-01 08:45:23","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":100137,"visible":true,"origin":"","legend":"","description":"","filename":"fasterViTSTR.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/59a6927e48ebc6476bec63d7.png"},{"id":92578359,"identity":"fc79000c-1047-4e2e-aee3-b0834a62cde7","added_by":"auto","created_at":"2025-10-01 08:45:23","extension":"eps","order_by":12,"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-7585117/v1/505c75c99b9f3d88df425532.eps"},{"id":92578360,"identity":"4d7e9e94-a3f5-4c00-a0a7-73c8e46f9185","added_by":"auto","created_at":"2025-10-01 08:45:23","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":184190,"visible":true,"origin":"","legend":"","description":"","filename":"final2.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/6c537448d481a36f7dfce5b1.png"},{"id":92578353,"identity":"46b11aed-d062-4103-a787-7fe16becc2a4","added_by":"auto","created_at":"2025-10-01 08:45:23","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":55041,"visible":true,"origin":"","legend":"","description":"","filename":"framework.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/db6a318a262486b0df67f58e.png"},{"id":92578729,"identity":"66546bf1-39ec-4d60-97d7-26702a52ce74","added_by":"auto","created_at":"2025-10-01 08:53:23","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":33207,"visible":true,"origin":"","legend":"","description":"","filename":"framework1.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/c49efc5cc8df1b7c12ef2795.png"},{"id":92578733,"identity":"4a004251-0322-4996-ab3f-60ac1739aae1","added_by":"auto","created_at":"2025-10-01 08:53:23","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":26363,"visible":true,"origin":"","legend":"","description":"","filename":"framework2.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/0e221e1a24d5fa42fd90b007.png"},{"id":92578741,"identity":"7f84b9fc-e596-468b-b90e-e4141067e2f5","added_by":"auto","created_at":"2025-10-01 08:53:24","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":463196,"visible":true,"origin":"","legend":"","description":"","filename":"hybrid.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/343e563d4e55ccfc92565d10.png"},{"id":92578356,"identity":"4b88d34f-9cf2-49bc-ad6d-e2a92c515ff1","added_by":"auto","created_at":"2025-10-01 08:45:23","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":464622,"visible":true,"origin":"","legend":"","description":"","filename":"hybridglobalattention.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/5293449a9374b04e1003cfd5.png"},{"id":92578366,"identity":"d380ca50-e2fb-4fa6-91e6-f6c417fe8a0a","added_by":"auto","created_at":"2025-10-01 08:45:23","extension":"bst","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":147148,"visible":true,"origin":"","legend":"","description":"","filename":"snapacite.bst","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/c3d252fe8895ce292ce6d6a0.bst"},{"id":92579204,"identity":"41a0fc53-3d37-4747-a291-e13380202e4d","added_by":"auto","created_at":"2025-10-01 09:01:23","extension":"bst","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":30423,"visible":true,"origin":"","legend":"","description":"","filename":"snaps.bst","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/8ce0fcc7113fe8f5143b8262.bst"},{"id":92578385,"identity":"bc0ed692-cb69-4251-bbb7-9e855a47a582","added_by":"auto","created_at":"2025-10-01 08:45:24","extension":"pdf","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":390370,"visible":true,"origin":"","legend":"","description":"","filename":"snarticle.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/951fafc4a2c29cb15fc56d2b.pdf"},{"id":92578734,"identity":"93041f8c-ae0a-4579-a482-e1f92edb5a5e","added_by":"auto","created_at":"2025-10-01 08:53:23","extension":"bst","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":35733,"visible":true,"origin":"","legend":"","description":"","filename":"snbasic.bst","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/0496cd44a6197018c53ed2ec.bst"},{"id":92578357,"identity":"8c4e672a-51ca-4fbe-949c-a522c493e2cc","added_by":"auto","created_at":"2025-10-01 08:45:23","extension":"bst","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":40398,"visible":true,"origin":"","legend":"","description":"","filename":"snchicago.bst","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/ddba7efdde6d6ab863a7e8e3.bst"},{"id":92578735,"identity":"5166eddb-bb78-4a21-aa9b-08409d0c1877","added_by":"auto","created_at":"2025-10-01 08:53:23","extension":"cls","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":55331,"visible":true,"origin":"","legend":"","description":"","filename":"snjnl.cls","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/a8b8762cfddc1e0de20da203.cls"},{"id":92578737,"identity":"52bd6b9d-0844-4508-8c59-f301bcc579f5","added_by":"auto","created_at":"2025-10-01 08:53:24","extension":"bst","order_by":25,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":64140,"visible":true,"origin":"","legend":"","description":"","filename":"snmathphysay.bst","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/9b73aea1c96c601c80d1ee2b.bst"},{"id":92578738,"identity":"462c9dba-e4c0-4cba-8646-bcb6f521c50a","added_by":"auto","created_at":"2025-10-01 08:53:24","extension":"bst","order_by":26,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":64141,"visible":true,"origin":"","legend":"","description":"","filename":"snmathphysnum.bst","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/974f5a29639f91c9034dd63e.bst"},{"id":92578362,"identity":"3c170901-e499-40a6-8a17-82e5ef7db7c1","added_by":"auto","created_at":"2025-10-01 08:45:23","extension":"bst","order_by":27,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":38349,"visible":true,"origin":"","legend":"","description":"","filename":"snnature.bst","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/fd2b9622478407993dcaa4b9.bst"},{"id":92578363,"identity":"da2680de-ff0b-4354-82d4-2b5ebb4a7fe2","added_by":"auto","created_at":"2025-10-01 08:45:23","extension":"bst","order_by":28,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":41304,"visible":true,"origin":"","legend":"","description":"","filename":"snvancouver.bst","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/5562d23e841c9574f948cbed.bst"},{"id":92578372,"identity":"ad188750-2686-4fca-83a5-83d854852f9b","added_by":"auto","created_at":"2025-10-01 08:45:24","extension":"pdf","order_by":29,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":415247,"visible":true,"origin":"","legend":"","description":"","filename":"usermanual.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/26968c3af9315c6eea9758eb.pdf"},{"id":92578371,"identity":"63c87c02-5387-4113-8624-234664c4fecd","added_by":"auto","created_at":"2025-10-01 08:45:24","extension":"png","order_by":30,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":136990,"visible":true,"origin":"","legend":"","description":"","filename":"vitstr.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/ab0a318604d29a241f515f23.png"},{"id":92579205,"identity":"1fdae407-167b-41ab-b295-9cc581c6ccfe","added_by":"auto","created_at":"2025-10-01 09:01:24","extension":"png","order_by":31,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":111191,"visible":true,"origin":"","legend":"","description":"","filename":"vitstr2.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/2b7810d00a48d28944c73c01.png"},{"id":92578365,"identity":"f58fc364-af32-4a2c-b844-c697c25a9a7c","added_by":"auto","created_at":"2025-10-01 08:45:23","extension":"png","order_by":32,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":70823,"visible":true,"origin":"","legend":"","description":"","filename":"OnlinePicture1ViTSTR.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/d43cfb955f7a17f6ecb13283.png"},{"id":92578369,"identity":"554535b2-112a-49ef-978d-a9ed73f6aff6","added_by":"auto","created_at":"2025-10-01 08:45:23","extension":"png","order_by":33,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":13999,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineSTRFramework1.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/21384a0ad690faefdaf73bb8.png"},{"id":92578370,"identity":"766f3fb1-8a5d-4bfa-a488-1f5a0e969a06","added_by":"auto","created_at":"2025-10-01 08:45:24","extension":"png","order_by":34,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":9405,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineSTRFramework2.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/63eae39e660e4f6ae67526e0.png"},{"id":92578368,"identity":"00b46a94-fe06-49bc-8c77-ef9711ed4f2a","added_by":"auto","created_at":"2025-10-01 08:45:23","extension":"png","order_by":35,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":104402,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineVANSTR.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/4bc5accf5632ff1a2adedb49.png"},{"id":92578386,"identity":"9e607ef0-512e-452f-9ab0-27372506bb47","added_by":"auto","created_at":"2025-10-01 08:45:24","extension":"png","order_by":36,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":59762,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineVANSTR2.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/51f6c99290a0de38b23339c8.png"},{"id":92578740,"identity":"5c786f67-3880-40d1-bf8c-181ef94dc144","added_by":"auto","created_at":"2025-10-01 08:53:24","extension":"png","order_by":37,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":166196,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinealmostfinal.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/2432335bd6012ab7ab8fed27.png"},{"id":92578383,"identity":"64486227-e48c-47ad-ba84-830104968b88","added_by":"auto","created_at":"2025-10-01 08:45:24","extension":"png","order_by":39,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":78626,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefinal2.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/ae71a1a1583c0e3dedbbffa4.png"},{"id":92578376,"identity":"b01b3e9d-dbda-468b-9360-ae886a109ef4","added_by":"auto","created_at":"2025-10-01 08:45:24","extension":"png","order_by":40,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":19051,"visible":true,"origin":"","legend":"","description":"","filename":"Onlineframework.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/06bfe4fe87f46fa54187c9b3.png"},{"id":92578380,"identity":"518888eb-f0dc-4758-8336-43053c2173c3","added_by":"auto","created_at":"2025-10-01 08:45:24","extension":"png","order_by":41,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":14160,"visible":true,"origin":"","legend":"","description":"","filename":"Onlineframework1.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/a92948dfc3f1f19e7914dc6b.png"},{"id":92578742,"identity":"3dd5549d-80f7-48c5-8270-ab04a2a9eb69","added_by":"auto","created_at":"2025-10-01 08:53:24","extension":"png","order_by":42,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":9555,"visible":true,"origin":"","legend":"","description":"","filename":"Onlineframework2.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/4a212e8a98b9e7c00030cbf1.png"},{"id":92579206,"identity":"2917370f-fe90-4357-8d05-65452b2f455c","added_by":"auto","created_at":"2025-10-01 09:01:24","extension":"png","order_by":44,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":90193,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinehybridglobalattention.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/aeff0747d66ede82f32719d7.png"},{"id":92578747,"identity":"b8471bc9-185e-43f3-86cf-d1653140b318","added_by":"auto","created_at":"2025-10-01 08:53:24","extension":"png","order_by":45,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":71931,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinevitstr.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/d32483bc771866c4b71938ab.png"},{"id":92578375,"identity":"85553e30-5f0b-4b48-ad52-6802c6242575","added_by":"auto","created_at":"2025-10-01 08:45:24","extension":"png","order_by":46,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":28512,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinevitstr2.png","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/3b581a438f23834b0db00b18.png"},{"id":92578377,"identity":"1b558cb1-17c7-4d4f-b0ba-944a3d18b75c","added_by":"auto","created_at":"2025-10-01 08:45:24","extension":"xml","order_by":47,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":165581,"visible":true,"origin":"","legend":"","description":"","filename":"64d3de10c7704d089e53c448b06622db1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/afec24303cc81fec22ba2175.xml"},{"id":92578382,"identity":"d0a0ef91-f914-4b77-84b8-38acb32674b4","added_by":"auto","created_at":"2025-10-01 08:45:24","extension":"html","order_by":48,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":140592,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1/61857fcf09b6297fca1b6a79.html"},{"id":92579210,"identity":"1a239718-63ef-4c85-877b-f62b1ff62d73","added_by":"auto","created_at":"2025-10-01 09:01:29","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":394654,"visible":true,"origin":"","legend":"","description":"","filename":"Paper6SignalImageandVideoProcessing1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7585117/v1_covered_2387e2f2-f0ce-4af0-a7e3-57bef9bc639e.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Scene Text Recognition via Alternating Hierarchical-Global Attention in Encoder-Only Transformers","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":"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":"Attention, Scene Text Recognition, Vision Transformers","lastPublishedDoi":"10.21203/rs.3.rs-7585117/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7585117/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eScene Text Recognition (STR) plays a pivotal role in various real-world applications, such as autonomous driving, document analysis, and assistive technologies, by enabling machines to extract and interpret text from natural images. Despite significant advancements in deep learning, existing transformer-based STR models, often suffer from high computational costs, large memory requirements, and slow inference due to the autoregressive nature of decoder-based architectures. This paper proposes two model variants, FasterViTSTR and DualFasterViTSTR, to address these challenges. FasterViTSTR is built upon the FasterViT architecture, modified to an encoder-only design for efficient STR. FasterViTSTR utilizes hierarchical attention to capture both local and global features, enhancing the model's ability to process variable-length sequences and complex spatial layouts. DualFasterViTSTR introduces a novel alternating hybrid attention mechanism, alternating between hierarchical and global attention layers, not only enhances performance but also improves efficiency. The proposed models are evaluated on multiple STR benchmark datasets, where the FasterViTSTR V2 variant achieves up to a 0.9% improvement in word accuracy over existing baselines, while the DualFasterViTSTR V2 variant delivers up to a 2.11% improvement, along with reduced inference time, FLOPs, and parameter count compared to FasterViTSTR. The implementation is made available in the repository : \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003egithub.com/ShashNagendra/STR-via-Alternating-Hierarchical-Global-Attention\u003c/span\u003e\u003cspan address=\"https://github.com/ShashNagendra/STR-via-Alternating-Hierarchical-Global-Attention\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e","manuscriptTitle":"Scene Text Recognition via Alternating Hierarchical-Global Attention in Encoder-Only Transformers","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-01 08:45:18","doi":"10.21203/rs.3.rs-7585117/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewersInvited","content":"","date":"2025-09-20T23:13:33+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-11T13:23:37+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-09-11T13:23:16+00:00","index":"","fulltext":""},{"type":"submitted","content":"Signal, Image and Video Processing","date":"2025-09-10T17:02:18+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":"3f1b0562-1d89-448c-a255-0f40445bee49","owner":[],"postedDate":"October 1st, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2025-10-01T08:45:18+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-01 08:45:18","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7585117","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7585117","identity":"rs-7585117","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}