MS-Adapter: Multi-scaled Adapter for Efficient DeepFake Detection

MS-Adapter: Multi-scaled Adapter for Efficient DeepFake 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 MS-Adapter: Multi-scaled Adapter for Efficient DeepFake Detection Ruofan Wang, Aimin Pan, Vladimir Y Mariano This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6354040/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Existing deepfake detection methods overly rely on low-level forgery patterns, leading to poor performance when encounteringunseen forgery types or low-quality images. Recently, Vision Transformer (ViT) pretrained on large-scale datasets havedemonstrated strong generalization capabilities across various image downstream tasks. However, parameter-efficient fine tuning methods for ViTs have shown limited effectiveness in DeepFake detection, mainly because ViTs rely on high-levelsemantics while struggling to capture fine grained local details. To address this issue, this paper proposes MS-Adapter, amulti-scale adapter network for efficient deepfake detection. By embedding multi-scale adapter modules within the pretrainedViT, MS-Adapter progressively extracts and fuses features from low-level forgery artifacts to high-level semantic forgery patternsacross multiple scales. At the same time, the Temporal Aggregation Transformer receives the frame-level features extractedby the Multi-Scale Adapter and performs temporal modeling on these features to enhance forgery detection performance.Experimental results demonstrate that MS-Adapter achieves superior performance on multiple datasets, including FF++,Celeb-DFv2, and DFDC, while requiring only a small number of trainable parameters. Physical sciences/Mathematics and computing/Computer science Physical sciences/Mathematics and computing/Information technology Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted 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-6354040","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":449694123,"identity":"4505dac7-0f46-4656-ab10-05769a557ee3","order_by":0,"name":"Ruofan Wang","email":"","orcid":"","institution":"Xinzhou Normal University","correspondingAuthor":false,"prefix":"","firstName":"Ruofan","middleName":"","lastName":"Wang","suffix":""},{"id":449694124,"identity":"a73cbe4c-8d04-40dd-836b-3e3051449459","order_by":1,"name":"Aimin Pan","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAyklEQVRIiWNgGAWjYLCCBAYJBn4GBgMgk5kELZINJGkBAYMDxGoxZz978MPDHIs84xvJ2yQYKqwTG9jPHsCrxbInL1kicZtEsdmNtDIJhjPpiQ08eQkE3JNjANKSuO12jpkEY9vhxAYJHgP8Ws6/Mf4B0rJ5NkjLP2K03ACqBGnZIA3S0kCEFssZb8wsQFpm3H9WbJFwLN24jScHvxZz/hzjmz+31SX29xzeeONDjbVsP/sZAg5D4SUAMRte9RhaRsEoGAWjYBRgAwC3RkPosLmLhAAAAABJRU5ErkJggg==","orcid":"","institution":"Wuhan Donghu College","correspondingAuthor":true,"prefix":"","firstName":"Aimin","middleName":"","lastName":"Pan","suffix":""},{"id":449694125,"identity":"b68bb189-eb8c-40d5-866b-113811059dfc","order_by":2,"name":"Vladimir Y Mariano","email":"","orcid":"","institution":"National University","correspondingAuthor":false,"prefix":"","firstName":"Vladimir","middleName":"Y","lastName":"Mariano","suffix":""}],"badges":[],"createdAt":"2025-04-01 14:23:17","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6354040/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6354040/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":82553570,"identity":"8ffc0808-adf8-45c2-9300-220335a0328c","added_by":"auto","created_at":"2025-05-12 21:31:32","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":303484,"visible":true,"origin":"","legend":"","description":"","filename":"ScientificReportsMSAdapter.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6354040/v1_covered_0c353e03-9d5c-46ec-9b49-40a7bd0d92ed.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"MS-Adapter: Multi-scaled Adapter for Efficient DeepFake Detection","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"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":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-6354040/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6354040/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eExisting deepfake detection methods overly rely on low-level forgery patterns, leading to poor performance when encounteringunseen forgery types or low-quality images. Recently, Vision Transformer (ViT) pretrained on large-scale datasets havedemonstrated strong generalization capabilities across various image downstream tasks. However, parameter-efficient fine tuning methods for ViTs have shown limited effectiveness in DeepFake detection, mainly because ViTs rely on high-levelsemantics while struggling to capture fine grained local details. To address this issue, this paper proposes MS-Adapter, amulti-scale adapter network for efficient deepfake detection. By embedding multi-scale adapter modules within the pretrainedViT, MS-Adapter progressively extracts and fuses features from low-level forgery artifacts to high-level semantic forgery patternsacross multiple scales. At the same time, the Temporal Aggregation Transformer receives the frame-level features extractedby the Multi-Scale Adapter and performs temporal modeling on these features to enhance forgery detection performance.Experimental results demonstrate that MS-Adapter achieves superior performance on multiple datasets, including FF++,Celeb-DFv2, and DFDC, while requiring only a small number of trainable parameters.\u003c/p\u003e","manuscriptTitle":"MS-Adapter: Multi-scaled Adapter for Efficient DeepFake Detection","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-04-30 14:30:14","doi":"10.21203/rs.3.rs-6354040/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"998d90d6-04d8-4bfa-b04f-dd8f01026e99","owner":[],"postedDate":"April 30th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":47852107,"name":"Physical sciences/Mathematics and computing/Computer science"},{"id":47852108,"name":"Physical sciences/Mathematics and computing/Information technology"}],"tags":[],"updatedAt":"2025-05-12T21:23:25+00:00","versionOfRecord":[],"versionCreatedAt":"2025-04-30 14:30:14","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6354040","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6354040","identity":"rs-6354040","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}