Score-based Likelihood Ratios for Deepfake Image Evidence

Score-based Likelihood Ratios for Deepfake Image Evidence | 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 Score-based Likelihood Ratios for Deepfake Image Evidence Tianli Guo, Jisong Li, Yunqi Tang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8551375/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 8 You are reading this latest preprint version Abstract Deepfake technology's rapid advancement challenges forensic evidence evaluation. Conventional binary detection models use fixed thresholds to classify images but lack probabilistic interpretation and cannot quantify evidential strength, limiting their forensic applicability. To address this, we propose a score-based likelihood ratio framework that combines deep feature extraction with probabilistic modeling for deepfake authentication.We created real-real and real-fake image pairs, extracted 2048-dimensional features using the SRM network, and computed pairwise similarity scores. Kernel density estimation modeled the score distributions for both pair types, with bandwidth optimized via the Silverman rule and ten-fold cross-validation for robustness. This produced an LR model that converts similarity scores into quantifiable evidential strength, enabling hypothesis testing between real and deepfake images. On the test set, the framework achieved an EER of 0.0192. Applying the Pool Adjacent Violators Algorithm reduced the classification log-likelihood ratio cost from 0.0202 to 0.0128. DET, Tippett, and ECE analyses confirmed satisfactory discrimination and calibration performance on the tested dataset. We further enhanced the model using the ELUB method, which helps improve the statistical consistency of LR outputs in the experimental setting. This study presents a rigorous, computation-driven approach to quantifying deepfake evidence. By integrating deep learning with calibrated probabilistic modeling, it aims to provide reliable and interpretable evidential strength, contributing to the potential advancement of LR-based analysis in legal and forensic settings. Physical sciences/Engineering Physical sciences/Mathematics and computing Deepfake images similarity score deep learning feature kernel density estimation likelihood ratio Full Text Additional Declarations No competing interests reported. Supplementary Files 0210SupplementaryFile.doc Cite Share Download PDF Status: Under Review Version 1 posted Reviewers agreed at journal 24 Feb, 2026 Reviewers agreed at journal 22 Feb, 2026 Reviewers agreed at journal 21 Feb, 2026 Reviewers invited by journal 20 Feb, 2026 Editor assigned by journal 12 Feb, 2026 Editor invited by journal 12 Feb, 2026 Submission checks completed at journal 09 Feb, 2026 First submitted to journal 09 Feb, 2026 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-8551375","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":596313885,"identity":"519c9c48-58c3-4889-99b5-4738b2cac551","order_by":0,"name":"Tianli Guo","email":"","orcid":"","institution":"People's Public Security University of China","correspondingAuthor":false,"prefix":"","firstName":"Tianli","middleName":"","lastName":"Guo","suffix":""},{"id":596313886,"identity":"67a3f199-a6a0-4623-8fe6-1b4fa54d1820","order_by":1,"name":"Jisong Li","email":"","orcid":"","institution":"People's Public Security University of China","correspondingAuthor":false,"prefix":"","firstName":"Jisong","middleName":"","lastName":"Li","suffix":""},{"id":596313887,"identity":"61775e38-0f3b-4e26-807c-7442eb829fdd","order_by":2,"name":"Yunqi Tang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA7UlEQVRIiWNgGAWjYDADNvYGGDMBv0oeuBaeA6RqYZCAqySgxZ797OHXvG13ZPsk3x57+LPNjoGfPceA4ecOPLbw5KVZ87Y9M26Tzks35m1LZpDseWPA2HsGn8NyzIAqDye2SeeYSTO2HWAwuJFjwMzYhkcL/xuoFskzZpI/gVrsCWqRyDF+DNYiwWMmwQuyRYKQlhtvzBjnnDts3MYD9AvPuWQeiTPPCg724tHC3p9j/OFN2WHZ+e1njz38UWYnx9+evPHBTzxagIBNChg5jA0MPGyQ8ACCA3g1MDAwf/yBpGUUjIJRMApGAQYAAGmATFmF4SXeAAAAAElFTkSuQmCC","orcid":"","institution":"People's Public Security University of China","correspondingAuthor":true,"prefix":"","firstName":"Yunqi","middleName":"","lastName":"Tang","suffix":""}],"badges":[],"createdAt":"2026-01-08 13:52:12","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8551375/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8551375/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":103406859,"identity":"c4b428d0-9ebb-47ca-b3ab-a624f95cfdbd","added_by":"auto","created_at":"2026-02-25 10:13:06","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":677302,"visible":true,"origin":"","legend":"","description":"","filename":"0210ScorebasedLikelihoodRatiosforDeepfakeImageEvidenceUsingDeepLearningFeatures.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8551375/v1_covered_97f63d75-9c92-442a-b847-632778556a5a.pdf"},{"id":103406655,"identity":"273215ea-f02d-4838-9c49-99b29172e238","added_by":"auto","created_at":"2026-02-25 10:12:27","extension":"doc","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":43520,"visible":true,"origin":"","legend":"","description":"","filename":"0210SupplementaryFile.doc","url":"https://assets-eu.researchsquare.com/files/rs-8551375/v1/c0d50711320ee82a8725cf2d.doc"}],"financialInterests":"No competing interests reported.","formattedTitle":"Score-based Likelihood Ratios for Deepfake Image Evidence","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":"humanities-and-social-sciences-communications","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"palcomms","sideBox":"Learn more about [Humanities \u0026 Social Sciences Communications](http://www.nature.com/palcomms/)","snPcode":"41599","submissionUrl":"https://submission.springernature.com/new-submission/41599/3","title":"Humanities and Social Sciences Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Deepfake images, similarity, score, deep learning feature, kernel density estimation, likelihood ratio","lastPublishedDoi":"10.21203/rs.3.rs-8551375/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8551375/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eDeepfake technology's rapid advancement challenges forensic evidence evaluation. Conventional binary detection models use fixed thresholds to classify images but lack probabilistic interpretation and cannot quantify evidential strength, limiting their forensic applicability. To address this, we propose a score-based likelihood ratio framework that combines deep feature extraction with probabilistic modeling for deepfake authentication.We created real-real and real-fake image pairs, extracted 2048-dimensional features using the SRM network, and computed pairwise similarity scores. Kernel density estimation modeled the score distributions for both pair types, with bandwidth optimized via the Silverman rule and ten-fold cross-validation for robustness. This produced an LR model that converts similarity scores into quantifiable evidential strength, enabling hypothesis testing between real and deepfake images. On the test set, the framework achieved an EER of 0.0192. Applying the Pool Adjacent Violators Algorithm reduced the classification log-likelihood ratio cost from 0.0202 to 0.0128. DET, Tippett, and ECE analyses confirmed satisfactory discrimination and calibration performance on the tested dataset. We further enhanced the model using the ELUB method, which helps improve the statistical consistency of LR outputs in the experimental setting. This study presents a rigorous, computation-driven approach to quantifying deepfake evidence. By integrating deep learning with calibrated probabilistic modeling, it aims to provide reliable and interpretable evidential strength, contributing to the potential advancement of LR-based analysis in legal and forensic settings.\u003c/p\u003e","manuscriptTitle":"Score-based Likelihood Ratios for Deepfake Image Evidence","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-25 10:09:39","doi":"10.21203/rs.3.rs-8551375/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"317466663215417927735231340625838937583","date":"2026-02-24T12:51:31+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"86997360401068083968926497523132785944","date":"2026-02-22T17:20:39+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"269759184555793201094074966180007495521","date":"2026-02-21T09:27:28+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-02-20T14:49:58+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-02-12T13:30:58+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-02-12T09:51:49+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-02-09T08:10:12+00:00","index":"","fulltext":""},{"type":"submitted","content":"Humanities and Social Sciences Communications","date":"2026-02-09T07:53:28+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"humanities-and-social-sciences-communications","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"palcomms","sideBox":"Learn more about [Humanities \u0026 Social Sciences Communications](http://www.nature.com/palcomms/)","snPcode":"41599","submissionUrl":"https://submission.springernature.com/new-submission/41599/3","title":"Humanities and Social Sciences Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"fc952fbd-22ea-43b1-a8ef-0046d5fced1e","owner":[],"postedDate":"February 25th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":63451216,"name":"Physical sciences/Engineering"},{"id":63451217,"name":"Physical sciences/Mathematics and computing"}],"tags":[],"updatedAt":"2026-05-07T06:08:30+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-25 10:09:39","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8551375","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8551375","identity":"rs-8551375","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}